PyPSA in one file#
One spec file, examples/pypsa.yaml, states the spec of the model that a
plain n.optimize() builds in the power system modelling framework PyPSA. The
file grows one step of the PyPSA ladder
(a rung) at a time. Each rung states one
n.optimize() keyword or feature in full.
The file also states two classes of model beyond a plain n.optimize(). The
two-stage stochastic class spans a scenario dimension, and
n.set_scenarios(...) switches it on (rung 14). The multi-period investment
class spans a period dimension, and the keyword multi_investment_periods
switches it on (rung 15).
Each rung table lists the rows PyPSA builds, linked to their entries in the
file. The rows and the line numbers are those of PyPSA master at commit
3625c816, under pypsa/optimization/. The banner of each rung states the
objective that PyPSA solved its reference network to.
The work that turns a PyPSA network into the data the file declares (data prep) is outside the file. Data prep lists the parameters whose values come from it.
Three rules apply throughout the file:
- Each bound is an explicit row, as PyPSA writes it, so its dual is the dual of a row.
- Each case of a PyPSA rule is a
where:condition on a column of data. - The names are those of PyPSA, spelled
Component_attribute. The typesetter prints them in the usual symbols, from the symbol tableexamples/symbols/pypsa.yaml.
Statuses#
A row has one of four statuses:
- A row is done where the file states it as the one entry that PyPSA builds.
- A row is split where the file states the same feasible region and optimum
in a different form, such as several
where:entries. - A row is out where the file leaves it out, because PyPSA builds it only under the keyword, scope or version in the note.
- A row is refused where PyPSA refuses the data, and Refusals lists what the file states for it.
A rung with no table adds no row. A name with {k}, {s}, {c} or {n} is a
family of rows that PyPSA numbers per segment, scenario, outaged component or
sub-network.
The rungs#
Every rung's network is
spine.build()plus the rung's ownn.addcalls, data inline; a keyword not passed is PyPSA's default. A banner states what PyPSA solved the rung to; how an engine attaches the network to the file, and what it makes of it, is that engine's own record.
The shared spine, spine.py
spine.py
"""The spine every rung starts from: two buses, a coal and a gas unit, one link, two loads.
Four hourly snapshots with three different weighting columns, none of them constant
and none 1.0, so a factor a formula drops or swaps cannot pass as identity.
"""
from __future__ import annotations
from datetime import datetime
#: Four hourly stamps — snapshots are timestamps, as PyPSA's are in practice and as the file declares them.
SNAPSHOTS = [datetime(2015, 1, 1, hour) for hour in range(4)]
WEIGHTINGS = {'objective': [2.0, 1.5, 2.5, 3.0], 'stores': [0.5, 2.0, 1.5, 2.5], 'generators': [1.5, 0.5, 3.0, 2.0]}
def build():
"""The spine as a fresh ``pypsa.Network``; each rung adds to what this returns."""
import pypsa
n = pypsa.Network()
n.set_snapshots(SNAPSHOTS)
for column, values in WEIGHTINGS.items():
n.snapshot_weightings[column] = values
n.add('Bus', 'north')
n.add('Bus', 'south')
n.add('Generator', 'coal', bus='north', p_nom=100, marginal_cost=10)
n.add('Generator', 'gas', bus='south', p_nom=100, marginal_cost=30)
n.add('Link', 'wire', bus0='north', bus1='south', p_nom=40, p_min_pu=-1, efficiency=0.9)
n.add('Load', 'north_load', bus='north', p_set=30)
n.add('Load', 'south_load', bus='south', p_set=40)
return n
Rung 1 — transport#
| PyPSA | status | note |
|---|---|---|
Generator-p, Link-p |
done | |
Generator-fix-p-lower |
done | |
Generator-fix-p-upper |
done | |
Link-fix-p-lower |
done | |
Link-fix-p-upper |
done | |
Bus-nodal_balance |
done | a loaded bus with nothing attached in Refusals |
Bus-meshed-*-nodal_balance |
out | the rows and duals of Bus-nodal_balance, split by component columns per bus into containers of the model-building library linopy, under the keyword meshed_thresholds, default [30, 100, 400] (#123) |
marginal_cost |
done | |
objective_constant |
split | an objective shift, which a test compares net of n._objective_constant, nonzero in rungs 11 (21915277.52) and 13 (160.0) |
Rungs 43, 50 and 60 extend the Bus-nodal_balance row.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective7182.222222222223, 45 rows.
The network, as PyPSA code
rung_01_transport.py
"""Rung 1: transport — two buses, two generators, one controllable link."""
from __future__ import annotations
from math import nan
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.links_t.p_set['wire'] = [10, nan, nan, nan]
n.add('Generator', 'must_run', bus='south', p_nom=10, marginal_cost=0, p_set=[5, 5, 5, 5])
return n
Rung 2 — storage#
| PyPSA | status | note |
|---|---|---|
StorageUnit-p_dispatch, -p_store, -state_of_charge, Store-e, Store-p |
done | |
StorageUnit-spill |
done | where: inflow > 0, missing: 0, with bounds on the variable, as PyPSA's |
StorageUnit-fix-*, Store-fix-e-* |
done | |
StorageUnit-energy_balance |
done | the charge carried into a snapshot as a quantity stated with cases: (cyclic, opening, carried), and (1-loss)**eh from data prep |
Store-energy_balance |
done | as StorageUnit-energy_balance |
StorageUnit-p_set, {c}-{attr}_set |
done | Generator-p_set, Link-p_set, StorageUnit-state_of_charge_set, Store-e_set and Line-s_set here, and Store-p_set, StorageUnit-p_dispatch_set and -p_store_set in rung 37 |
marginal_cost_storage, spill_cost |
done |
Rung 60 extends the StorageUnit-energy_balance row. Rung 76 extends the
Store-energy_balance row and adds the bounds of Store-p and Store-p_store.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective4456.659315422356, 111 rows.
The network, as PyPSA code
rung_02_storage.py
"""Rung 2: storage — a cyclic battery, an inflow reservoir with a set state of charge, and a store."""
from __future__ import annotations
from math import nan
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.generators_t.marginal_cost['gas'] = [15, 15, 60, 60]
n.add(
'StorageUnit',
'battery',
bus='south',
p_nom=20,
max_hours=4,
efficiency_store=0.95,
efficiency_dispatch=0.9,
standing_loss=0.01,
cyclic_state_of_charge=True,
marginal_cost=0.5,
p_set=[0, nan, nan, nan],
)
n.add(
'StorageUnit',
'reservoir',
bus='south',
p_nom=10,
max_hours=2,
spill_cost=2,
state_of_charge_initial=5,
marginal_cost_storage=0.1,
inflow=[12, 12, 12, 12],
state_of_charge_set=[nan, nan, nan, 10],
)
n.add(
'Store',
'cavern',
bus='south',
e_nom=40,
e_initial=25,
standing_loss=0.005,
marginal_cost=0.2,
e_set=[nan, nan, nan, 20],
)
return n
Rung 3 — expansion#
| PyPSA | status | note |
|---|---|---|
{c}-p_nom, -s_nom, -e_nom |
done | {c}_p_nom_ext for an extendable build, the parameter for a fixed build |
{c}-ext-{attr}-lower/upper |
done | |
{c}-ext-p_nom-lower/upper |
done | |
{c}-p_nom_set |
done | |
Generator-e_sum_min/max |
done | |
| capital cost | done | periodized_cost from data prep: capital_cost, or the annuity of overnight_cost, plus fom_cost |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective7633.908502024292, 200 rows.
The network, as PyPSA code
rung_03_expansion.py
"""Rung 3: expansion — extendable capacity, energy-sum bounds, fixed and set nominal capacities."""
from __future__ import annotations
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add('Bus', 'island')
n.add('Carrier', 'onwind')
n.add('Carrier', 'solarpv')
n.add('Carrier', 'dc')
n.add('Carrier', 'phs')
n.add('Carrier', 'h2')
n.add(
'Generator',
'wind',
bus='north',
carrier='onwind',
p_nom_extendable=True,
capital_cost=50,
p_nom_min=5,
p_nom_max=80,
marginal_cost=0,
e_sum_min=40,
ramp_limit_up=0.4,
ramp_limit_down=0.4,
p_max_pu=[0.3, 0.8, 0.5, 0.9],
)
n.add(
'Generator',
'solar',
bus='north',
carrier='solarpv',
p_nom_extendable=True,
capital_cost=60,
p_nom_max=40,
marginal_cost=0,
p_nom_set=15,
p_max_pu=[0.5, 0.6, 0.4, 0.2],
)
n.add('Generator', 'diesel', bus='island', marginal_cost=40, p_nom=60, e_sum_max=70)
n.add(
'Link',
'cable',
bus0='north',
bus1='island',
carrier='dc',
length=120,
p_nom_extendable=True,
capital_cost=20,
p_nom_max=30,
efficiency=0.95,
p_nom_set=25,
ramp_limit_up=0.3,
ramp_limit_down=0.3,
)
n.add('Load', 'island_load', bus='island', p_set=10)
n.add(
'StorageUnit',
'pump',
bus='north',
carrier='phs',
p_nom_extendable=True,
capital_cost=15,
p_nom_max=30,
max_hours=4,
efficiency_store=0.9,
efficiency_dispatch=0.9,
cyclic_state_of_charge=True,
p_nom_set=20,
)
n.add('StorageUnit', 'ice', bus='island', max_hours=2, p_nom=8, state_of_charge_initial=6)
n.add(
'Store',
'tank',
bus='north',
carrier='h2',
e_nom_extendable=True,
capital_cost=2,
e_nom_max=80,
e_cyclic=True,
e_nom_set=50,
)
n.add('Store', 'keg', bus='island', e_nom=15, e_initial=5)
n.add(
'GlobalConstraint',
'tech_wind',
type='tech_capacity_expansion_limit',
carrier_attribute='onwind',
sense='==',
constant=50,
)
n.add(
'GlobalConstraint',
'tech_solar',
type='tech_capacity_expansion_limit',
carrier_attribute='solarpv',
sense='>=',
constant=10,
)
n.add(
'GlobalConstraint',
'tech_dc',
type='tech_capacity_expansion_limit',
carrier_attribute='dc',
sense='<=',
constant=28,
)
n.add(
'GlobalConstraint',
'tech_phs',
type='tech_capacity_expansion_limit',
carrier_attribute='phs',
sense='<=',
constant=25,
)
n.add(
'GlobalConstraint',
'tech_h2',
type='tech_capacity_expansion_limit',
carrier_attribute='h2',
sense='>=',
constant=30,
)
n.add(
'GlobalConstraint',
'vol_dc',
type='transmission_volume_expansion_limit',
carrier_attribute='dc',
sense='<=',
constant=3500,
)
n.add(
'GlobalConstraint',
'cost_dc',
type='transmission_expansion_cost_limit',
carrier_attribute='dc',
sense='>=',
constant=400,
)
n.add(
'GlobalConstraint',
'cost_dc_exact',
type='transmission_expansion_cost_limit',
carrier_attribute='dc',
sense='==',
constant=500,
)
return n
Rung 4 — ramps#
The build, the allowance and the output carried into a snapshot are each a
quantity stated with cases:. A fixed, an extendable and a committed unit
therefore share one entry. A missing limit is the full build. Rung 8 adds the rows of an
extendable committed unit, which use a large constant bound (big-M).
| PyPSA | status | note |
|---|---|---|
{c}-p-ramp_limit_up/down |
done | a lone start-up or shut-down ramp in rung 28 |
Rungs 45 and 46 extend these rows.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective8785.0, 64 rows.
The network, as PyPSA code
rung_04_ramps.py
"""Rung 4: ramps — ramp limits on fixed and extendable generators and links."""
from __future__ import annotations
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add('Bus', 'east')
n.add('Generator', 'coal_slow', bus='north', p_nom=80, marginal_cost=8, ramp_limit_up=0.2, ramp_limit_down=0.2)
n.add('Link', 'tie', bus0='north', bus1='east', p_nom=50, efficiency=1, ramp_limit_up=0.4, ramp_limit_down=0.4)
n.add('Load', 'east_load', bus='east', p_set=[5, 20, 25, 10])
n.add('Load', 'swing', bus='north', p_set=[0, 25, 45, 0])
return n
Rung 5 — global constraints#
PyPSA names every global constraint row GlobalConstraint-{name}. The type and
the comparator are data. A sense from data is a
limit of the language (#70). The file states each type as
three entries, one per sense.
| PyPSA type | status | note |
|---|---|---|
primary_energy |
split | an entry per sense, with carrier weights from data prep, one period in rung 35 and per scenario in rung 40 |
operational_limit |
split | an entry per sense, one period in rung 35, per scenario in rung 40 |
transmission_volume_expansion_limit |
split | an entry per sense, membership from PyPSA's carrier string in data prep, per scenario in rung 40 |
transmission_expansion_cost_limit |
split | an entry per sense, with the cost as the capital_cost property, without fom_cost |
tech_capacity_expansion_limit |
split | an entry per sense |
Bus-nom_min/max_{carrier} |
out | deprecated in PyPSA |
Rung 60 extends the primary_energy row. Rung 15 states the
Carrier-growth_limit row, and rung 21 extends it to every extendable
component.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective10282.833333333334, 110 rows.
The network, as PyPSA code
rung_05_global_constraints.py
"""Rung 5: global constraints — one row per limit type and sense."""
from __future__ import annotations
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add('Carrier', 'coalc', co2_emissions=0.9)
n.add('Carrier', 'gasc', co2_emissions=0.4)
n.add('Carrier', 'windc')
n.add('Generator', 'coal5', bus='north', carrier='coalc', p_nom=60, marginal_cost=9, efficiency=0.35)
n.add('Generator', 'gas5', bus='north', carrier='gasc', p_nom=60, marginal_cost=25, efficiency=0.5)
n.add('Generator', 'wind5', bus='north', carrier='windc', p_nom=60, marginal_cost=40)
n.add('Load', 'extra5', bus='north', p_set=50)
n.add('StorageUnit', 'res5', bus='north', carrier='gasc', p_nom=20, max_hours=4, state_of_charge_initial=30)
n.add('Store', 'tank5', bus='north', carrier='coalc', e_nom=40, e_initial=25)
n.add(
'GlobalConstraint',
'co2_cap',
type='primary_energy',
carrier_attribute='co2_emissions',
sense='<=',
constant=150,
)
n.add(
'GlobalConstraint',
'co2_floor',
type='primary_energy',
carrier_attribute='co2_emissions',
sense='>=',
constant=20,
)
n.add(
'GlobalConstraint',
'co2_exact',
type='primary_energy',
carrier_attribute='co2_emissions',
sense='==',
constant=120,
)
n.add('GlobalConstraint', 'op_wind', type='operational_limit', carrier_attribute='windc', sense='==', constant=30)
n.add('GlobalConstraint', 'op_coal', type='operational_limit', carrier_attribute='coalc', sense='<=', constant=200)
n.add('GlobalConstraint', 'op_gas', type='operational_limit', carrier_attribute='gasc', sense='>=', constant=10)
return n
Rung 6 — KVL#
A line's flow obeys the Kirchhoff voltage law (KVL). Around every independent cycle of the network, the flows times their series impedance sum to zero. The set of independent cycles (the cycle basis) is data, which you compute in data prep.
| PyPSA | status | note |
|---|---|---|
Line-s, Line-fix-s-* |
done | the ext and nominal rows as in rung 3 |
Kirchhoff-Voltage-Law |
done | the cycle basis from data prep |
Rungs 60 and 66 extend the Kirchhoff-Voltage-Law row.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective23962.0, 123 rows.
The network, as PyPSA code
rung_06_kvl.py
"""Rung 6: KVL — passive lines under Kirchhoff's voltage law."""
from __future__ import annotations
from math import nan
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add('Bus', 'a')
n.add('Bus', 'b')
n.add('Bus', 'c')
n.add('Generator', 'hydro', bus='a', p_nom=80, marginal_cost=10)
n.add('Generator', 'diesel6', bus='b', p_nom=80, marginal_cost=50)
n.add('Load', 'town', bus='c', p_set=45)
n.add('Line', 'ab', bus0='a', bus1='b', carrier='AC', length=30, x=0.1, r=0.01, s_nom=60)
n.add('Line', 'bc', bus0='b', bus1='c', carrier='AC', length=40, x=0.2, r=0.01, s_nom=60, s_set=[16, nan, nan, nan])
n.add('Line', 'ca', bus0='c', bus1='a', carrier='AC', length=35, x=0.1, r=0.01, s_nom=60)
n.add(
'Line',
'ca2',
bus0='c',
bus1='a',
carrier='AC',
length=50,
x=0.15,
r=0.01,
s_nom_extendable=True,
capital_cost=10,
s_nom_max=40,
s_nom_set=30,
)
n.add(
'Line',
'ca3',
bus0='c',
bus1='a',
carrier='AC',
length=80,
x=0.12,
r=0.01,
s_nom_extendable=True,
capital_cost=8,
s_nom_max=40,
)
n.add(
'GlobalConstraint',
'vol_ac',
type='transmission_volume_expansion_limit',
carrier_attribute='AC',
sense='==',
constant=2300,
)
n.add(
'GlobalConstraint',
'vol_ac_floor',
type='transmission_volume_expansion_limit',
carrier_attribute='AC',
sense='>=',
constant=1000,
)
n.add(
'GlobalConstraint',
'cost_ac',
type='transmission_expansion_cost_limit',
carrier_attribute='AC',
sense='<=',
constant=500,
)
n.add(
'GlobalConstraint',
'cost_ac_floor',
type='transmission_expansion_cost_limit',
carrier_attribute='AC',
sense='>=',
constant=100,
)
n.add(
'GlobalConstraint',
'tech_ac',
type='tech_capacity_expansion_limit',
carrier_attribute='AC',
sense='<=',
constant=60,
)
return n
Rung 7 — commitment#
| PyPSA | status | note |
|---|---|---|
{c}-status, -start_up, -shut_down |
done | Generator, with Link in rung 25 and Process in rung 26 |
{c}-com-p-lower/upper |
done | not for a modular build, which takes com-mod-p-* (PyPSA/PyPSA#1901) |
{c}-*-p-fixed-upper |
done | status, start and stop each at most one, as explicit rows |
{c}-com-transition-start-up/shut-down |
done | the carried-in state stated with cases:, no entry for the first snapshot, and zero carried in for a unit built in a later period (rung 64) |
{c}-com-up-time, -down-time |
done | sum_back(window=min_up_time) from the first snapshot. The up or down time brought in counts as a start or a stop before it, from data prep (rung 24). A modular unit's down time counts modules |
stand_by_cost, start_up_cost, shut_down_cost |
done | the period weight and no snapshot weight on a start or a stop (rung 48), and a cost that may change per snapshot (rung 62) |
{c}-com-p-before/-current/-partly-* |
done | under linearized_unit_commitment, in a file laid over this one (a patch): rungs 12, 44, 47 and 74 for Generator, rung 67 for Link and Process |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective7775.0, 118 rows.
The network, as PyPSA code
rung_07_commitment.py
"""Rung 7: commitment — committable units with up and down times and ramp limits at the transitions."""
from __future__ import annotations
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add(
'Generator',
'uc',
bus='north',
committable=True,
p_nom=50,
marginal_cost=5,
p_min_pu=0.4,
min_up_time=3,
min_down_time=2,
up_time_before=1,
ramp_limit_up=0.5,
ramp_limit_down=0.5,
ramp_limit_start_up=0.6,
ramp_limit_shut_down=0.6,
start_up_cost=100,
shut_down_cost=50,
stand_by_cost=5,
)
n.add(
'Generator',
'cold',
bus='south',
committable=True,
p_nom=30,
marginal_cost=60,
p_min_pu=0.3,
min_up_time=2,
min_down_time=1,
up_time_before=0,
ramp_limit_up=0.5,
ramp_limit_down=0.5,
start_up_cost=80,
)
n.add('Load', 'swing7', bus='north', p_set=[25, 45, 45, 10])
return n
Rung 8 — modular and big-M#
| PyPSA | status | note |
|---|---|---|
{c}-n_mod, {c}-p_nom_modularity |
done | Generator, Link and Process, and none for a build that is active in no snapshot (rung 63) |
{c}-*-p_nom-variable-upper |
done | a modular unit is on only where a module is built |
{c}-*-p-fixed-upper, modular |
done | the cap as the whole count of modules, p_nom / p_nom_mod from data prep, also for a fixed build such as array (#123), a fractional count in Refusals |
{c}-com-mod-p-lower/upper |
done | one module's share times the status, for a fixed build too, in place of the com-p-* rows (PyPSA/PyPSA#1901) |
{c}-com-ext-p-* (big-M) |
done | a cap row beside a big-M row, with M from data prep, PyPSA's rule for an infinite build cap included (rung 64) |
{c}-com-ext-p-lower-nonneg |
done | (p_min_pu >= 0).all() from data prep |
{c}-p-ramp_limit_*-bigM |
done | run and start rows up, run and shut rows down, each over the carried-in output stated with cases:, and the ordinary rows for a modular build (rung 27) |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective15915.0, 183 rows.
The network, as PyPSA code
rung_08_modular_big_m.py
"""Rung 8: modular and big-M — capacity in whole modules, built or already standing, and a committable unit whose capacity is also built."""
from __future__ import annotations
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add('Bus', 'mill')
n.add(
'Generator',
'block',
bus='mill',
p_nom_extendable=True,
committable=True,
p_nom_mod=25,
p_nom_max=100,
capital_cost=30,
marginal_cost=20,
p_min_pu=0.2,
up_time_before=0,
)
n.add(
'Generator',
'flex',
bus='mill',
p_nom_extendable=True,
committable=True,
p_nom_max=80,
capital_cost=50,
marginal_cost=10,
p_min_pu=0.3,
up_time_before=0,
ramp_limit_up=0.25,
ramp_limit_down=0.25,
)
n.add(
'Generator',
'sink',
bus='mill',
p_nom_extendable=True,
committable=True,
p_nom_max=30,
capital_cost=40,
marginal_cost=15,
p_min_pu=-0.2,
up_time_before=0,
)
n.add(
'Generator',
'array',
bus='mill',
committable=True,
p_nom=90,
p_nom_mod=30,
marginal_cost=12,
p_min_pu=0.2,
up_time_before=0,
)
n.add('Load', 'mill_load', bus='mill', p_set=[40, 80, 120, 60])
return n
Rung 9 — multi-link#
| PyPSA | status | note |
|---|---|---|
| nodal balance, ports 1..n | done | one term over link_output, so a link of any number of output ports needs no further entry (#124) |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective11714.4, 92 rows.
The network, as PyPSA code
rung_09_multilink.py
"""Rung 9: a multi-link with four output ports — power and heat sold, waste heat vented, and a station service the link draws back."""
from __future__ import annotations
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add('Bus', 'gasb')
n.add('Bus', 'power')
n.add('Bus', 'heat')
n.add('Bus', 'flue')
n.add('Bus', 'aux')
n.add('Generator', 'well', bus='gasb', p_nom=100, marginal_cost=5)
n.add('Generator', 'grid_import', bus='power', p_nom=50, marginal_cost=60)
n.add('Generator', 'vent', bus='flue', p_nom=100, p_min_pu=-1, p_max_pu=0)
n.add('Generator', 'aux_supply', bus='aux', p_nom=10, marginal_cost=2)
n.add(
'Link',
'chp',
bus0='gasb',
bus1='power',
bus2='heat',
bus3='flue',
bus4='aux',
efficiency=0.4,
efficiency2=0.45,
efficiency3=0.1,
efficiency4=-0.02,
p_nom=60,
marginal_cost=1,
)
n.add('Load', 'homes', bus='power', p_set=20)
n.add('Load', 'district', bus='heat', p_set=18)
return n
Rung 10 — quadratic costs#
The PyPSA marginal_cost_quadratic is a marginal cost that is quadratic in
output. It adds one squared cost per component to the objective, and weights
each snapshot by the hours the snapshot weighs. Here a generator and a link
have it, and rung 36 puts it on a process, a storage unit and a store. On a
plain run, it is zero, so the squared cost is zero and the objective stays
linear.
| PyPSA | status | note |
|---|---|---|
marginal_cost_quadratic |
done | degree 2 in the objective, on Generator and Link |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective12587.437500000098, 60 rows.
The network, as PyPSA code
rung_10_quadratic_costs.py
"""Rung 10: quadratic costs — a marginal cost quadratic in output."""
from __future__ import annotations
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add('Bus', 'village')
n.add('Generator', 'steam', bus='north', p_nom=80, marginal_cost=5, marginal_cost_quadratic=0.08)
n.add('Generator', 'engine', bus='north', p_nom=80, marginal_cost=20, marginal_cost_quadratic=0.01)
n.add(
'Link',
'wire2',
bus0='north',
bus1='village',
p_nom=40,
p_min_pu=-1,
efficiency=0.9,
marginal_cost=1,
marginal_cost_quadratic=0.02,
)
n.add('Load', 'village_load', bus='village', p_set=15)
n.add('Load', 'extra10', bus='north', p_set=[30, 50, 40, 60])
return n
Rung 11 — ac-dc-meshed#
This rung is the whole ac_dc_meshed example of PyPSA. It has meshed
alternating-current (AC) and direct-current (DC) networks, extendable lines,
links and generators, carriers and a CO2 budget. It has the rows of rungs 1, 3,
5 and 6 together, and it is the first rung with an objective constant.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective-3474256.0405499237, 468 rows.
The network, as PyPSA code
rung_11_ac_dc_meshed.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 11: PyPSA's own `ac_dc_meshed` example, whole — meshed AC and DC, extendable lines, links and generators, carriers, a CO2 budget."""
from __future__ import annotations
from datetime import datetime
#: Ten hourly stamps, the example's own. Every weighting column there is 1.0, which is
#: also the default, so no row below sets one.
SNAPSHOTS = [datetime(2015, 1, 1, hour) for hour in range(10)]
#: Wind availability per snapshot, for the three generators that carry a profile.
P_MAX_PU = {
'Manchester Wind': [0.930019875, 0.4857475804, 0.2336917351, 0.2576042221, 0.6269055694, 0.6035984088, 0.6789075462, 0.3613026112, 0.6216040549, 0.5215183715],
'Norway Wind': [0.9745832033, 0.4812903778, 0.4072258018, 0.5999649628, 0.524468219, 0.0096927054, 0.2204533621, 0.8239185004, 0.5562297265, 0.4394160378],
'Frankfurt Wind': [0.5590784039, 0.7529103711, 0.1234650887, 0.9666766524, 0.8590078044, 0.5261537924, 0.077893008, 0.0590234716, 0.2485544952, 0.1080601728],
} # fmt: skip
#: Demand per snapshot, for each of the six loads.
P_SET = {
'London': [35.7962441027, 976.8245614698, 250.5873120464, 130.7531445827, 151.1001686, 931.857051942, 289.8482871447, 864.3433217147, 689.5772637703, 627.8789859434],
'Frankfurt': [398.0478469638, 432.4361062425, 379.8039282662, 868.3617642835, 548.7707546221, 828.6652426012, 449.2907519075, 699.1637663734, 915.8667802518, 414.8876464034],
'Norway': [820.035835936, 854.8340468618, 42.550744351, 647.5482327851, 884.0738733306, 509.0624485516, 595.6079648147, 291.6424496984, 2.1534925491, 760.7401765038],
'Norwich': [415.4625642653, 262.6061464526, 418.4763531902, 552.9595393098, 218.159858091, 791.9762655836, 531.8706808219, 23.5134667186, 970.0590684572, 0.9248336907],
'Bremen': [640.0863775411, 703.554333706, 440.8361303183, 612.5763056818, 803.4367808051, 605.4006873582, 641.0905902397, 408.0085411725, 912.2477761646, 898.0530916423],
'Manchester': [857.5514402011, 750.5996237166, 156.5648760141, 527.8708221189, 83.8977589634, 676.6233193474, 731.1371004827, 553.3448891847, 298.338082262, 768.2905859888],
} # fmt: skip
def build():
"""The example network, stated as the calls that build it.
A rung states its data inline, so that the PyPSA model under review is the
script — ``reference.py`` says so and ``test_pypsa_references.py`` checks
it. The numbers here are PyPSA's own ``ac_dc_meshed``, which is where this
rung's published objective comes from; ``reference.py`` pins the version
they were read at.
"""
import pypsa
n = pypsa.Network()
n.set_snapshots(SNAPSHOTS)
# Bus
n.add('Bus', 'London', v_nom=380.0, x=-0.13, y=51.5)
n.add('Bus', 'Norwich', v_nom=380.0, x=1.3, y=52.6)
n.add('Bus', 'Norwich DC', v_nom=200.0, x=1.3, y=52.5, carrier='DC')
n.add('Bus', 'Manchester', v_nom=380.0, x=-2.2, y=53.47)
n.add('Bus', 'Bremen', v_nom=380.0, x=8.8, y=53.08)
n.add('Bus', 'Bremen DC', v_nom=200.0, x=8.8, y=52.98, carrier='DC')
n.add('Bus', 'Frankfurt', v_nom=380.0, x=8.7, y=50.12)
n.add('Bus', 'Norway', v_nom=380.0, x=10.75, y=60.0)
n.add('Bus', 'Norway DC', v_nom=200.0, x=10.75, y=60.0, carrier='DC')
# Carrier
n.add('Carrier', 'gas', co2_emissions=0.24, color='red')
n.add('Carrier', 'wind', color='blue')
n.add('Carrier', 'battery', color='green')
n.add('Carrier', 'load', color='black')
n.add('Carrier', 'AC', color='orange')
n.add('Carrier', 'DC', color='purple')
# Generator
n.add(
'Generator',
'Manchester Wind',
bus='Manchester',
p_nom=80.0,
p_nom_extendable=True,
p_nom_min=100.0,
p_max_pu=P_MAX_PU['Manchester Wind'],
carrier='wind',
marginal_cost=0.11,
capital_cost=2793.6516029328,
)
n.add(
'Generator',
'Manchester Gas',
bus='Manchester',
p_nom=50000.0,
p_nom_extendable=True,
carrier='gas',
marginal_cost=4.5323676307,
capital_cost=196.6151679691,
efficiency=0.3500264336,
)
n.add(
'Generator',
'Norway Wind',
bus='Norway',
p_nom=100.0,
p_nom_extendable=True,
p_nom_min=100.0,
p_max_pu=P_MAX_PU['Norway Wind'],
carrier='wind',
marginal_cost=0.09,
capital_cost=2184.3747960912,
)
n.add(
'Generator',
'Norway Gas',
bus='Norway',
p_nom=20000.0,
p_nom_extendable=True,
carrier='gas',
marginal_cost=5.8928445406,
capital_cost=158.2512497168,
efficiency=0.3568363832,
)
n.add(
'Generator',
'Frankfurt Wind',
bus='Frankfurt',
p_nom=110.0,
p_nom_extendable=True,
p_nom_min=100.0,
p_max_pu=P_MAX_PU['Frankfurt Wind'],
carrier='wind',
marginal_cost=0.1,
capital_cost=2129.4561224763,
)
n.add(
'Generator',
'Frankfurt Gas',
bus='Frankfurt',
p_nom=80000.0,
p_nom_extendable=True,
carrier='gas',
marginal_cost=4.0863219899,
capital_cost=102.6769530076,
efficiency=0.3516658529,
)
# Line
n.add(
'Line',
'0',
bus0='London',
bus1='Manchester',
x=0.7968782824,
s_nom=40000.0,
s_nom_extendable=True,
capital_cost=0.1367157553,
carrier='AC',
)
n.add(
'Line',
'1',
bus0='Manchester',
bus1='Norwich',
x=0.3915599178,
s_nom=40000.0,
s_nom_extendable=True,
capital_cost=0.1334916779,
carrier='AC',
)
n.add(
'Line',
'2',
bus0='Bremen DC',
bus1='Norwich DC',
r=0.2126041927,
s_nom=40000.0,
s_nom_extendable=True,
capital_cost=0.0086734246,
carrier='AC',
)
n.add(
'Line',
'3',
bus0='Norwich DC',
bus1='Norway DC',
r=0.4861637504,
s_nom=40000.0,
s_nom_extendable=True,
capital_cost=0.1291260515,
carrier='AC',
)
n.add(
'Line',
'4',
bus0='Norway DC',
bus1='Bremen DC',
r=0.4287266497,
s_nom=40000.0,
s_nom_extendable=True,
capital_cost=0.0624298729,
carrier='AC',
)
n.add(
'Line',
'5',
bus0='Norwich',
bus1='London',
x=0.2388003463,
s_nom=40000.0,
s_nom_extendable=True,
capital_cost=0.0218524519,
carrier='AC',
)
n.add(
'Line',
'6',
bus0='Bremen',
bus1='Frankfurt',
x=0.4,
s_nom=40000.0,
s_nom_extendable=True,
capital_cost=0.2,
carrier='AC',
)
# Link
n.add(
'Link',
'Norwich Converter',
bus0='Norwich',
bus1='Norwich DC',
carrier='DC',
p_nom=1000.0,
p_nom_extendable=True,
p_min_pu=-0.9,
p_max_pu=0.9,
capital_cost=0.21,
)
n.add(
'Link',
'Norway Converter',
bus0='Norway',
bus1='Norway DC',
carrier='DC',
p_nom=1000.0,
p_nom_extendable=True,
p_min_pu=-0.9,
p_max_pu=0.9,
capital_cost=0.2,
)
n.add(
'Link',
'Bremen Converter',
bus0='Bremen',
bus1='Bremen DC',
carrier='DC',
p_nom=1000.0,
p_nom_extendable=True,
p_min_pu=-0.9,
p_max_pu=0.9,
capital_cost=0.19,
)
n.add(
'Link',
'DC link',
bus0='London',
bus1='Bremen',
carrier='DC',
p_nom=1000.0,
p_nom_extendable=True,
p_min_pu=-0.9,
p_max_pu=0.9,
capital_cost=0.8765342,
)
# Load
n.add('Load', 'London', bus='London', carrier='load', p_set=P_SET['London'])
n.add('Load', 'Frankfurt', bus='Frankfurt', carrier='load', p_set=P_SET['Frankfurt'])
n.add('Load', 'Norway', bus='Norway', carrier='load', p_set=P_SET['Norway'])
n.add('Load', 'Norwich', bus='Norwich', carrier='load', p_set=P_SET['Norwich'])
n.add('Load', 'Bremen', bus='Bremen', carrier='load', p_set=P_SET['Bremen'])
n.add('Load', 'Manchester', bus='Manchester', carrier='load', p_set=P_SET['Manchester'])
# GlobalConstraint
n.add('GlobalConstraint', 'co2_limit', sense='<=', constant=1000.0)
return n
Rung 13 — transmission losses#
This rung runs n.optimize(transmission_losses=...), under which a line
dissipates a loss that its flow causes. The loss stays above a fan of cuts to
the quadratic loss curve r_pu_eff * p**2. Half of it is charged at either end
of the line.
PyPSA has two modes (constraints.py:2722). Both build the same rows
loss + slope * flow >= offset and loss - slope * flow >= offset, one pair
per cut:
{'mode': 'tangents', 'segments': K}takesKtangents atp_k = k / Kof the rating, with slope2 r p_k.True, or{'mode': 'secants', 'atol': 1, 'rtol': 0.1, 'max_segments': 20}, takes the secants between consecutive breakpointsp_k, p_k+1. Each has sloper (p_k + p_k+1)and offset-r p_k p_k+1. The breakpoints start atp_0 = 0and step bymax(k / (k - 1), 1 + 2 (rtol + sqrt(rtol + rtol**2)))until they cover the rating.
The mode sets only the values of Line_loss_slope and Line_loss_offset over
the segment axis. You compute them, and the breakpoint loop, in data prep.
The loss variable, its cap and the cut rows exist only where
transmission_losses is on. On a plain run, the flag is off and the data has
no segments. The loss is then absent and zero in the balance, and the rows
collapse to the lossless ones.
| PyPSA | status | note |
|---|---|---|
Line-loss, Transformer-loss |
done | absent, and zero in the balance, where lossless |
Line-fix-s-*, Line-ext-s-*, Transformer-fix-s-*, Transformer-ext-s-* |
done | the loss counted against the rating |
Bus-nodal_balance |
done | half of each incident line's and transformer's loss at either end |
Line-loss_upper, Transformer-loss_upper |
done | loss_max from data prep, and its refusals in Refusals |
Line-loss_tangents-{k}-1, Transformer-loss_tangents-{k}-1 |
split | a row per segment in PyPSA, one entry over segment in the file |
Line-loss_tangents-{k}--1, Transformer-loss_tangents-{k}--1 |
split | |
Line-loss_secants-pos, Line-loss_secants-neg, Transformer-loss_secants-pos, Transformer-loss_secants-neg |
done | the same two entries in the secant mode, with slope, offset and the breakpoint loop from data prep (rungs 19 and 23) |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective10645.295879552297, 150 rows.
The network, as PyPSA code
rung_13_losses.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 13: transmission losses in tangent form — a loss per line."""
from __future__ import annotations
import spine
OPTIMIZE = {'transmission_losses': {'mode': 'tangents', 'segments': 2}}
def build():
"""The spine plus a 110 kV triangle of lines, one of them extendable — ohms a real line has, so the loss stays a few percent of the flow."""
n = spine.build()
n.add('Bus', ['a', 'b', 'c'], v_nom=110)
n.add('Generator', 'hydro13', bus='a', p_nom=80, marginal_cost=10)
n.add('Generator', 'diesel13', bus='b', p_nom=80, marginal_cost=50)
n.add('Line', 'ab13', bus0='a', bus1='b', carrier='AC', x=30, r=6, s_nom=60)
n.add('Line', 'bc13', bus0='b', bus1='c', carrier='AC', x=60, r=9.7, s_nom=60)
n.add(
'Line',
'ca13',
bus0='c',
bus1='a',
carrier='AC',
x=45,
r=6,
s_nom=40,
s_nom_extendable=True,
s_nom_max=90,
capital_cost=4,
)
n.add('Load', 'town13', bus='c', p_set=[35, 55, 15, 45])
return n
Rung 19 — transmission losses in secant form#
Rung 19 is rung 13's network in the secant mode. It has the same loss rows, but PyPSA's tolerance loop places the cuts, not one cut per segment.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective10840.926895402912, 150 rows.
The network, as PyPSA code
rung_19_losses_secants.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 19: transmission losses in secant form — the same loss per line, its cuts placed by PyPSA's tolerance loop."""
from __future__ import annotations
import spine
OPTIMIZE = {'transmission_losses': {'mode': 'secants', 'atol': 1, 'rtol': 0.1, 'max_segments': 20}}
def build():
"""Rung 13's 110 kV triangle, unchanged, so the two modes differ only in the cuts."""
n = spine.build()
n.add('Bus', ['a', 'b', 'c'], v_nom=110)
n.add('Generator', 'hydro19', bus='a', p_nom=80, marginal_cost=10)
n.add('Generator', 'diesel19', bus='b', p_nom=80, marginal_cost=50)
n.add('Line', 'ab19', bus0='a', bus1='b', carrier='AC', x=30, r=6, s_nom=60)
n.add('Line', 'bc19', bus0='b', bus1='c', carrier='AC', x=60, r=9.7, s_nom=60)
n.add(
'Line',
'ca19',
bus0='c',
bus1='a',
carrier='AC',
x=45,
r=6,
s_nom=40,
s_nom_extendable=True,
s_nom_max=90,
capital_cost=4,
)
n.add('Load', 'town19', bus='c', p_set=[35, 55, 15, 45])
return n
Rung 14 — two-stage stochastic#
This rung has two futures and a risk preference, set by n.set_scenarios(...)
and n.set_risk_preference(alpha, omega). Everything over a snapshot spans a
scenario as well. Capacity does not, because it is chosen once before the
future is known. The operating cost is the expectation over the scenarios'
weights.
A risk preference adds the conditional value at risk (CVaR) rows: an excess per
scenario and the tail's average, blended into the objective at omega. PyPSA
(optimize.py:497) builds the CVaR columns and rows under any risk preference,
and none without one. The file states them only where CVaR_omega has a
value, and CVaR_omega has a value only under a risk preference. Rung 70 sets
omega = 0. On a plain run, there is no risk preference and no CVaR row.
A parameter spans scenario exactly when PyPSA reads it per scenario. PyPSA
reads component data through c.da, one value per scenario
(components/array.py:347-406), so almost every parameter spans one. This
rung's wind p_max_pu differs by scenario, and the file states it over
scenario.
Some parameters span no scenario. PyPSA (constants.py:26-58,
consistency.py:1250) refuses a difference by scenario in the attributes that
fix the network's shape, such as bus, carrier, lifetime, active,
committable or p_nom_extendable. It refuses one in *_nom_set, delay and
cyclic_delay too. These parameters, and the values you derive from them in
data prep, span no scenario. PyPSA (components.py:1000-1003,
constraints.py:477-482, global_constraints.py:239-243) also reduces three
quantities over the scenarios, so they span none:
maintainable, to a union over the scenarios(p_min_pu >= 0).all(), over every scenario- a carrier's growth limits, to their least value
PyPSA builds the cycle basis (networks.py:1428-1435) and the outage factors
(abstract.py:652) from the first scenario. It refuses a v_nom, x_pu_eff
or r_pu_eff that differs by scenario (constants.py:33-35,
PyPSA/PyPSA#2104), so the first scenario's impedances are those of every
scenario. The cycle weights and the branch outage distribution factors (BODF)
of a branch therefore span no scenario.
In the file, a transformer's phase shift spans no scenario (#783). PyPSA
(variables.py:446-456, constraints.py:1744) reads it per scenario.
| PyPSA | status | note |
|---|---|---|
Generator-p, Link-p |
done | over scenario, except Generator-p_nom, which is chosen once |
Generator-fix-p-*, -ext-p-*, Link-fix-p-*, Bus-nodal_balance |
done | rungs 1 and 3, over scenario |
CVaR-a, CVaR-theta, CVaR |
done | |
CVaR-excess-{s} |
split | a row per scenario in PyPSA, one entry over scenario in the file, none without a risk preference |
CVaR-def |
done | 1 / (1 - alpha) from data prep, none without a risk preference |
| objective | done | capacity once, at its capital cost in expectation over the scenarios, and operation at (1 - omega) in expectation and omega at the tail |
Generator-fix-p-* and the other rows over component data |
done | data per scenario, such as p_max_pu, with operating data in rung 41, first-stage bounds and capital cost in rung 42 |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective9267.386666666665, 87 rows.
The network, as PyPSA code
rung_14_stochastic.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 14: two futures and a risk preference — capacity chosen once, dispatch per scenario."""
from __future__ import annotations
import spine
def build():
"""The spine plus an extendable wind unit whose availability and the south's load differ between a calm and a stormy future."""
n = spine.build()
n.add('Generator', 'wind14', bus='south', p_nom_extendable=True, p_nom_max=100, marginal_cost=1, capital_cost=20)
n.add('Load', 'port14', bus='south')
n.set_scenarios({'calm': 0.6, 'stormy': 0.4})
n.c.loads.dynamic.p_set[('calm', 'port14')] = [10, 20, 15, 10]
n.c.loads.dynamic.p_set[('stormy', 'port14')] = [40, 60, 50, 30]
n.c.generators.dynamic.p_max_pu[('calm', 'wind14')] = [0.9, 0.7, 0.8, 0.6]
n.c.generators.dynamic.p_max_pu[('stormy', 'wind14')] = [0.3, 0.2, 0.4, 0.1]
n.set_risk_preference(alpha=0.5, omega=0.3)
return n
Rung 15 — investment periods#
This rung runs n.optimize(multi_investment_periods=True), where each snapshot
belongs to an investment period. An asset is active in the periods that its
build year and lifetime span. Its capacity is paid once per period that it is
active in. Each period has a weight, and a carrier may grow only so much per
period. The snapshots an asset is active in are the parameter {c}_active,
from data prep.
| PyPSA | status | note |
|---|---|---|
Generator-p |
done | where the generator is active in the snapshot's period, by active from data prep |
Generator-fix-p-*, -ext-p-*, -ext-p_nom-* |
done | rungs 1 and 3, where active is true |
Carrier-growth_limit |
done | every extendable component of the carrier, counted in the first period a build is active in, with edge=0 at the first period |
| objective | done | period weight on operation, capacity once per period it is active in |
Rungs 29, 32, 35, 38, 39, 49 and 66 extend these rows.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective12747.19109626398, 80 rows.
The network, as PyPSA code
rung_15_multi_period.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 15: two investment periods — build years, lifetimes, period weights and a carrier's growth limit."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def build():
"""A whole network, not the spine: eight snapshots over two periods, a unit that retires, two wind builds capped by growth."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(4)] + [(2030, datetime(2030, 1, 1, t)) for t in range(4)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [10.0, 10.0]
n.snapshot_weightings['objective'] = [2.0, 1.5, 2.5, 2.0, 2.0, 1.5, 2.5, 2.0]
n.add('Bus', ['north', 'south'])
n.add('Carrier', 'wind', max_growth=50, max_relative_growth=0.5)
n.add('Carrier', 'gas')
n.add('Generator', 'old_gas', bus='north', carrier='gas', p_nom=40, marginal_cost=30, build_year=2010, lifetime=15)
n.add(
'Generator',
'wind20',
bus='north',
carrier='wind',
p_nom_extendable=True,
p_nom_max=200,
marginal_cost=1,
capital_cost=100,
build_year=2020,
lifetime=30,
p_max_pu=[0.8, 0.6, 0.7, 0.5, 0.8, 0.6, 0.7, 0.5],
)
n.add(
'Generator',
'wind30',
bus='south',
carrier='wind',
p_nom_extendable=True,
p_nom_max=200,
marginal_cost=1,
capital_cost=80,
build_year=2030,
lifetime=30,
p_max_pu=[0.9, 0.7, 0.6, 0.8, 0.9, 0.7, 0.6, 0.8],
)
n.add(
'Generator',
'gas30',
bus='south',
carrier='gas',
p_nom_extendable=True,
p_nom_max=200,
marginal_cost=40,
capital_cost=50,
build_year=2030,
lifetime=30,
)
n.add('Link', 'wire15', bus0='north', bus1='south', p_nom=60, p_min_pu=-1, efficiency=0.95)
n.add('Load', 'town15', bus='north', p_set=[20, 30, 25, 20, 35, 45, 40, 30])
n.add('Load', 'port15', bus='south', p_set=[10, 20, 15, 10, 30, 40, 35, 25])
return n
Rung 16 — link delay#
A source feeds two sinks over links whose energy arrives late. With the PyPSA
delay, a port delivers a number of snapshots late. cyclic_delay sets what
happens to flow still in transit at the end of the horizon: it wraps to the
start, or it is lost. The delay is a number per link and the wrap a kind per
link. The balance states them with cases: over a
shift by
offset=Link_output_delay. One case wraps (edge='wrap'),
and the other vacates (edge=0).
This is the only rung with a uniform generators weighting, and PyPSA measures
delay in units of that weighting. With a uniform weighting, a delay of n is
a shift of exactly n snapshot positions, which a positional shift gives.
With a weighting that is not uniform, PyPSA resamples by elapsed time, not by
position. A shift that varies along the snapshot axis is a
limit of the language (#299). The file states an
assumption,
Link_output_delay_under_uniform_weighting,
and a tool refuses such a delay when data is attached.
| PyPSA | status | note |
|---|---|---|
link delay, cyclic_delay |
done | a cases: on cyclic_delay over shift(offset=delay), at uniform generators weighting |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective5262.5, 52 rows.
The network, as PyPSA code
rung_16_link_delay.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 16: link delay — a source feeding two sinks over links whose energy arrives late, one wrapping cyclically and one losing what is still in transit at the horizon's edge."""
from __future__ import annotations
from datetime import datetime
#: Four hourly stamps. The `generators` weighting is uniform here, and only here
#: on the ladder, because PyPSA measures `delay` in those units: a uniform column
#: makes a delay of `n` a shift of exactly `n` snapshot positions, which is what a
#: positional `shift(offset=n)` reproduces. The `objective` and `stores` columns
#: stay non-uniform, so no cost or storage factor passes as identity.
SNAPSHOTS = [datetime(2015, 1, 1, hour) for hour in range(4)]
WEIGHTINGS = {'objective': [2.0, 1.5, 2.5, 3.0], 'stores': [0.5, 2.0, 1.5, 2.5], 'generators': [1.0, 1.0, 1.0, 1.0]}
#: Each sink carries the same demand, so the only thing that separates their cost
#: is how each link treats the horizon's edge.
DEMAND = [20.0, 15.0, 25.0, 10.0]
def build():
"""A source, two delayed links, and two sinks, stated as the calls that build it.
``pipe_wrap`` delays by two snapshots and wraps cyclically, so every unit the
cheap source sends reaches its sink and the expensive backup stays dark.
``pipe_lose`` delays by one and does not wrap, so the flow that would arrive
in the first snapshot is lost and that snapshot's demand falls to the backup.
The two links differ in both a per-link number (`delay`) and a per-link kind
(`cyclic_delay`), which is what the spec's ``cases:`` turns on.
"""
import pypsa
n = pypsa.Network()
n.set_snapshots(SNAPSHOTS)
for column, values in WEIGHTINGS.items():
n.snapshot_weightings[column] = values
n.add('Bus', 'source')
n.add('Bus', 'sink_wrap')
n.add('Bus', 'sink_lose')
n.add('Generator', 'spring', bus='source', p_nom=200, marginal_cost=5)
n.add('Generator', 'backup_wrap', bus='sink_wrap', p_nom=200, marginal_cost=100)
n.add('Generator', 'backup_lose', bus='sink_lose', p_nom=200, marginal_cost=100)
n.add('Link', 'pipe_wrap', bus0='source', bus1='sink_wrap', p_nom=100, delay=2, cyclic_delay=True)
n.add('Link', 'pipe_lose', bus0='source', bus1='sink_lose', p_nom=100, delay=1, cyclic_delay=False)
n.add('Load', 'load_wrap', bus='sink_wrap', p_set=DEMAND)
n.add('Load', 'load_lose', bus='sink_lose', p_set=DEMAND)
return n
Rung 17 — process#
A process is a generalised converter that moves an internal power from bus0
to the buses it feeds. Each port draws or delivers at its own rate. A
non-extendable process has a fixed capacity, and an extendable one chooses its
capacity between p_nom_min and p_nom_max. A ramp limit caps the change in
internal power between snapshots. A p_set fixes an internal power schedule,
and a p_nom_set fixes the built capacity of an extendable process. The rows
are those of the generator and the link, applied to a converter.
| PyPSA | status | note |
|---|---|---|
Process-p, Process-p_nom |
done | internal power and capacity, as a link |
Process-fix-p-*, -ext-p-*, -ext-p_nom-* |
done | rungs 1 and 3, over a converter |
Process-p-ramp_limit_* |
done | rung 4, on a non-committable converter, and committed in rung 26 |
Process-p_set |
done | a fixed internal power schedule |
Process-p_nom_set |
done | a fixed built capacity |
Bus-nodal_balance |
done | each port enters at its rate |
| objective | done | marginal cost on internal power, and capital cost on capacity |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective9730.0, 70 rows.
The network, as PyPSA code
rung_17_process.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 17: process — generalized converters, one fixed and ramping, one extendable, one on a set schedule, all feeding a hub."""
from __future__ import annotations
from math import nan
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add('Bus', 'hub')
n.add(
'Process',
'conv_fix',
bus0='north',
bus1='hub',
p_nom=50,
marginal_cost=2,
ramp_limit_up=0.3,
ramp_limit_down=0.3,
)
n.add(
'Process',
'conv_ext',
bus0='south',
bus1='hub',
p_nom_extendable=True,
capital_cost=20,
p_nom_min=5,
p_nom_max=40,
marginal_cost=1,
p_nom_set=25,
)
n.add('Process', 'conv_set', bus0='north', bus1='hub', p_nom=20, marginal_cost=3, p_set=[10, nan, nan, nan])
n.add('Load', 'hub_load', bus='hub', p_set=[15, 20, 25, 10])
return n
Rung 18 — transformer#
A transformer is a passive branch between two buses, as a line is. Its flow
follows its effective series reactance and a phase shift, which is fixed or
optimised. It obeys KVL around every independent cycle, and a transformer in no
cycle has no KVL row. A non-extendable transformer has a fixed nominal apparent
power, and an extendable one chooses it between s_nom_min and s_nom_max. An
s_set fixes a flow schedule, and an s_nom_set fixes the built capacity of
an extendable transformer.
| PyPSA | status | note |
|---|---|---|
Transformer-s, Transformer-s_nom |
done | flow and capacity, as a line |
Transformer-fix-s-*, -ext-s-*, -ext-s_nom-* |
done | rungs 1 and 3, over a transformer |
Transformer-s_set |
done | a fixed flow schedule |
Transformer-s_nom_set |
done | a fixed built capacity |
Kirchhoff-Voltage-Law |
done | rung 6, over x_pu_eff and a phase shift |
Transformer-phase_shift |
done | rung 20, an optimised phase shift |
| objective | done | capital on capacity |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective12274.401472395122, 106 rows.
The network, as PyPSA code
rung_18_transformer.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 18: transformer — passive branches under KVL in a meshed triangle, one fixed and on a set flow, two extendable in parallel."""
from __future__ import annotations
from math import nan
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add('Bus', 'a')
n.add('Bus', 'b')
n.add('Bus', 'c')
n.add('Generator', 'hydro18', bus='a', p_nom=80, marginal_cost=10)
n.add('Generator', 'diesel18', bus='b', p_nom=80, marginal_cost=50)
n.add('Load', 'town18', bus='c', p_set=45)
n.add('Line', 'ab', bus0='a', bus1='b', carrier='AC', length=30, x=0.1, r=0.01, s_nom=60)
n.add(
'Transformer',
'bc',
bus0='b',
bus1='c',
x=0.1,
r=0.01,
s_nom=60,
phase_shift=10,
s_set=[16, nan, nan, nan],
)
n.add(
'Transformer',
'ca',
bus0='c',
bus1='a',
x=0.15,
r=0.01,
s_nom=60,
s_nom_extendable=True,
capital_cost=10,
s_nom_min=5,
s_nom_max=40,
s_nom_set=30,
tap_ratio=1.05,
)
n.add(
'Transformer',
'ca2',
bus0='c',
bus1='a',
x=0.12,
r=0.01,
s_nom=60,
s_nom_extendable=True,
capital_cost=8,
s_nom_min=5,
s_nom_max=40,
)
return n
Rung 20 — phase shifter#
Where the phase_shift_min of a phase-shifting transformer is below its
phase_shift_max, its voltage angle shift is a decision per snapshot. The shift
lies between the two bounds, in degrees. It enters the same KVL cycle sum as a
fixed one, so it redistributes the flows around a cycle without moving active
power. In this rung, with the shift, the transformer runs at its rating while
the upstream unit serves the whole load as the load varies. The decided shift replaces
the fixed phase_shift.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective16455.0, 88 rows.
The network, as PyPSA code
rung_20_phase_shifter.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 20: phase shifter — a transformer whose per-snapshot phase shift is optimised, holding it at its rating and rerouting the surplus around the cycle.
The parallel lines carry low reactance, so a few degrees of shift move tens of
megawatts: the phase-shifting transformer keeps its flow at its ``s_nom`` while
the upstream hydro serves the whole varying load, and the costly local unit
stays off. A fixed shift could not follow the load, so the shift is a decision.
"""
from __future__ import annotations
import spine
def build():
"""The spine plus a triangle where a phase-shifting transformer reroutes cheap power around a binding leg, as a ``pypsa.Network``."""
n = spine.build()
n.add('Bus', ['a', 'b', 'c'])
n.add('Generator', 'hydro20', bus='a', p_nom=300, marginal_cost=10)
n.add('Generator', 'diesel20', bus='c', p_nom=300, marginal_cost=200)
n.add('Load', 'town20', bus='c', p_set=[90, 75, 120, 105])
n.add('Line', 'ab20', bus0='a', bus1='b', carrier='AC', x=0.002, r=0.0002, s_nom=120)
n.add('Line', 'bc20', bus0='b', bus1='c', carrier='AC', x=0.002, r=0.0002, s_nom=120)
n.add(
'Transformer',
'ca20',
bus0='c',
bus1='a',
x=0.002,
r=0.0002,
s_nom=40,
phase_shift_min=-30,
phase_shift_max=30,
)
return n
Rung 21 — carrier growth#
The max_growth of a carrier caps what the carrier adds in an investment
period. The cap covers every extendable component of that carrier, not only
the generators. Here a battery carrier caps a storage unit and a store built in
the first period, and the two builds fill the cap together. A store built in the
later period adds its allowance plus half of what the carrier added before.
PyPSA counts only the components that have a carrier attribute, and the file
states the same sum, so a transformer counts in no carrier.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective8452.5, 86 rows.
The network, as PyPSA code
rung_21_carrier_growth.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 21: a carrier's growth limit binds its non-generator builds — a storage unit and two stores over two periods."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def build():
"""A whole network, not the spine: cheap solar only at the first snapshot of each period, so battery capacity pays off but its growth is capped."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(2)] + [(2030, datetime(2030, 1, 1, t)) for t in range(2)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [10.0, 10.0]
n.snapshot_weightings['objective'] = [2.0, 1.5, 2.5, 2.0]
n.add('Bus', 'grid')
n.add('Carrier', 'solar')
n.add('Carrier', 'gas')
n.add('Carrier', 'battery', max_growth=20, max_relative_growth=0.5)
n.add('Generator', 'solar', bus='grid', carrier='solar', p_nom=100, marginal_cost=1, p_max_pu=[1, 0, 1, 0])
n.add('Generator', 'backup', bus='grid', carrier='gas', p_nom=200, marginal_cost=80)
n.add(
'StorageUnit',
'bat20',
bus='grid',
carrier='battery',
p_nom_extendable=True,
p_nom_max=100,
max_hours=4,
capital_cost=50,
build_year=2020,
lifetime=30,
)
n.add(
'Store',
'tank20',
bus='grid',
carrier='battery',
e_nom_extendable=True,
e_nom_max=10,
capital_cost=10,
build_year=2020,
lifetime=30,
)
n.add(
'Store',
'tank30',
bus='grid',
carrier='battery',
e_nom_extendable=True,
e_nom_max=100,
capital_cost=8,
build_year=2030,
lifetime=30,
)
n.add('Load', 'town', bus='grid', p_set=[40, 60, 40, 80])
return n
Rung 22 — transformer losses#
PyPSA applies the loss of rung 13 to every passive branch, so a transformer dissipates a loss as a line does. A transformer has:
- its own loss variable
- the loss counted against its rating
- its cap and its fan of cuts
- half of the loss at either end in the balance
The loss curve is rung 13's. A transformer's r_pu_eff
(power_flow.py:845) is its resistance over its given s_nom, times its tap
ratio. The given s_nom sets it also for an extendable transformer, whose
build does not move the curve. Here the loss of the extendable transformer
counts against its rating, so its build exceeds the flow it carries.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective10643.477135410736, 174 rows.
The network, as PyPSA code
rung_22_transformer_losses.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 22: transformer losses in tangent form — a loss per transformer, as per line."""
from __future__ import annotations
import spine
OPTIMIZE = {'transmission_losses': {'mode': 'tangents', 'segments': 3}}
def build():
"""The spine plus a triangle of one 110 kV line and two transformers, one extendable and off-nominal tap, per-unit resistances a real transformer has, so its loss stays a few percent of the flow."""
n = spine.build()
n.add('Bus', ['a', 'b', 'c'], v_nom=110)
n.add('Generator', 'hydro22', bus='a', p_nom=80, marginal_cost=10)
n.add('Generator', 'diesel22', bus='b', p_nom=80, marginal_cost=50)
n.add('Line', 'ab22', bus0='a', bus1='b', carrier='AC', x=30, r=6, s_nom=60)
n.add('Transformer', 'bc22', bus0='b', bus1='c', x=0.1, r=0.03, s_nom=60)
n.add(
'Transformer',
'ca22',
bus0='c',
bus1='a',
x=0.12,
r=0.02,
s_nom=40,
s_nom_extendable=True,
s_nom_max=90,
capital_cost=4,
tap_ratio=1.05,
)
n.add('Load', 'town22', bus='c', p_set=[35, 55, 15, 45])
return n
Rung 23 — transformer losses in secant form#
Rung 23 is rung 22's network in the secant mode. It has the same loss rows, but PyPSA's tolerance loop places the cuts.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective10821.999155213578, 142 rows.
The network, as PyPSA code
rung_23_transformer_losses_secants.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 23: transformer losses in secant form — the same loss per transformer, its cuts placed by PyPSA's tolerance loop."""
from __future__ import annotations
import spine
OPTIMIZE = {'transmission_losses': {'mode': 'secants', 'atol': 1, 'rtol': 0.1, 'max_segments': 20}}
def build():
"""Rung 22's triangle, unchanged, so the two modes differ only in the cuts."""
n = spine.build()
n.add('Bus', ['a', 'b', 'c'], v_nom=110)
n.add('Generator', 'hydro23', bus='a', p_nom=80, marginal_cost=10)
n.add('Generator', 'diesel23', bus='b', p_nom=80, marginal_cost=50)
n.add('Line', 'ab23', bus0='a', bus1='b', carrier='AC', x=30, r=6, s_nom=60)
n.add('Transformer', 'bc23', bus0='b', bus1='c', x=0.1, r=0.03, s_nom=60)
n.add(
'Transformer',
'ca23',
bus0='c',
bus1='a',
x=0.12,
r=0.02,
s_nom=40,
s_nom_extendable=True,
s_nom_max=90,
capital_cost=4,
tap_ratio=1.05,
)
n.add('Load', 'town23', bus='c', p_set=[35, 55, 15, 45])
return n
Rung 24 — must stay down#
A committable unit that stopped down_time_before snapshots before the horizon
stays off until its min_down_time has passed. PyPSA counts that stop in the
down-time row of the first min_down_time - down_time_before snapshots, which
holds the status at zero there. Here the cheapest
unit in the network brought one snapshot of a three-snapshot down time into the
horizon. It stays off for two snapshots, and the dearer coal unit serves the
load.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective9007.5, 64 rows.
The network, as PyPSA code
rung_24_must_stay_down.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 24: must stay down — a committable unit still serving the down time it brought in stays off."""
from __future__ import annotations
import spine
def build():
"""The spine plus a cheap committable unit that stopped one snapshot before the horizon and must stay off for three."""
n = spine.build()
n.add(
'Generator',
'warm',
bus='north',
committable=True,
p_nom=50,
marginal_cost=5,
p_min_pu=0.2,
min_down_time=3,
up_time_before=0,
down_time_before=1,
start_up_cost=20,
)
n.add('Load', 'swing24', bus='north', p_set=[25, 45, 45, 10])
return n
Rung 25 — committable links#
A committable link has the whole on-off scheduling of a generator (unit
commitment), over its flow.
PyPSA builds the same status, transition, up time, down time, big-M,
modular and ramp rows for a Link as for a Generator. It prices the
starts, stops and stand-by snapshots of a link the same way.
Here only committable links serve an east bus. hvdc brought one snapshot of a
three-snapshot up time into the horizon. It stays on for two snapshots,
although a cheaper link could carry the load. cold_tie brought one snapshot
of a three-snapshot down time, so it stays off for two snapshots. Started then,
it would have to stay on into the last snapshot, for its two-snapshot up time.
The load there is below its minimum, so it does not start at all.
The other links are committable builds that are extendable, modular, or both.
| PyPSA | status | note |
|---|---|---|
Link-status, -start_up, -shut_down, -n_mod |
done | as the generator's, rungs 7 and 8 |
Link-com-p-*, -com-mod-p-*, -com-ext-p-* |
done | no Link-fix-p-* or Link-ext-p-* rows for a committable link, as for a generator |
Link-*-p-fixed-upper, -*-p_nom-variable-upper |
done | |
Link-com-transition-*, -com-up-time, -com-down-time |
done | |
Link-p-ramp_limit_*, -*-bigM |
done | the generator's allowance stated with cases:, and its big-M rows, over flow |
Link-p_nom_modularity |
done | |
stand_by_cost, start_up_cost, shut_down_cost |
done |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective13982.0, 227 rows.
The network, as PyPSA code
rung_25_committable_link.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 25: committable links — a link held on by the up time it brought in, one held off by its down time, one kept on by its own up time, and committable builds that are extendable, modular or both."""
from __future__ import annotations
import spine
def build():
"""The spine plus an east bus that only committable links serve."""
n = spine.build()
n.add('Bus', 'east')
n.add(
'Link',
'hvdc',
bus0='north',
bus1='east',
committable=True,
p_nom=60,
p_min_pu=0.3,
marginal_cost=8,
min_up_time=3,
min_down_time=2,
up_time_before=1,
ramp_limit_up=0.5,
ramp_limit_down=0.5,
ramp_limit_start_up=0.6,
ramp_limit_shut_down=0.6,
start_up_cost=100,
shut_down_cost=50,
stand_by_cost=5,
)
n.add(
'Link',
'cold_tie',
bus0='north',
bus1='east',
committable=True,
p_nom=40,
p_min_pu=0.2,
min_up_time=2,
min_down_time=3,
up_time_before=0,
down_time_before=1,
start_up_cost=20,
)
n.add(
'Link',
'ext_tie',
bus0='north',
bus1='east',
committable=True,
p_nom_extendable=True,
p_nom_max=30,
capital_cost=5,
p_min_pu=0.2,
marginal_cost=2,
up_time_before=0,
ramp_limit_up=0.5,
ramp_limit_down=0.5,
)
n.add(
'Link',
'mod_tie',
bus0='south',
bus1='east',
committable=True,
p_nom_extendable=True,
p_nom_mod=10,
p_nom_max=40,
capital_cost=3,
p_min_pu=0.5,
)
n.add(
'Link',
'mod_fix',
bus0='north',
bus1='east',
committable=True,
p_nom=20,
p_nom_mod=10,
p_min_pu=0.5,
marginal_cost=1,
)
n.add('Load', 'east_load', bus='east', p_set=[20, 70, 60, 5])
return n
Rung 26 — committable processes#
A committable process has the same unit commitment over its internal power
p. The status gates p, not a port, so every port follows the status at its
own rate. This rung states the links of rung 25 again as processes. Each
draws a quarter more from the north than it delivers to the east. The same
up and down time rules hold.
| PyPSA | status | note |
|---|---|---|
Process-status, -start_up, -shut_down, -n_mod |
done | as the link's, rung 25 |
Process-com-p-*, -com-mod-p-*, -com-ext-p-* |
done | no Process-fix-p-* or Process-ext-p-* rows for a committable process |
Process-*-p-fixed-upper, -*-p_nom-variable-upper |
done | |
Process-com-transition-*, -com-up-time, -com-down-time |
done | |
Process-p-ramp_limit_*, -*-bigM |
done | the generator's allowance stated with cases:, and its big-M rows, over internal power |
Process-p_nom_modularity |
done | |
stand_by_cost, start_up_cost, shut_down_cost |
done |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective15865.124999999998, 227 rows.
The network, as PyPSA code
rung_26_committable_process.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 26: committable processes — rung 25's committable links restated as processes that draw a quarter more than they deliver."""
from __future__ import annotations
import spine
def build():
"""The spine plus an east bus that only committable processes serve."""
n = spine.build()
n.add('Bus', 'east')
n.add(
'Process',
'warm_conv',
bus0='north',
bus1='east',
rate0=-1.25,
committable=True,
p_nom=60,
p_min_pu=0.3,
marginal_cost=8,
min_up_time=3,
min_down_time=2,
up_time_before=1,
ramp_limit_up=0.5,
ramp_limit_down=0.5,
ramp_limit_start_up=0.6,
ramp_limit_shut_down=0.6,
start_up_cost=100,
shut_down_cost=50,
stand_by_cost=5,
)
n.add(
'Process',
'cold_conv',
bus0='north',
bus1='east',
rate0=-1.25,
committable=True,
p_nom=40,
p_min_pu=0.2,
min_up_time=2,
min_down_time=3,
up_time_before=0,
down_time_before=1,
start_up_cost=20,
)
n.add(
'Process',
'ext_conv',
bus0='north',
bus1='east',
rate0=-1.25,
committable=True,
p_nom_extendable=True,
p_nom_max=30,
capital_cost=5,
p_min_pu=0.2,
marginal_cost=2,
up_time_before=0,
ramp_limit_up=0.5,
ramp_limit_down=0.5,
)
n.add(
'Process',
'mod_conv',
bus0='south',
bus1='east',
rate0=-1.25,
committable=True,
p_nom_extendable=True,
p_nom_mod=10,
p_nom_max=40,
capital_cost=3,
p_min_pu=0.5,
)
n.add(
'Process',
'mod_fix',
bus0='north',
bus1='east',
rate0=-1.25,
committable=True,
p_nom=20,
p_nom_mod=10,
p_min_pu=0.5,
marginal_cost=1,
)
n.add('Load', 'east_load', bus='east', p_set=[20, 70, 60, 5])
return n
Rung 27 — modular ramps#
A committable, extendable and modular unit gets no big-M ramp rows. PyPSA gives
it the ordinary {c}-p-ramp_limit_* rows of a committed unit, with one module
p_nom_mod in place of p_nom. The status counts the modules that are on, so
the allowance grows with each module. This rung has one such generator, link
and process on a peak bus. Each has a start-up and shut-down ramp of 0.6 and a
ramp limit of one half. The ramp rows are tight at the optimum at the rise and
at the fall of the load.
| PyPSA | status | note |
|---|---|---|
{c}-p-ramp_limit_up/down, modular |
done | the committed allowance over p_nom_committed, which is one module where the build is extendable and modular, p_nom otherwise |
{c}-p-ramp_limit_*-bigM, modular |
done | none for a modular build |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective45469.49999999998, 161 rows.
The network, as PyPSA code
rung_27_modular_ramp.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 27: modular ramps — a committable extendable modular unit ramps against one module through the ordinary ramp rows, not the big-M ones."""
from __future__ import annotations
import spine
def build():
"""The spine plus a peak bus served by a committable modular generator, link and process, each ramp-limited, with a dear backup."""
n = spine.build()
n.add('Bus', 'peak')
common = {
'committable': True,
'p_nom_extendable': True,
'p_nom_mod': 20,
'p_nom_max': 60,
'capital_cost': 2,
'p_min_pu': 0.2,
'up_time_before': 0,
'ramp_limit_up': 0.5,
'ramp_limit_down': 0.5,
'ramp_limit_start_up': 0.6,
'ramp_limit_shut_down': 0.6,
}
n.add('Generator', 'mod_gen', bus='peak', marginal_cost=3, **common)
n.add('Link', 'mod_link', bus0='north', bus1='peak', marginal_cost=4, **common)
n.add('Process', 'mod_proc', bus0='south', bus1='peak', rate0=-1.25, marginal_cost=5, **common)
n.add('Generator', 'peak_backup', bus='peak', p_nom=200, marginal_cost=500)
n.add('Load', 'peak_load', bus='peak', p_set=[10, 90, 150, 20])
return n
Rung 28 — a start-up ramp alone#
PyPSA builds a ramp row where either the ramp limit or the start-up ramp is
given. It takes the missing one as 1.0, the full build. The down row is the
same with the shut-down ramp. This rung has a committable generator, link and
process with only two ramps, for start-up (0.4) and for shut-down (0.5). The
start-up ramp caps the snapshot each unit turns on, and the shut-down ramp caps
the snapshot before it turns off. PyPSA (constraints.py:1116-1117) also
builds the row at the first snapshot for either ramp alone.
A unit that is not committable starts the horizon on, whatever its
up_time_before (constraints.py:323). Data prep writes {c}_status_initial
as one for it (PyPSA/PyPSA#1943).
| PyPSA | status | note |
|---|---|---|
{c}-p-ramp_limit_up/down, start-up or shut-down ramp alone |
done | a where: on either limit, with a missing one as 1 in ramp_up_rate and its three siblings |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective83282.99999999983, 152 rows.
The network, as PyPSA code
rung_28_start_up_ramp.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 28: a start-up ramp alone — a committable unit with only a start-up and a shut-down ramp still gets ramp rows, at the full build between them."""
from __future__ import annotations
import spine
def build():
"""The spine plus a pulse bus served by committable generator, link and process that carry only start-up and shut-down ramps, with a dear backup."""
n = spine.build()
n.add('Bus', 'pulse')
common = {
'committable': True,
'p_nom': 40,
'p_min_pu': 0.1,
'up_time_before': 0,
'ramp_limit_start_up': 0.4,
'ramp_limit_shut_down': 0.5,
}
n.add('Generator', 'pulse_gen', bus='pulse', marginal_cost=3, **common)
n.add('Link', 'pulse_link', bus0='north', bus1='pulse', marginal_cost=4, **common)
n.add('Process', 'pulse_proc', bus0='south', bus1='pulse', rate0=-1.25, marginal_cost=5, **common)
n.add('Generator', 'pulse_backup', bus='pulse', p_nom=200, marginal_cost=500)
n.add('Load', 'pulse_load', bus='pulse', p_set=[0, 60, 110, 0])
return n
Rung 29 — storage per investment period#
This rung runs n.optimize(multi_investment_periods=True) with storage that
treats each investment period as a horizon of its own. A storage unit with
cyclic_state_of_charge_per_period and a store with e_cyclic_per_period close
each period on itself. The first snapshot of a period then takes the level of
the last snapshot of that period. A storage unit with
state_of_charge_initial_per_period and a store with e_initial_per_period
open each period on their initial level. The per-period cyclic flag overrides
the global one and the per-period initial flag. PyPSA reads the four flags only
under multi_investment_periods, so on a plain run they are false.
PyPSA also builds no ramp row at the first snapshot of a later period, with or without these flags. The ramp-limited coal unit in this rung raises its output from 59.2 to 90 across the period boundary, above its limit of 10 per snapshot. Inside each period, its ramp rows are tight at the optimum.
| PyPSA | status | note |
|---|---|---|
StorageUnit-energy_balance, Store-energy_balance, per period |
done | two more cases in the charge carried in: a shift(…, edge='wrap', within=snapshot_period[period]) and the initial level where position() within snapshot_period[period] is 0 |
{c}-p-ramp_limit_*, -bigM, at a period start |
done | a where: that is false at every period start but the horizon's first |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective7438.461538461539, 244 rows.
The network, as PyPSA code
rung_29_storage_per_period.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 29: storage per investment period — a storage unit and a store that cycle within each period, two that reopen on their initial level, and a ramp that restarts at a period start."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def build():
"""A whole network, not the spine: eight snapshots over two periods, four storages that each close or reopen per period, a ramp-limited coal unit."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(4)] + [(2030, datetime(2030, 1, 1, t)) for t in range(4)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [10.0, 10.0]
n.snapshot_weightings['objective'] = [2.0, 1.5, 2.5, 2.0, 2.0, 1.5, 2.5, 2.0]
n.snapshot_weightings['stores'] = [0.5, 2.0, 1.5, 2.5, 0.5, 2.0, 1.5, 2.5]
n.add('Bus', 'hub')
n.add('Generator', 'coal29', bus='hub', p_nom=100, marginal_cost=10, ramp_limit_up=0.1, ramp_limit_down=0.1)
n.add('Generator', 'peak29', bus='hub', p_nom=200, marginal_cost=[80, 20, 90, 30, 80, 20, 90, 30])
n.add('StorageUnit', 'su_cycle', bus='hub', p_nom=15, max_hours=4, cyclic_state_of_charge_per_period=True)
n.add(
'StorageUnit',
'su_reset',
bus='hub',
p_nom=15,
max_hours=4,
state_of_charge_initial=20,
state_of_charge_initial_per_period=True,
)
n.add('Store', 'e_cycle', bus='hub', e_nom=30, e_cyclic_per_period=True)
n.add('Store', 'e_reset', bus='hub', e_nom=30, e_initial=10, e_initial_per_period=True)
n.add('Load', 'hub_load', bus='hub', p_set=[40, 60, 70, 40, 90, 110, 120, 90])
return n
Rung 30 — security-constrained#
This rung runs n.optimize.optimize_security_constrained(branch_outages=...).
After any one outage of a listed passive branch, every branch of the same
sub-network stays within its rating. PyPSA (abstract.py:130-198,
:544-556) computes the sub-network's BODF and copies each flow limit row.
Each copy adds the outaged branch's flow, times its factor, to the left-hand
side. The copy has the same sense, right-hand side and extendable rating as
the row.
The outaged branch is monitored too, at the factor -1.
A copy has no loss. PyPSA (abstract.py:645-650) builds the model without
transmission_losses and linearized_unit_commitment, and hands any other
keyword to the solver (abstract.py:666). A security run is therefore lossless
and integer, and transmission_losses is false there. A network with no
passive branch is the exception. There PyPSA (abstract.py:625-632) runs a
plain n.optimize() with every keyword, and no outage exists.
An outage is a line or a transformer. A branch_outages of bare names lists
only lines. The file states the copies over an outage axis, with the factors Line_BODF and
Transformer_BODF from data prep. You compute the factors from the branches
active in each period, so they span period. A branch not active in a period
has no factor there, so it has no copy there. On a plain run, the data has no
outage, so no copy exists and the rows collapse to the standard ones.
In this rung, PyPSA outages two lines and a transformer of a meshed triangle,
and monitors the third line only. PyPSA solves the same network at
objective 17380.0 with a plain n.optimize(), and the outages raise it to
22113.33 (#620). The cheap unit at a falls to 32 in every snapshot. The
unit at b covers the rest. The build of the extendable line ca is 20.7
instead of 2 in the plain run. Six of the 120 copied rows are tight at the optimum:
Transformer-fix-s-lower against a line outage, and Line-ext-s-lower against
a transformer outage.
| PyPSA | status | note |
|---|---|---|
Line-fix-s-*-security-for-{c}-outage-in-sub-network-{n}, Line-ext-s-*-security-… |
split | a row per outaged component and sub-network in PyPSA, one entry over the outage axis in the file |
Transformer-fix-s-*-security-…, Transformer-ext-s-*-security-… |
split | the same for a transformer |
Line-fix-s-*-security-…, Transformer-fix-s-*-security-… |
done | none for a branch not active in a period, which has no factor there (rung 66) |
Line-fix-s-*-security-…, Transformer-fix-s-*-security-… |
done | in each period, the factors of that period's own active branches (rung 68) |
Line-fix-s-*-security-…, Transformer-fix-s-*-security-… |
done | over scenarios, the first scenario's factors, so Line_BODF spans no scenario (rung 56) |
…-period-{p} copies |
split | under multi_investment_periods, one entry over snapshot, each snapshot in one period |
Line-loss, Transformer-loss and the relaxed commitment rows |
done | none in a security-constrained run, which has no transmission_losses or linearized_unit_commitment, and no rung of its own |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective22113.333333333332, 240 rows.
The network, as PyPSA code
rung_30_security_constrained.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 30: security-constrained — a meshed triangle of lines and two transformers, one extendable, that must carry their flow within their rating after any one of three outages."""
from __future__ import annotations
import pandas as pd
import spine
BRANCH_OUTAGES = pd.MultiIndex.from_tuples([('Line', 'ab'), ('Line', 'ca'), ('Transformer', 'ca_t')])
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add('Bus', 'a')
n.add('Bus', 'b')
n.add('Bus', 'c')
n.add('Generator', 'hydro30', bus='a', p_nom=100, marginal_cost=10)
n.add('Generator', 'diesel30', bus='b', p_nom=100, marginal_cost=50)
n.add('Generator', 'peak30', bus='c', p_nom=100, marginal_cost=200)
n.add('Load', 'town30', bus='c', p_set=[40, 60, 80, 50])
n.add('Line', 'ab', bus0='a', bus1='b', x=0.1, s_nom=60)
n.add('Line', 'bc', bus0='b', bus1='c', x=0.1, s_nom=60, s_max_pu=0.9)
n.add('Line', 'ca', bus0='c', bus1='a', x=0.1, s_nom_extendable=True, capital_cost=5, s_nom_max=200)
n.add('Transformer', 'ca_t', bus0='c', bus1='a', x=0.2, s_nom=30)
n.add('Transformer', 'bc_t', bus0='b', bus1='c', x=0.3, s_nom_extendable=True, capital_cost=3, s_nom_max=50)
return n
Rung 32 — storage that stands in one period only#
This rung runs n.optimize(multi_investment_periods=True) with storage that is
not per period and is not active in every period. PyPSA
(constraints.py:2181-2183, 2251-2254) opens such a storage at the first
snapshot it is active in. A storage that is not cyclic opens on its initial
level. A cyclic one opens on the level of the last snapshot it is active in.
PyPSA reads the previous level through a forward fill over the snapshots the
storage is not active in.
A build year and a lifetime make those snapshots one stretch at each end of the
horizon. The file states the same rows with two parameters from data prep.
{c}_opens_late is the opening snapshot past the horizon's first.
{c}_inactive_snapshots is the number of snapshots the storage is not active
in, and a cyclic storage reaches back that many snapshots further. On a plain
run, the two are false and zero, so the rows collapse to the standard ones.
Three assumptions per class state the stretch and the opening snapshot:
StorageUnit_stands_in_one_run,
-opens_late_only_where_it_opens
and -opens_late_where_it_opens for
the storage unit. No assumption ties {c}_inactive_snapshots to the count of
inactive snapshots. A count
compared against a parameter is a
limit of the language.
This rung's network has a cyclic store that retires after the 2020 period. A
cyclic storage unit and a store with an initial level of 5 open in the 2030
period. Without the opening rows, the level before 2030 would be absent, and
each storage would open on any level it chose. With them, the objective rises
to 7230.49.
| PyPSA | status | note |
|---|---|---|
StorageUnit-energy_balance, Store-energy_balance, at the first snapshot a storage is active in |
done | the cyclic and opening cases true at position(snapshot) == 0 or at {c}_opens_late, the cyclic one shifting one snapshot and then {c}_inactive_snapshots more, edge='wrap' |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective7230.4866975671375, 108 rows.
The network, as PyPSA code
rung_32_storage_later_period.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 32: storage that stands in one period only — two built in the later period open at its first snapshot, and a cyclic one that retires closes on its own last snapshot."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def build():
"""A whole network, not the spine: eight snapshots over two periods, two storages built in 2030, one cyclic store that retires after 2020."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(4)] + [(2030, datetime(2030, 1, 1, t)) for t in range(4)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [10.0, 10.0]
n.snapshot_weightings['objective'] = [2.0, 1.5, 2.5, 2.0, 2.0, 1.5, 2.5, 2.0]
n.snapshot_weightings['stores'] = [0.5, 2.0, 1.5, 2.5, 0.5, 2.0, 1.5, 2.5]
n.add('Bus', 'hub')
n.add('Generator', 'base32', bus='hub', p_nom=100, marginal_cost=10)
n.add('Generator', 'peak32', bus='hub', p_nom=200, marginal_cost=[80, 20, 90, 30, 80, 20, 90, 30])
n.add(
'StorageUnit',
'su_late',
bus='hub',
p_nom=15,
max_hours=4,
standing_loss=0.02,
cyclic_state_of_charge=True,
build_year=2030,
lifetime=30,
)
n.add('Store', 'e_late', bus='hub', e_nom=30, e_initial=5, build_year=2030, lifetime=30)
n.add('Store', 'e_retire', bus='hub', e_nom=30, standing_loss=0.01, e_cyclic=True, build_year=2020, lifetime=10)
n.add('Load', 'hub_load', bus='hub', p_set=[40, 60, 70, 40, 90, 110, 120, 90])
return n
Rung 33 — maintenance#
A Generator, Link or Process with maintainable=True is taken off for
maintenance_events events (default 1) within the horizon. Each event
(constraints.py:826-859) covers the snapshot it starts in and the snapshots
after it, until their generators weightings reach maintenance_duration
hours. While in maintenance, the component (constraints.py:150-160, 236-249)
loses the share maintenance_pu (default 1) of its build from both of its
bounds. The status maintenance (variables.py:207-264) is continuous in
[0, 1], and the binary starts make it 0 or 1.
An extendable build multiplies a variable by a variable. PyPSA
(constraints.py:869-908) writes that product as maintenance_capacity, and
bounds it with four linear rows on a product (McCormick rows). The rows are
against p_nom_min and p_nom_max, so PyPSA refuses an extendable
maintainable build with an infinite p_nom_max.
The window of an event depends on the weightings, so its width varies along
snapshot. sum_back takes a width that does not vary along the dimension it
sums, so the file states the windows as data. A table that pairs labels
(a relation), *_maintenance_cover,
pairs each start with the snapshots it covers. The condition
*_maintenance_start_blocked marks a start whose window runs past the
horizon's end, or into a snapshot the component is not active in. You
compute both in data prep from maintenance_duration, the weightings and
active, and with them each row is the one PyPSA builds.
Across investment periods, an event may cross a period boundary, and the count is over the whole horizon, as in PyPSA. Over scenarios, every maintenance variable is second stage, with one schedule per scenario.
Here every maintainable component is new, and a peaker at 60 covers the east
bus. With the same network and no maintainable, PyPSA solves to
5388.833333333333.
| PyPSA | status | note |
|---|---|---|
Generator-maintenance, -maintenance_start, -maintenance_capacity, and the Link and Process ones |
done | zero where the component is not maintainable |
Generator-maint-event-count, -maint-window, -maint-start-horizon, and the Link and Process ones |
done | the window and the blocked starts from data prep |
Generator-maintcap_*, and the Link and Process ones |
done | -maintcap_lower_nommin only where p_nom_min > 0 |
Generator-fix-p-*, -ext-p-*, and the Link and Process ones, in maintenance |
done | the bound less maintenance_pu of the build |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective7367.560185185185, 179 rows.
The network, as PyPSA code
rung_33_maintenance.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 33: maintenance — a fixed and an extendable generator, link and process, each taken off for events that the generator weightings time."""
from __future__ import annotations
import spine
def build():
"""The spine plus an east bus and six maintainable components, fixed and extendable, with a dear peaker to cover them."""
n = spine.build()
n.add('Bus', 'east')
n.add('Generator', 'hydro', bus='north', p_nom=50, marginal_cost=2, maintainable=True, maintenance_duration=3)
n.add(
'Generator',
'wind_ext',
bus='south',
p_nom_extendable=True,
p_nom_min=10,
p_nom_max=60,
capital_cost=20,
marginal_cost=1,
p_max_pu=[0.9, 0.6, 0.8, 0.7],
maintainable=True,
maintenance_duration=2,
maintenance_pu=0.5,
)
n.add(
'Link',
'tie_fix',
bus0='north',
bus1='east',
p_nom=30,
efficiency=0.95,
maintainable=True,
maintenance_duration=1,
maintenance_events=2,
)
n.add(
'Link',
'tie_ext',
bus0='south',
bus1='east',
p_nom_extendable=True,
p_nom_min=5,
p_nom_max=40,
capital_cost=4,
efficiency=0.9,
p_min_pu=-1,
maintainable=True,
maintenance_duration=3,
)
n.add(
'Process',
'conv_fix',
bus0='north',
bus1='east',
rate0=-1.25,
p_nom=25,
maintainable=True,
maintenance_duration=3,
maintenance_pu=0.6,
)
n.add(
'Process',
'conv_ext',
bus0='south',
bus1='east',
rate0=-1.25,
p_nom_extendable=True,
p_nom_min=5,
p_nom_max=30,
capital_cost=3,
maintainable=True,
maintenance_duration=1,
)
n.add('Generator', 'east_peak', bus='east', p_nom=100, marginal_cost=60)
n.add('Load', 'east_load', bus='east', p_set=[50, 60, 40, 55])
return n
Rung 34 — maintenance of committable units#
A committable build scales its bounds by the status, so maintenance takes its
share off the status. PyPSA (variables.py:307-344, constraints.py:507-540)
writes the product of the status and maintenance as maintenance_status,
with three McCormick rows. A unit in maintenance can then also be off. A
modular committable build (constraints.py:583-621) uses the same product,
bounded by its module count in maintenance, Generator_maintenance_modules:
p_nom_max / p_nom_mod for an extendable build and p_nom / p_nom_mod for a
fixed one (rung 80). An extendable committable
build that is not modular (constraints.py:440-450) uses
maintenance_capacity in its big-M rows instead.
-com-ext-p-upper-bigM does not change. Ramps, the up and down times and the
storage rows do not contain maintenance. With the same network and no
maintainable, PyPSA solves to 5370.0.
| PyPSA | status | note |
|---|---|---|
Generator-maintenance_status, and the Link and Process ones |
done | |
Generator-maint-status-*, -maint-modstatus-*, and the Link and Process ones |
done | maint-status for a fixed build that is not modular, maint-modstatus for a modular build, fixed or not (PyPSA/PyPSA#1901) |
Generator-com-p-*, -com-mod-p-*, -com-ext-p-lower, -com-ext-p-upper-cap, and the Link and Process ones, in maintenance |
done |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective6420.048611111111, 503 rows.
The network, as PyPSA code
rung_34_committable_maintenance.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 34: maintenance of committable units — a fixed, an extendable and a modular committable generator, link and process, each taken off for one event."""
from __future__ import annotations
import spine
def build():
"""The spine plus an east bus and nine maintainable committable components, with a dear peaker to cover them."""
n = spine.build()
n.add('Bus', 'east')
committable = {'committable': True, 'maintainable': True}
n.add(
'Generator',
'unit',
bus='north',
p_nom=60,
p_min_pu=0.3,
marginal_cost=5,
start_up_cost=10,
maintenance_duration=2,
**committable,
)
n.add(
'Generator',
'unit_ext',
bus='south',
p_nom_extendable=True,
p_nom_min=5,
p_nom_max=50,
capital_cost=10,
p_min_pu=0.2,
marginal_cost=4,
maintenance_duration=3,
maintenance_pu=0.5,
**committable,
)
n.add(
'Generator',
'unit_mod',
bus='south',
p_nom_extendable=True,
p_nom_mod=10,
p_nom_max=30,
capital_cost=8,
p_min_pu=0.5,
marginal_cost=3,
maintenance_duration=1,
**committable,
)
for component, ports in (('Link', {}), ('Process', {'rate0': -1.25})):
n.add(
component,
f'{component.lower()}_unit',
bus0='north',
bus1='east',
p_nom=40,
p_min_pu=0.2,
marginal_cost=1,
maintenance_duration=2,
**ports,
**committable,
)
n.add(
component,
f'{component.lower()}_ext',
bus0='south',
bus1='east',
p_nom_extendable=True,
p_nom_min=5,
p_nom_max=30,
capital_cost=5,
p_min_pu=0.2,
maintenance_duration=3,
maintenance_pu=0.5,
**ports,
**committable,
)
n.add(
component,
f'{component.lower()}_mod',
bus0='north',
bus1='east',
p_nom_extendable=True,
p_nom_mod=10,
p_nom_max=20,
capital_cost=3,
p_min_pu=0.5,
maintenance_duration=1,
**ports,
**committable,
)
n.add('Generator', 'east_peak', bus='east', p_nom=100, marginal_cost=60)
n.add('Load', 'east_load', bus='east', p_set=[50, 60, 40, 55])
return n
Rung 35 — a global constraint for one investment period#
This rung runs n.optimize(multi_investment_periods=True) with
primary_energy and operational_limit rows that name an investment_period.
PyPSA (global_constraints.py:378-390, 601-613) sums such a row over the
snapshots of that period only, and over the whole horizon where the row names
none. Each snapshot (:333, :640) counts with its generator weighting times the
years weighting of its period. Storage closes in such a row as
follows:
- A storage unit or store that reopens per period closes at the last snapshot
of each counted period, weighted by that period's years (
:488-491,:694-697). - One that carries its level across periods closes once, at the last counted
snapshot that it is active in (
:474-477,:680-683, rung 69).
The file states the counted snapshots as GlobalConstraint_counts_snapshot,
from data prep, and the years as period_weight_years. On a plain run, the
counted snapshots are all true and the years are one, so the rows collapse to
the standard ones.
This rung's network caps CO2 in 2030 alone and again over the horizon. It
limits a hydro carrier with a per-period storage unit and store in the 2020
period. The two periods weigh 5 and 10 years. With the same network and no
investment_period, PyPSA solves to 10675.0.
| PyPSA | status | note |
|---|---|---|
primary_energy, operational_limit over one investment period |
done | GlobalConstraint_energy_weight, zero outside the counted snapshots, with the years weighing each snapshot |
primary_energy, operational_limit over one investment period |
done | the closing level of storage as StorageUnit_closing_weight and Store_closing_weight, at each counted period's last snapshot where the storage reopens per period, at the last counted snapshot it is active in otherwise |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective4886.764705882353, 171 rows.
The network, as PyPSA code
rung_35_period_global_constraints.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 35: a global constraint for one investment period — a CO2 cap on 2030 alone, one over the horizon, and a 2020 limit on a carrier with storage that reopens per period."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def build():
"""A whole network, not the spine: eight snapshots over two periods of unequal years, two emitting units, a hydro carrier with a unit and storage."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(4)] + [(2030, datetime(2030, 1, 1, t)) for t in range(4)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [5.0, 10.0]
n.snapshot_weightings['generators'] = [1.0, 2.0, 1.0, 2.0, 1.0, 2.0, 1.0, 2.0]
n.add('Carrier', 'coal35', co2_emissions=1.0)
n.add('Carrier', 'gas35', co2_emissions=0.4)
n.add('Carrier', 'hydro35')
n.add('Bus', 'hub')
n.add('Generator', 'coal35', bus='hub', carrier='coal35', p_nom=100, marginal_cost=10, efficiency=0.4)
n.add('Generator', 'gas35', bus='hub', carrier='gas35', p_nom=100, marginal_cost=30, efficiency=0.5)
n.add('Generator', 'clean35', bus='hub', p_nom=200, marginal_cost=60)
n.add('Generator', 'river35', bus='hub', carrier='hydro35', p_nom=30, marginal_cost=5)
n.add(
'StorageUnit',
'dam35',
bus='hub',
carrier='hydro35',
p_nom=15,
max_hours=4,
state_of_charge_initial=20,
state_of_charge_initial_per_period=True,
)
n.add('Store', 'pond35', bus='hub', carrier='hydro35', e_nom=30, e_initial=10, e_initial_per_period=True)
n.add('Load', 'town35', bus='hub', p_set=[60, 80, 70, 50, 90, 110, 100, 80])
n.add(
'GlobalConstraint',
'co2_2030',
type='primary_energy',
carrier_attribute='co2_emissions',
sense='<=',
constant=3500,
investment_period=2030,
)
n.add(
'GlobalConstraint',
'co2_all',
type='primary_energy',
carrier_attribute='co2_emissions',
sense='<=',
constant=6000,
)
n.add(
'GlobalConstraint',
'hydro_2020',
type='operational_limit',
carrier_attribute='hydro35',
sense='<=',
constant=600,
investment_period=2020,
)
return n
Rung 36 — quadratic costs on a process and on storage#
PyPSA (variables.csv:22, :30, :33) declares marginal_cost_quadratic on
three other components, and this rung puts it on all three. A process pays on
its internal power p, and a storage unit on p_dispatch only. A store pays on
its net p, so charging costs as much as delivering. Each quadratic cost
(optimize.py:349-366) is the square times marginal_cost_quadratic, weighted
as the linear cost is. On a plain run, marginal_cost_quadratic is zero, so
the quadratic costs are zero.
The cost of this rung's storage unit changes per snapshot. With the same network
and no quadratic cost, PyPSA solves to 17641.666666666668.
| PyPSA | status | note |
|---|---|---|
marginal_cost_quadratic on Process, StorageUnit and Store |
done | degree 2 in the objective, with no charge on p_store |
marginal_cost_quadratic |
refused | under a risk preference, in Refusals |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective19185.241281403858, 92 rows.
The network, as PyPSA code
rung_36_quadratic_storage_process.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 36: quadratic costs on a process, a storage unit and a store — the storage unit pays on dispatch only, the store on its net power both ways."""
from __future__ import annotations
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add('Bus', 'hub')
n.add('Process', 'conv36', bus0='north', bus1='hub', p_nom=60, marginal_cost=1, marginal_cost_quadratic=0.05)
n.add(
'StorageUnit',
'battery36',
bus='south',
p_nom=20,
max_hours=4,
state_of_charge_initial=40,
marginal_cost=0.5,
marginal_cost_quadratic=[0.2, 0.1, 0.3, 0.1],
)
n.add('Store', 'tank36', bus='hub', e_nom=60, e_initial=20, marginal_cost_quadratic=0.4)
n.add('Load', 'hub_load', bus='hub', p_set=[20, 45, 30, 50])
n.add('Load', 'peak36', bus='south', p_set=[10, 40, 20, 60])
return n
Rung 37 — storage dispatch pinned to a schedule#
PyPSA pins a store's power delivered to p_set. It pins a storage unit's
dispatch to p_dispatch_set and its charging to p_store_set, each on its
own. PyPSA builds a row only where a value is given and the storage is active.
On a plain run, the data has no value, so PyPSA builds no row.
The rung pins a store to deliver 10 in the first snapshot and to take 5 in the
last. It pins the dispatch of a storage unit in the first snapshot, and its
charging in the second. With the same network and no pins, PyPSA solves to
5811.111111111111.
| PyPSA | status | note |
|---|---|---|
Store-p_set, StorageUnit-p_dispatch_set, StorageUnit-p_store_set |
done | where: a value is given and the storage is active, as in PyPSA (optimize.py:917, :922, constraints.py:2156-2220) |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective6395.833333333333, 84 rows.
The network, as PyPSA code
rung_37_fixed_storage_dispatch.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 37: storage dispatch pinned — a store's power delivered, and a storage unit's dispatch and charging, each on its own schedule."""
from __future__ import annotations
from math import nan
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.generators_t.marginal_cost['gas'] = [15, 60, 15, 60]
n.add('Store', 'tank37', bus='south', e_nom=40, e_initial=20, p_set=[10, nan, nan, -5])
n.add(
'StorageUnit',
'battery37',
bus='south',
p_nom=20,
max_hours=2,
state_of_charge_initial=10,
p_dispatch_set=[6, nan, nan, nan],
p_store_set=[nan, 4, nan, nan],
)
return n
Rung 38 — delays per investment period#
This rung runs n.optimize(multi_investment_periods=True) with delayed ports.
PyPSA applies a link's or a process's delay in each investment period on its
own (constraints.py:1370-1378, multiports.py:212-219). A cyclic_delay
port wraps from the end of its own period. A port that is not cyclic loses the
flow still in transit at the first snapshots of every period. PyPSA measures
the delay in generators weighting per period and rounds it down to a snapshot
start (multiports.py:106-123). Scenarios do not change the source snapshot.
The file states a shift by the same whole number of snapshots in every period.
It states an assumption for it,
Link_output_delay_under_uniform_weighting.
When data is attached, a tool refuses a delayed port whose generators
weighting has more than one value over the horizon (#299).
Link_output_arrival and Process_output_arrival shift with
within=snapshot_period[period]. On a plain run, the data has one period, so
the shift is the flat one.
The rung adds a link that delays by two and wraps, and a process that delays
by one and does not wrap. Both run on two periods of four snapshots. A shift
over the flat horizon would let 2030 receive the flow that 2020 sent. With that
flat shift patched into PyPSA's source index, PyPSA solves the network to
10543.75 (#620).
| PyPSA | status | note |
|---|---|---|
link and process delay, cyclic_delay per investment period |
done | shift(offset=delay, within=snapshot_period[period]), with edge='wrap' closing each period and edge=0 vacating each period's first snapshots |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective12918.75, 104 rows.
The network, as PyPSA code
rung_38_delay_per_period.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 38: delays per investment period — a link that wraps its delayed flow within each period, and a process that loses what is still in transit at each period's start."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
#: The `generators` weighting is uniform, as on rung 16, so a delay of `n` is a
#: shift of exactly `n` positions. The `objective` column stays non-uniform.
WEIGHTINGS = {'objective': [2.0, 1.5, 2.5, 3.0, 2.0, 1.5, 2.5, 3.0], 'generators': [1.0] * 8}
#: Demand differs from snapshot to snapshot, so which snapshot a delayed flow is
#: read from changes what it costs.
DEMAND = [20.0, 15.0, 25.0, 10.0, 30.0, 35.0, 5.0, 40.0]
def build():
"""A whole network, not the spine: eight snapshots over two periods, a source, a delayed link and a delayed process.
``pipe_wrap`` delays by two snapshots and wraps cyclically, so the first two
snapshots of each period read the last two of that same period, never the
other period. ``conv_lose`` delays by one and does not wrap, so the first
snapshot of each period, 2030 included, receives nothing and its demand falls
to the backup. The source is capped, so where each delayed flow is read from
decides how much of the backup runs.
"""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(4)] + [(2030, datetime(2030, 1, 1, t)) for t in range(4)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [10.0, 10.0]
for column, values in WEIGHTINGS.items():
n.snapshot_weightings[column] = values
n.add('Bus', ['source', 'sink_wrap', 'sink_lose'])
n.add('Generator', 'spring38', bus='source', p_nom=60, marginal_cost=5)
n.add('Generator', 'backup_wrap38', bus='sink_wrap', p_nom=200, marginal_cost=100)
n.add('Generator', 'backup_lose38', bus='sink_lose', p_nom=200, marginal_cost=100)
n.add('Link', 'pipe_wrap', bus0='source', bus1='sink_wrap', p_nom=30, delay=2, cyclic_delay=True)
n.add('Process', 'conv_lose', bus0='source', bus1='sink_lose', p_nom=30, delay1=1, cyclic_delay1=False)
n.add('Load', 'load_wrap', bus='sink_wrap', p_set=DEMAND)
n.add('Load', 'load_lose', bus='sink_lose', p_set=DEMAND)
return n
Rung 39 — a negative relative growth#
This rung runs n.optimize(multi_investment_periods=True) with a carrier whose
max_relative_growth is negative. PyPSA clips the share at zero before it
builds Carrier-growth_limit (global_constraints.py:250), so a negative
share adds nothing and does not tighten the limit. Carrier_relative_growth
states the clip as a case: the given share where it is positive, zero
otherwise. On a plain run, the data has one period, and PyPSA builds no
growth row.
The rung adds a battery carrier with max_growth=20 and
max_relative_growth=-0.5, and two stores, built in 2020 and 2030
respectively. Taking the share as given would add half of the 2020 build to the
left side of the 2030 row. With that row patched into PyPSA through
extra_functionality, PyPSA solves the network to 9347.5 (#620).
| PyPSA | status | note |
|---|---|---|
Carrier-growth_limit, max_relative_growth.clip(min=0) |
done | Carrier_relative_growth is Carrier_max_relative_growth where it is positive, 0 otherwise |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective8600.0, 56 rows.
The network, as PyPSA code
rung_39_negative_relative_growth.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 39: a negative relative growth adds nothing to a carrier's growth limit — PyPSA clips it at zero."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def build():
"""A whole network, not the spine: a battery carrier with `max_relative_growth=-0.5` builds in both periods."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(2)] + [(2030, datetime(2030, 1, 1, t)) for t in range(2)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [10.0, 10.0]
n.snapshot_weightings['objective'] = [2.0, 1.5, 2.5, 2.0]
n.add('Bus', 'grid')
n.add('Carrier', 'solar')
n.add('Carrier', 'gas')
n.add('Carrier', 'battery', max_growth=20, max_relative_growth=-0.5)
n.add('Generator', 'solar', bus='grid', carrier='solar', p_nom=100, marginal_cost=1, p_max_pu=[1, 0, 1, 0])
n.add('Generator', 'backup', bus='grid', carrier='gas', p_nom=200, marginal_cost=80)
n.add(
'Store',
'tank20',
bus='grid',
carrier='battery',
e_nom_extendable=True,
e_nom_max=100,
capital_cost=10,
build_year=2020,
lifetime=30,
)
n.add(
'Store',
'tank30',
bus='grid',
carrier='battery',
e_nom_extendable=True,
e_nom_max=100,
capital_cost=8,
build_year=2030,
lifetime=30,
)
n.add('Load', 'town', bus='grid', p_set=[40, 60, 40, 80])
return n
Rung 40 — a global constraint per scenario#
The reference script calls n.set_scenarios(...) with GlobalConstraint rows whose
constant and sense differ between the scenarios. PyPSA builds one
GlobalConstraint-{name} row per scenario for primary_energy,
operational_limit and transmission_volume_expansion_limit. It reads each
scenario's own sense and constant (global_constraints.py:366-376,
:576-577, :779-780, :814-822, :854-864). The file states
GlobalConstraint_constant and GlobalConstraint_sense over scenario as
well, so each row takes its own value in each future. A row with no scenario
axis in its total, such as the transmission volume, repeats the same capacity
sum under each scenario's constant. On a plain run, the data has one scenario,
and the rows collapse to the standard ones.
The rung puts an extendable line under a volume limit of 60 in the calm
future and 20 in the stormy one. It adds a CO2 row that is at most 250 in
the calm future and exactly 200 in the stormy one. All three are tight at
the optimum. With the same network and the calm values in both futures, PyPSA
solves to 12943.333333333334.
| PyPSA | status | note |
|---|---|---|
primary_energy, operational_limit, transmission_volume_expansion_limit with a constant and sense per scenario |
done | The file states GlobalConstraint_constant and GlobalConstraint_sense over scenario. |
primary_energy, operational_limit, transmission_volume_expansion_limit |
done | PyPSA reads a carrier_attribute or investment_period per scenario (global_constraints.py:823-830). The file states the weights and GlobalConstraint_counts_snapshot over scenario. |
Rungs 52 and 53 extend these rows.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective15106.666666666666, 86 rows.
The network, as PyPSA code
rung_40_scenario_global_constraints.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 40: a global constraint takes its own constant and sense in each scenario."""
from __future__ import annotations
import spine
#: each scenario's own constant and sense, per row
PER_SCENARIO = {
('calm', 'volume40'): {'constant': 60},
('stormy', 'volume40'): {'constant': 20},
('calm', 'co2_40'): {'constant': 250, 'sense': '<='},
('stormy', 'co2_40'): {'constant': 200, 'sense': '=='},
}
def build():
"""The spine over two futures, with an extendable line under a volume limit and a CO2 row that differ by scenario."""
n = spine.build()
n.add('Carrier', 'AC')
n.add('Carrier', 'coalc', co2_emissions=0.5)
n.c.generators.static.loc['coal', 'carrier'] = 'coalc'
n.add(
'Line',
'tie40',
bus0='north',
bus1='south',
x=0.1,
carrier='AC',
length=2,
s_nom_extendable=True,
capital_cost=1,
)
n.add('Load', 'port40', bus='south', p_set=30)
n.add(
'GlobalConstraint',
'volume40',
type='transmission_volume_expansion_limit',
carrier_attribute='AC',
sense='<=',
constant=60,
)
n.add(
'GlobalConstraint', 'co2_40', type='primary_energy', carrier_attribute='co2_emissions', sense='<=', constant=250
)
n.set_scenarios({'calm': 0.6, 'stormy': 0.4})
for row, values in PER_SCENARIO.items():
for column, value in values.items():
n.c.global_constraints.static.loc[row, column] = value
return n
Rung 41 — operating data per scenario#
The reference script calls n.set_scenarios(...) with a gas unit's marginal_cost and a
link's efficiency set per scenario. PyPSA reads both through c.da, one
value per scenario, into the objective and into Bus-nodal_balance
(components/array.py:347-406). The file states Generator_marginal_cost and
Link_efficiency over scenario. On a plain run, the data has one scenario,
and the rows collapse to the standard ones.
The rung adds a gas unit that costs 20 in the calm future and 80 in the
stormy one. A second link delivers 0.9 and 0.6 of its flow. Each value
changes the objective. With the calm value in both futures, PyPSA solves to
these objectives.
| calm value in both futures | objective |
|---|---|
gas unit marginal_cost |
16308.0 |
link efficiency |
16992.0 |
| both | 15660.0 |
| PyPSA | status | note |
|---|---|---|
objective, Bus-nodal_balance with a cost and an efficiency per scenario |
done | Generator_marginal_cost and Link_efficiency over scenario |
Bus-nodal_balance |
refused | a link delay or cyclic_delay that differs by scenario, Refusals |
Rungs 55, 57 and 58 extend these rows.
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective17964.0, 96 rows.
The network, as PyPSA code
rung_41_scenario_operational_data.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 41: a unit's cost and a link's efficiency differ by scenario."""
from __future__ import annotations
import spine
#: each scenario's own value, per component and attribute
PER_SCENARIO = {
('calm', 'gas41'): {'marginal_cost': 20},
('stormy', 'gas41'): {'marginal_cost': 80},
('calm', 'wire41'): {'efficiency': 0.9},
('stormy', 'wire41'): {'efficiency': 0.6},
}
def build():
"""The spine over two futures, with a gas unit that costs more and a link that delivers less in the stormy one."""
n = spine.build()
n.add('Generator', 'gas41', bus='south', p_nom=100, marginal_cost=20)
n.add('Link', 'wire41', bus0='north', bus1='south', p_nom=40, efficiency=0.9)
n.add('Load', 'port41', bus='south', p_set=60)
n.set_scenarios({'calm': 0.6, 'stormy': 0.4})
for (scenario, name), values in PER_SCENARIO.items():
component = n.c.generators if name == 'gas41' else n.c.links
for column, value in values.items():
component.static.loc[(scenario, name), column] = value
return n
Rung 42 — first-stage data per scenario#
The reference script calls n.set_scenarios(...) with an extendable unit whose
capital_cost and p_nom_max differ between the scenarios. The build is
chosen once, but PyPSA reads its bounds per scenario. It writes
Generator-ext-p_nom-lower and -upper once per scenario
(constraints.py:949-962), so the tightest cap bounds the build. PyPSA prices
the build at each scenario's capital cost and weights the costs by the scenario
weights (optimize.py:438-445, :484-490). The build pays its capital cost in
expectation.
The file states Generator_p_nom_max, Generator_capital_cost and the bound
rows over scenario. It states the capital costs in the objective under
scenario_weight. On a plain run, the data has one scenario of weight one, and
the rows and the objective collapse to the standard ones.
The rung's wind unit costs 20 in the calm future and 60 in the stormy one,
weighted 0.6 and 0.4. It may be built to 100 and 30. PyPSA builds 30
at the expected cost of 36. With a cost of 36 and a cap of 30 in both
futures, PyPSA solves the same network to the same objective. With the calm
values in both futures, PyPSA solves to 1943.0.
| PyPSA | status | note |
|---|---|---|
{c}-ext-p_nom-lower/upper with a bound per scenario |
done | one row per scenario, over scenario, the tightest bounding the build |
| capital cost per scenario | done | scenario_weight times each scenario's capital cost |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective5049.999999999999, 84 rows.
The network, as PyPSA code
rung_42_scenario_first_stage.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 42: an extendable unit's capital cost and build cap differ by scenario."""
from __future__ import annotations
import spine
#: each scenario's own value, per attribute of the extendable unit
PER_SCENARIO = {
'calm': {'capital_cost': 20, 'p_nom_max': 100},
'stormy': {'capital_cost': 60, 'p_nom_max': 30},
}
def build():
"""The spine over two futures, with an extendable wind unit that costs more and may be built less in the stormy one."""
n = spine.build()
n.add('Generator', 'wind42', bus='south', p_nom_extendable=True, p_nom_max=100, marginal_cost=1, capital_cost=20)
n.set_scenarios({'calm': 0.6, 'stormy': 0.4})
for scenario, values in PER_SCENARIO.items():
for column, value in values.items():
n.c.generators.static.loc[(scenario, 'wind42'), column] = value
return n
Rung 43 — a component's sign#
This rung runs n.optimize() with a generator, a load, a storage unit and a
store whose sign is not PyPSA's default. PyPSA multiplies each of these in
Bus-nodal_balance by that sign:
- a generator's
p - a storage unit's
p_dispatchandp_store - a store's
p(constraints.py:1474-1475) - a load's
p_set, on the constant side (constraints.py:1586)
PyPSA reads sign nowhere else in the model it builds. The default is 1 for
a generator, a storage unit and a store, and -1 for a load. PyPSA refuses a
sign that differs by scenario (constants.py:46). The file states
Generator_sign, Load_sign, StorageUnit_sign and Store_sign without
scenario. On a plain run, the data has PyPSA's defaults, and the row
collapses to the standard one.
The rung adds a unit that draws 20 from its bus at a cost of -50. It adds
a storage unit that opens full and a full store. These three have a sign of
-1. The rung also adds a load of 10 with a sign of 1, which feeds its
bus. Each sign changes the objective. With the default sign restored, PyPSA solves to these
objectives.
| default sign restored on | objective |
|---|---|
| the unit | -5652.78 |
| the load | 3925.0 |
| the storage unit | 1187.5 |
| the store | 925.0 |
| all four | -5255.56 |
| PyPSA | status | note |
|---|---|---|
Bus-nodal_balance with a component sign |
done | Generator_sign, StorageUnit_sign and Store_sign times each variable, and Load_sign times the load, negated |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective2125.0, 88 rows.
The network, as PyPSA code
rung_43_sign.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 43: a component's `sign` turns its term in the bus balance around."""
from __future__ import annotations
import spine
#: the sign each component enters the bus balance with, against PyPSA's default
SIGNS = {'generators': ('flex43', -1), 'loads': ('feed43', 1), 'storage_units': ('su43', -1), 'stores': ('e43', -1)}
def build():
"""The spine with a unit that draws power, a load that feeds it, and a storage unit and a store drawn the other way round."""
n = spine.build()
n.add('Generator', 'flex43', bus='south', p_nom=20, marginal_cost=-50, sign=-1)
n.add('Load', 'feed43', bus='north', p_set=10, sign=1)
n.add('StorageUnit', 'su43', bus='south', p_nom=10, max_hours=2, state_of_charge_initial=20, sign=-1)
n.add('Store', 'e43', bus='north', e_nom=30, e_initial=30, sign=-1)
return n
Rung 45 — ramp limits per snapshot#
This rung runs n.optimize() with a generator, a link and a process whose
ramp_limit_up and ramp_limit_down change over time. PyPSA declares both as
a constant or a time series (static or series) for all three components. It
declares ramp_limit_start_up and ramp_limit_shut_down static. The row
between two snapshots takes the limit at the later snapshot
(constraints.py:1104-1105, 1089-1090, 1119-1120). PyPSA makes the "no
limit" test per snapshot (constraints.py:1110-1111), and takes a missing
value as the full build there (constraints.py:1116-1119).
The file states {c}_ramp_limit_up and {c}_ramp_limit_down over snapshot,
and {c}_ramp_up_rate and {c}_ramp_down_rate with them. A snapshot without a
value has no ramp row, unless the row has a start-up or shut-down ramp there.
On a plain run, the data has the same value at each snapshot, and the rows
collapse to the standard ones.
The rung adds a steep bus with a load of 90 at the third snapshot. Each unit
may raise its output by 0.1 of its build, and by 0.5 into the third
snapshot. It may lower its output by 0.1, and without a limit into the last
snapshot. With the constant limit 0.1 in each direction, the backup covers
the peak, and PyPSA solves to 90871.0 (#620).
| PyPSA | status | note |
|---|---|---|
{c}-p-ramp_limit_up/down with a limit per snapshot |
done | {c}_ramp_limit_up, {c}_ramp_limit_down and their rates over snapshot, with no row at a snapshot without a value, unless the row has a start-up or shut-down ramp there |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective10996.0, 83 rows.
The network, as PyPSA code
rung_45_ramp_per_snapshot.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 45: ramp limits per snapshot — a generator, a link and a process whose ramp limits change over time, and lift at one snapshot."""
from __future__ import annotations
import math
import spine
def build():
"""The spine plus a steep bus served by a generator, a link and a process with ramp limits per snapshot, with a dear backup."""
n = spine.build()
n.add('Bus', 'steep')
up = [0.1, 0.1, 0.5, 0.1]
down = [0.1, 0.1, 0.1, math.nan]
n.add('Generator', 'steep_gen', bus='steep', p_nom=60, marginal_cost=3, ramp_limit_up=up, ramp_limit_down=down)
n.add(
'Link',
'steep_link',
bus0='north',
bus1='steep',
p_nom=40,
marginal_cost=4,
ramp_limit_up=up,
ramp_limit_down=down,
)
n.add(
'Process',
'steep_proc',
bus0='south',
bus1='steep',
rate0=-1.25,
p_nom=40,
marginal_cost=5,
ramp_limit_up=up,
ramp_limit_down=down,
)
n.add('Generator', 'steep_backup', bus='steep', p_nom=200, marginal_cost=500)
n.add('Load', 'steep_load', bus='steep', p_set=[10, 20, 90, 10])
return n
Rung 46 — the output brought in#
This rung runs n.optimize() with units that have p_init, the output they
brought into the horizon. PyPSA reads p_init only where a unit came in
running (up_time_before > 0, or not committable), and zero where it came in
off (constraints.py:338). It builds the ramp rows at the first snapshot where
that value exists (constraints.py:1151). A unit that came in running
without p_init has no ramp row there. PyPSA carries the value into the first
snapshot's rows with the status the unit came in with
(constraints.py:1158-1163).
PyPSA builds these rows the same way for a Generator, a Link and a Process. It
does so for a fixed, an extendable and a committable build, and for the big-M
rows of a committable extendable build (constraints.py:1004-1007). PyPSA reads
p_init per scenario. Under multi_investment_periods, it reads p_init only
at the horizon's first snapshot, as PyPSA builds no ramp row at a later period
start (constraints.py:1159-1162).
The file states {c}_p_init, and the output carried in at the first snapshot
is {c}_status_initial * {c}_p_init. The first-snapshot where: is
{c}_status_initial == 0 OR {c}_p_init. On a plain run, the data has no
p_init, and the rows collapse to the standard ones.
PyPSA warns where a committable generator came in off and has a p_init, and
ignores the value (consistency.py:669-680). The product with
{c}_status_initial is zero there too. A unit that is not committable comes in
running, whatever its up_time_before (rung 71).
PyPSA sets the first-snapshot condition without the active condition. A unit
that is not active at the first snapshot still gets a row there, with no
variable in it. The file does not state that row.
The rung adds a warm bus with a load of 150, then 60, served by six
ramp-limited units:
- a fixed generator
- a link from
p_init = 0 - a dear process that came in at full output
- an extendable generator
- a committable generator
- a committable extendable generator
Each p_init is tight at the optimum. Without any of them, PyPSA solves to
9921.0 (#620).
| PyPSA | status | note |
|---|---|---|
{c}-p-ramp_limit_*, -bigM, at the first snapshot |
done | {c}_previous_p opens on {c}_status_initial * {c}_p_init, and the row exists where {c}_status_initial == 0 OR {c}_p_init |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective101294.125, 200 rows.
The network, as PyPSA code
rung_46_initial_output.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 46: the output brought in — units that came in running with a given `p_init` ramp from it into the first snapshot."""
from __future__ import annotations
import spine
def build():
"""The spine plus a warm bus served by fixed, extendable and committable units that each carry a `p_init`, with a dear backup."""
n = spine.build()
n.add('Bus', 'warm')
ramps = {'ramp_limit_up': 0.25, 'ramp_limit_down': 0.25}
n.add('Generator', 'warm_gen', bus='warm', p_nom=60, marginal_cost=3, p_init=10, **ramps)
n.add('Link', 'warm_link', bus0='north', bus1='warm', p_nom=40, marginal_cost=4, p_init=0, **ramps)
n.add(
'Process', 'warm_proc', bus0='south', bus1='warm', rate0=-1.25, p_nom=40, marginal_cost=600, p_init=40, **ramps
)
n.add(
'Generator',
'warm_ext',
bus='warm',
p_nom_extendable=True,
p_nom_max=40,
capital_cost=5,
marginal_cost=2,
p_init=5,
**ramps,
)
n.add(
'Generator',
'warm_com',
bus='warm',
committable=True,
p_nom=50,
p_min_pu=0.2,
marginal_cost=2.5,
p_init=30,
ramp_limit_start_up=0.4,
ramp_limit_shut_down=0.4,
**ramps,
)
n.add(
'Generator',
'warm_com_ext',
bus='warm',
committable=True,
p_nom_extendable=True,
p_nom_max=40,
capital_cost=5,
marginal_cost=2.2,
p_init=5,
**ramps,
)
n.add('Generator', 'warm_backup', bus='warm', p_nom=300, marginal_cost=500)
n.add('Load', 'warm_load', bus='warm', p_set=[150, 150, 60, 60])
return n
Rung 48 — a start and a stop, weighted by the period#
This rung runs n.optimize(multi_investment_periods=True) with a committable
unit that starts and stops. PyPSA adds start_up_cost * start_up and
shut_down_cost * shut_down to the objective with the objective weight of the
period, and without the weight of the snapshot (optimize.py:294-297,
:464-465). It weights every other operating cost by both. PyPSA/PyPSA#2011
added the period weight, which fixed PyPSA/PyPSA#1986. Over
scenarios, it weights the start and stop costs by the scenario weight, as every
operating cost. The file states the two costs in scenario_opex with the
period weight, and the scenario weight applies on top.
The rung adds a committable peaker that starts once and stops once in each of
two periods. The period weights are 1.0 and 0.5, and the snapshot weights
are not 1.0. PyPSA solves to 7125.0, and to 6525.0 without the start and
stop costs. The difference is 600.0, the four events weighted by their period.
Without the period weight, the difference is 800.0 (#620).
| PyPSA | status | note |
|---|---|---|
start_up_cost, shut_down_cost under multi_investment_periods and snapshot weights |
done | the period weight and the scenario weight, no snapshot weight |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective7125.0, 72 rows.
The network, as PyPSA code
rung_48_start_up_period_weight.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 48: a start and a stop take their period's weight and no snapshot weight, over two weighted periods."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def build():
"""A whole network, not the spine: a committable peaker that starts and stops once in each period."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(3)] + [(2030, datetime(2030, 1, 1, t)) for t in range(3)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [10.0, 10.0]
n.snapshot_weightings['objective'] = [2.0, 1.5, 2.5, 2.0, 1.5, 2.5]
n.add('Bus', 'grid')
n.add('Generator', 'base48', bus='grid', p_nom=60, marginal_cost=10)
n.add('Generator', 'dear48', bus='grid', p_nom=100, marginal_cost=90)
n.add(
'Generator',
'peak48',
bus='grid',
p_nom=50,
marginal_cost=20,
committable=True,
p_min_pu=0.4,
start_up_cost=300,
shut_down_cost=100,
up_time_before=0,
)
n.add('Load', 'town48', bus='grid', p_set=[50, 100, 50, 50, 100, 50])
return n
Rung 49 — a growth limit in one period#
This rung runs n.optimize() with a carrier that has max_growth. PyPSA
builds Carrier-growth_limit only under multi_investment_periods, and
returns before it reads the carrier otherwise
(global_constraints.py:232-233). A single-period run has no growth limit.
The file states the row where Carrier_max_growth has a value.
Carrier_max_growth has no value on a single-period run, so no row is built.
The rung adds a cheap extendable wind unit with max_growth = 10 to the shared
network most rungs start from (the spine). PyPSA builds it to 67 and solves
to 335.0, the same as without the limit. With the same network as one
investment period under multi_investment_periods, PyPSA caps the build at
10 and solves to 2583.33 (#620).
| PyPSA | status | note |
|---|---|---|
Carrier-growth_limit without multi_investment_periods |
done | The file states no row, as Carrier_max_growth has no value on a single-period run. |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective335.0, 42 rows.
The network, as PyPSA code
rung_49_single_period_growth.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 49: a carrier's growth limit without `multi_investment_periods` — PyPSA builds no row, so the build passes the cap."""
from __future__ import annotations
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add('Carrier', 'wind', max_growth=10)
n.add('Generator', 'wind49', bus='south', carrier='wind', p_nom_extendable=True, p_nom_max=100, capital_cost=5)
return n
Rung 50 — a load that is not active#
This rung runs n.optimize() with a load whose active is false. PyPSA
conditions the load side of Bus-nodal_balance on active
(constraints.py:1585-1586), so an inactive load draws nothing. A load has no
build year and no lifetime. Its active is the static flag alone, in every
snapshot and every period (descriptors.py:135-136). PyPSA refuses an
active that differs by scenario (constants.py:54).
The file states Load_active over the load, and Load_demand, the load's
signed demand where it is active and zero where it is not. The balance
contains Load_demand. On a plain run, every load is active in the data, and
the row collapses to the standard one.
The rung adds an inactive load to the spine's south bus. PyPSA solves to
7380.0, the same as without the load. With the load active, PyPSA solves to
14730.0 (#620).
| PyPSA | status | note |
|---|---|---|
Bus-nodal_balance with a load that is not active |
done | Load_demand is zero where Load_active is false |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective7380.0, 32 rows.
The network, as PyPSA code
rung_50_inactive_load.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 50: a load that is not `active` — PyPSA drops it from its bus's balance."""
from __future__ import annotations
import spine
def build():
"""The spine plus this rung's additions, as a ``pypsa.Network``."""
n = spine.build()
n.add('Load', 'idle50', bus='south', p_set=[30, 10, 20, 40], active=False)
return n
Rung 51 — a growth limit after an asset retires#
This rung runs n.optimize(multi_investment_periods=True) with a carrier that
has max_growth, and an extendable asset of that carrier that retires before
the last period. In the file, a build counts only in the first period it is
active in: {c}_first_active is one there and zero elsewhere. PyPSA takes
active & (active.cumsum() == 1), which is false after the asset retires
(global_constraints.py:278).
The rung adds two solar units under one carrier with max_growth = 10. The
old one is active in 2020 only, the new one in 2030 only. Both build to 10,
and PyPSA solves to 3432.5. A limit that counts the old unit again in 2030
caps the new one at 10 minus the old one's build. PyPSA then solves to
5185.0. Without max_growth, PyPSA solves to 248.75.
| PyPSA | status | note |
|---|---|---|
Carrier-growth_limit with an asset that retires |
done | {c}_first_active is zero after the first period an asset is active in |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective3432.5, 26 rows.
The network, as PyPSA code
rung_51_growth_retired_asset.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 51: a carrier's growth limit counts an asset in the first period it stands in only, not again after it retires.
The old unit retires before the new one is built, so each builds to the
limit in its own period.
"""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def network(carriers: dict[str, str]):
"""Two periods, a solar unit that stands in 2020 only and one built in 2030, each under the carrier named for it."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(2)] + [(2030, datetime(2030, 1, 1, t)) for t in range(2)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [10.0, 10.0]
n.snapshot_weightings['objective'] = [2.0, 1.5, 2.5, 2.0]
n.add('Bus', 'grid')
n.add('Carrier', 'gas')
for carrier in sorted(set(carriers.values())):
n.add('Carrier', carrier, max_growth=10)
for name, build_year in (('solar_old', 2020), ('solar_new', 2030)):
n.add(
'Generator',
name,
bus='grid',
carrier=carriers[name],
p_nom_extendable=True,
p_nom_max=50,
marginal_cost=1,
capital_cost=5,
build_year=build_year,
lifetime=10,
)
n.add('Generator', 'backup', bus='grid', carrier='gas', p_nom=100, marginal_cost=80)
n.add('Load', 'town', bus='grid', p_set=[15, 20, 15, 20])
return n
def build():
"""Both solar units under one carrier with `max_growth = 10`; the old one retires after 2020."""
return network({'solar_old': 'solar', 'solar_new': 'solar'})
Rung 52 — a transmission cost limit per scenario#
The reference script calls n.set_scenarios(...) with a
transmission_expansion_cost_limit row. The file states the row in every
scenario, as it states every global constraint. PyPSA builds one row per
scenario too, with that scenario's capital cost
(global_constraints.py:814-864, :908).
The rung adds an extendable DC link to the spine under a cost limit of 150,
so the link builds 15. The two futures are identical, and PyPSA solves to
15630.0, the objective of the network without scenarios. Without the row,
PyPSA solves to 11890.0.
| PyPSA | status | note |
|---|---|---|
transmission_expansion_cost_limit on a network with scenarios |
done | the row per scenario |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective15630.0, 86 rows.
The network, as PyPSA code
rung_52_scenario_cost_limit.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 52: a `transmission_expansion_cost_limit` row holds in every scenario.
The two futures are identical, so the expected cost is the cost of the same
network without scenarios.
"""
from __future__ import annotations
import spine
def network():
"""The spine plus an extendable DC link whose build a cost limit of 150 caps."""
n = spine.build()
n.add('Carrier', 'DC')
n.add(
'Link',
'hvdc52',
bus0='north',
bus1='south',
carrier='DC',
p_nom_extendable=True,
p_nom_max=100,
capital_cost=10,
)
n.add('Load', 'port52', bus='south', p_set=40)
n.add(
'GlobalConstraint',
'cost52',
type='transmission_expansion_cost_limit',
carrier_attribute='DC',
sense='<=',
constant=150,
)
return n
def build():
"""The same network over two identical futures."""
n = network()
n.set_scenarios({'calm': 0.6, 'stormy': 0.4})
return n
Rung 53 — a transmission volume limit per scenario and period#
The reference script calls n.set_scenarios(...) and runs
n.optimize(multi_investment_periods=True) with a
transmission_volume_expansion_limit row. The file states the row in every
scenario. PyPSA builds one row per scenario too, over the lines active in the
row's periods, with that scenario's length (global_constraints.py:836,
:882-883).
The rung is two periods with an extendable line of length 3 under a volume
limit of 60, so the line builds 20. The two futures are identical, and
PyPSA solves to 15005.0, the objective of the network without scenarios.
Without the row, PyPSA solves to 1465.0.
| PyPSA | status | note |
|---|---|---|
transmission_volume_expansion_limit on a network with scenarios and multi_investment_periods |
done | the row per scenario |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective15005.0, 70 rows.
The network, as PyPSA code
rung_53_scenario_period_volume_limit.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 53: a `transmission_volume_expansion_limit` row holds in every scenario under `multi_investment_periods`.
The two futures are identical, so the expected cost is the cost of the same
network without scenarios.
"""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def network():
"""Two periods, a cheap unit behind an extendable line whose volume a limit of 60 caps."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(2)] + [(2030, datetime(2030, 1, 1, t)) for t in range(2)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [10.0, 10.0]
n.snapshot_weightings['objective'] = [2.0, 1.5, 2.5, 2.0]
n.add('Bus', ['hill', 'town'])
n.add('Carrier', 'AC')
n.add('Generator', 'hydro53', bus='hill', p_nom=100, marginal_cost=5)
n.add('Generator', 'diesel53', bus='town', p_nom=100, marginal_cost=90)
n.add(
'Line',
'tie53',
bus0='hill',
bus1='town',
x=0.1,
carrier='AC',
length=3,
s_nom_extendable=True,
s_nom_max=100,
capital_cost=1,
build_year=2020,
lifetime=30,
)
n.add('Load', 'town_load', bus='town', p_set=[40, 50, 60, 45])
n.add(
'GlobalConstraint',
'volume53',
type='transmission_volume_expansion_limit',
carrier_attribute='AC',
sense='<=',
constant=60,
)
return n
def build():
"""The same network over two identical futures."""
n = network()
n.set_scenarios({'calm': 0.6, 'stormy': 0.4})
return n
Rung 55 — a transformer cycle per scenario#
The reference script calls n.set_scenarios(...) with two transformers in parallel, a
cycle. The file states the Kirchhoff voltage row in every scenario, and PyPSA
builds it in every scenario too (constraints.py:1738-1750). PyPSA takes a
phase shift per scenario, and the file states one for every scenario (#783).
The rung adds two transformers of reactance 0.1 and 0.2, each rated 30,
to the spine. The row splits the flow two to one, so the first caps the pair at
45. The two futures are identical, and PyPSA solves to 13105.0, the
objective of the network without scenarios. With one transformer rated 60,
PyPSA solves to 11705.0.
| PyPSA | status | note |
|---|---|---|
Kirchhoff-Voltage-Law with a transformer on a network with scenarios |
done | the row per scenario, with one phase shift for every scenario (#783) |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective13105.0, 152 rows.
The network, as PyPSA code
rung_55_scenario_transformer_cycle.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 55: a cycle of two transformers takes its Kirchhoff voltage row in every scenario.
The two futures are identical, so the expected cost is the cost of the same
network without scenarios.
"""
from __future__ import annotations
import spine
def network():
"""The spine plus two parallel transformers of unequal reactance, so the one that takes more flow caps the pair."""
n = spine.build()
n.add('Bus', ['a', 'b'])
n.add('Generator', 'hydro55', bus='a', p_nom=100, marginal_cost=10)
n.add('Generator', 'diesel55', bus='b', p_nom=100, marginal_cost=50)
n.add('Load', 'town55', bus='b', p_set=[50, 40, 55, 45])
n.add('Transformer', 't55', bus0='a', bus1='b', x=0.1, s_nom=30)
n.add('Transformer', 't55_2', bus0='a', bus1='b', x=0.2, s_nom=30)
return n
def build():
"""The same network over two identical futures."""
n = network()
n.set_scenarios({'calm': 0.6, 'stormy': 0.4})
return n
Rung 56 — a security-constrained run per scenario#
This rung runs n.optimize.optimize_security_constrained(...) on a network
with scenarios. The file states the outage copies in every scenario. PyPSA
builds them in every scenario too, with the outage factors of the first
scenario (abstract.py:610-619, :652-664).
The rung adds two parallel lines rated 30 and 35 to the spine, and outages
each. Each line must carry the whole import alone, so the import falls to 30.
The two futures are identical, and PyPSA solves to 18205.0, the objective of
the network without scenarios. With a plain n.optimize(), PyPSA solves to
11705.0.
| PyPSA | status | note |
|---|---|---|
Line-fix-s-*-security-for-{c}-outage-in-sub-network-{n} on a network with scenarios |
done | the copies per scenario |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective18205.0, 216 rows.
The network, as PyPSA code
rung_56_scenario_security_constrained.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 56: a security-constrained run over scenarios copies its rows into every scenario.
The two futures are identical, so the expected cost is the cost of the same
network without scenarios.
"""
from __future__ import annotations
import spine
BRANCH_OUTAGES = ['l56', 'l56_2']
def network():
"""The spine plus two parallel lines, so each must carry the whole import alone when the other is out."""
n = spine.build()
n.add('Bus', ['a', 'b'])
n.add('Generator', 'hydro56', bus='a', p_nom=100, marginal_cost=10)
n.add('Generator', 'diesel56', bus='b', p_nom=100, marginal_cost=50)
n.add('Load', 'town56', bus='b', p_set=[50, 40, 55, 45])
n.add('Line', 'l56', bus0='a', bus1='b', x=0.1, s_nom=30)
n.add('Line', 'l56_2', bus0='a', bus1='b', x=0.1, s_nom=35)
return n
def build():
"""The same network over two identical futures."""
n = network()
n.set_scenarios({'calm': 0.6, 'stormy': 0.4})
return n
Rung 57 — a fixed build per scenario#
The reference script calls n.set_scenarios(...) with p_nom_set on an
extendable unit.
The file states Generator-p_nom_set once, and PyPSA builds it once per
scenario (constraints.py:1915-1924). PyPSA refuses a p_nom_set that
differs by scenario (constants.py:42-44), so the rows are equal. The file
states Generator_p_nom_set and every other *_nom_set without scenario.
The rung adds a cheap extendable wind unit to the spine, pinned to 20. The
two futures are identical, and PyPSA solves to 4800.0, the objective of the
network without scenarios. Without p_nom_set, PyPSA builds the unit to 67
and solves to 335.0.
| PyPSA | status | note |
|---|---|---|
{c}-p_nom_set on a network with scenarios |
done | one row, the same feasible set as PyPSA's row per scenario |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective4800.000000000001, 86 rows.
The network, as PyPSA code
rung_57_scenario_nom_set.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 57: `p_nom_set` pins an extendable build on a network with scenarios.
The two futures are identical, so the expected cost is the cost of the same
network without scenarios.
"""
from __future__ import annotations
import spine
def network():
"""The spine plus a cheap extendable wind unit whose build is pinned below what it would choose."""
n = spine.build()
n.add(
'Generator',
'wind57',
bus='south',
p_nom_extendable=True,
p_nom_max=100,
capital_cost=5,
p_nom_set=20,
)
return n
def build():
"""The same network over two identical futures."""
n = network()
n.set_scenarios({'calm': 0.6, 'stormy': 0.4})
return n
Rung 58 — a committable unit per scenario#
The reference script calls n.set_scenarios(...) with a committable unit. The file
states the status rows in every scenario, and PyPSA builds them in every
scenario too.
The rung adds a cheap committable unit to the spine that cannot run below 40
% of its build, with a start-up cost of 100. The two futures are identical,
and PyPSA solves to 7430.0, the objective of the network without scenarios.
With the same unit not committable, PyPSA solves to 7330.0.
| PyPSA | status | note |
|---|---|---|
{c}-com-* on a network with scenarios |
done | the rows per scenario |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective7430.0, 128 rows.
The network, as PyPSA code
rung_58_scenario_committable.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 58: a committable unit takes its status rows in every scenario.
The two futures are identical, so the expected cost is the cost of the same
network without scenarios.
"""
from __future__ import annotations
import spine
def network():
"""The spine plus a cheap committable unit that cannot run below 40 % of its build."""
n = spine.build()
n.add(
'Generator',
'uc58',
bus='north',
committable=True,
p_nom=50,
marginal_cost=5,
p_min_pu=0.4,
min_up_time=2,
up_time_before=0,
start_up_cost=100,
)
n.add('Load', 'swing58', bus='north', p_set=[5, 45, 45, 10])
return n
def build():
"""The same network over two identical futures."""
n = network()
n.set_scenarios({'calm': 0.6, 'stormy': 0.4})
return n
Rung 60 — efficiencies per snapshot#
This rung runs n.optimize() with a link, a process, a storage unit, a fuel
unit and a transformer whose coefficients change over time. PyPSA declares
these coefficients static or series and reads each per snapshot:
- a link's
efficiency,efficiency2, … - a process's
rate0,rate1, … - a storage unit's
efficiency_storeandefficiency_dispatch - a generator's
efficiency - a transformer's
phase_shift
PyPSA's balance takes a port's coefficient at the snapshot the flow arrives.
It shifts the flow first and multiplies it by the coefficient after
(constraints.py:1565-1569). The energy balance of a storage unit takes both
efficiencies per snapshot (constraints.py:2167-2168). A primary_energy row
divides a generator's output by its efficiency at the same snapshot
(global_constraints.py:434-439). A fixed phase shift enters the cycle sum per
snapshot (constraints.py:1738-1744).
The file states Link_efficiency, Process_rate,
StorageUnit_efficiency_store, StorageUnit_efficiency_dispatch,
Generator_primary_energy_weight and Transformer_phase_shift_weight over
snapshot. Link_output_arrival and Process_output_arrival contain the
shifted flow times the coefficient. The file states a standing loss per
snapshot in StorageUnit_retention and Store_retention. On a plain run, the
data has the same value at each snapshot, and the rows are the standard ones.
The rung adds a sink fed from a source over a link and a process that both
deliver one snapshot late. Beside them are a storage unit and a fuel unit under a
primary_energy cap. A separate triangle has a transformer with a fixed phase
shift per snapshot. The generators weighting is uniform, as in rung 16.
PyPSA solves to 130562.09, and each coefficient changes the objective. With
the coefficient held at its mean at each snapshot, PyPSA solves to these
objectives (#620).
| coefficient held at its mean | objective |
|---|---|
link efficiency |
130774.59 |
process rate1 |
130622.09 |
storage unit efficiency_store |
129770.09 |
storage unit efficiency_dispatch |
129966.86 |
generator efficiency |
131280.86 |
transformer phase_shift |
130313.38 |
The snapshot at which a delayed port takes the coefficient also changes the
result. With each series moved one snapshot later, the port takes it at the
snapshot the flow departs. PyPSA then solves to 130676.26 for the link and to
130742.09 for the process.
| PyPSA | status | note |
|---|---|---|
Bus-nodal_balance with an efficiency or a rate per snapshot |
done | Link_efficiency and Process_rate span snapshot, taken at the snapshot a delayed flow arrives |
StorageUnit-energy_balance with efficiencies per snapshot |
done | StorageUnit_efficiency_store and StorageUnit_efficiency_dispatch span snapshot |
primary_energy with a generator efficiency per snapshot |
done | Generator_primary_energy_weight spans snapshot |
Kirchhoff-Voltage-Law with a fixed phase shift per snapshot |
done | Transformer_phase_shift_weight spans snapshot |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective130562.09418409193, 133 rows.
The network, as PyPSA code
rung_60_efficiency_per_snapshot.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 60: efficiencies per snapshot — a link, a process, a storage unit, a fuel unit under a primary-energy cap and a transformer with a fixed phase shift, each reading its coefficient per snapshot."""
from __future__ import annotations
from datetime import datetime
#: Four hourly stamps. The `generators` weighting is uniform, as rung 16's, so a
#: `delay` of one is a shift of one snapshot position. The `objective` and
#: `stores` columns stay non-uniform, so no cost or storage factor passes as identity.
SNAPSHOTS = [datetime(2015, 1, 1, hour) for hour in range(4)]
WEIGHTINGS = {'objective': [2.0, 1.5, 2.5, 3.0], 'stores': [0.5, 2.0, 1.5, 2.5], 'generators': [1.0, 1.0, 1.0, 1.0]}
LINK_EFFICIENCY = [2.0, 3.5, 2.5, 3.0]
PROCESS_RATE = [0.9, 0.5, 0.8, 0.6]
EFFICIENCY_STORE = [0.9, 0.5, 0.9, 0.5]
EFFICIENCY_DISPATCH = [0.6, 0.9, 0.7, 0.95]
FUEL_EFFICIENCY = [0.3, 0.6, 0.4, 0.5]
PHASE_SHIFT = [0.0, 0.5, -0.3, 1.0]
def build():
"""A source feeding a sink over a delayed link and a delayed process, a storage unit and a capped fuel unit at the sink, and a triangle with a fixed phase shift per snapshot."""
import pypsa
n = pypsa.Network()
n.set_snapshots(SNAPSHOTS)
for column, values in WEIGHTINGS.items():
n.snapshot_weightings[column] = values
n.add('Carrier', 'fuel60', co2_emissions=1.0)
n.add('Bus', ['source', 'sink'])
n.add('Generator', 'spring60', bus='source', p_nom=200, marginal_cost=5)
n.add('Generator', 'backup60', bus='sink', p_nom=200, marginal_cost=100)
n.add(
'Generator', 'fuel_unit60', bus='sink', carrier='fuel60', p_nom=40, marginal_cost=20, efficiency=FUEL_EFFICIENCY
)
n.add('Link', 'heat_pump60', bus0='source', bus1='sink', p_nom=10, efficiency=LINK_EFFICIENCY, delay=1)
n.add('Process', 'boiler60', bus0='source', bus1='sink', p_nom=20, rate1=PROCESS_RATE, delay1=1)
n.add(
'StorageUnit',
'battery60',
bus='sink',
p_nom=20,
max_hours=2,
cyclic_state_of_charge=True,
efficiency_store=EFFICIENCY_STORE,
efficiency_dispatch=EFFICIENCY_DISPATCH,
)
n.add('Load', 'sink_load60', bus='sink', p_set=[60.0, 90.0, 50.0, 100.0])
n.add(
'GlobalConstraint',
'fuel_cap60',
type='primary_energy',
carrier_attribute='co2_emissions',
sense='<=',
constant=150,
)
n.add('Bus', ['a', 'b', 'c'])
n.add('Generator', 'hydro60', bus='a', p_nom=300, marginal_cost=10)
n.add('Generator', 'diesel60', bus='c', p_nom=300, marginal_cost=200)
n.add('Load', 'town60', bus='c', p_set=[90.0, 75.0, 120.0, 105.0])
n.add('Line', 'ab60', bus0='a', bus1='b', carrier='AC', x=0.002, r=0.0002, s_nom=120)
n.add('Line', 'bc60', bus0='b', bus1='c', carrier='AC', x=0.002, r=0.0002, s_nom=120)
n.add('Transformer', 'ca60', bus0='c', bus1='a', x=0.002, r=0.0002, s_nom=40, phase_shift=PHASE_SHIFT)
return n
Rung 61 — voltage angle limits#
This rung runs n.optimize() with a cap on the voltage angle difference across
a line or a transformer. PyPSA master adds v_ang_max, in degrees, on Line and
Transformer (PyPSA#1910). PyPSA builds two rows for each active AC line and
transformer with a finite v_ang_max (constraints.py:1764-1881, called at
optimize.py:951). The rows bound the angle difference by the cap either
way.
The angle difference is x_pu_eff times the flow, plus the shift in radians
for a transformer. PyPSA divides through by x_pu_eff, so the flow takes
coefficient one. A fixed shift moves the bounds: Transformer-v_ang-lower and
Transformer-v_ang-upper. A decided shift is a variable of the row:
Transformer-v_ang-var-lower and Transformer-v_ang-var-upper. PyPSA ignores
v_ang_min and warns that it is deprecated, so the file has no parameter for
it.
This rung has no scenarios, and no rung states the angle rows on a network with scenarios. PyPSA takes the phase shift per scenario, and the file states one for every scenario (#783).
The rung adds a line with a cap of one degree beside the link of the spine. The lossy link carries the rest of the flow south. Two triangles each carry cheap upstream power to a town past a costly local unit. In the first, the transformer has a fixed shift of two degrees. In the second, the shift is a decision. In one cycle, the angle caps of all branches couple, so each triangle has one cap.
PyPSA solves to 121834.62, and each cap is tight at the optimum. With a cap
removed, PyPSA solves to these objectives (#620).
| cap removed | objective |
|---|---|
line north_south61 |
121161.89 |
transformer ca61, fixed shift |
35214.37 |
transformer fd61, decided shift |
111382.98 |
| all three | 24090.0 |
| PyPSA | status | note |
|---|---|---|
Line-v_ang-lower, Line-v_ang-upper |
done | Line_carrier == 'AC' in the where, and no row for an infinite Line_v_ang_max |
Transformer-v_ang-lower, Transformer-v_ang-upper |
done | the fixed shift is Transformer_phase_shift_fixed, per snapshot |
Transformer-v_ang-var-lower, Transformer-v_ang-var-upper |
done | the decided shift is the Transformer_phase_shift variable, in degrees |
{c}-v_ang-* on a network with scenarios |
done | the rows per scenario, with one phase shift for every scenario (#783) |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective121834.62083205313, 176 rows.
The network, as PyPSA code
rung_61_voltage_angle_limits.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 61: voltage angle limits — a line, a transformer with a fixed phase shift and one with a decided shift, each holding the angle difference across it within its `v_ang_max`.
A line beside the spine's link may carry only the flow that keeps the angle
across it within one degree, so the lossy link carries the rest. Two triangles
carry cheap upstream power to a town past a costly local unit. In the first a
transformer has a fixed shift, in the second the shift is a decision; the
angle cap on each transformer binds before its thermal rating does, so the
local unit runs. In one cycle the angle caps of all branches couple, so each
triangle carries one cap only. No scenario: with scenarios PyPSA builds no
angle row.
"""
from __future__ import annotations
import spine
def build():
"""The spine plus a line beside its link and two triangles, each with one voltage angle limit that binds, as a ``pypsa.Network``."""
n = spine.build()
n.add('Line', 'north_south61', bus0='north', bus1='south', carrier='AC', x=0.002, s_nom=100, v_ang_max=1)
n.add('Bus', ['a', 'b', 'c'])
n.add('Generator', 'hydro61', bus='a', p_nom=300, marginal_cost=10)
n.add('Generator', 'diesel61', bus='c', p_nom=300, marginal_cost=200)
n.add('Load', 'town61', bus='c', p_set=[90, 75, 120, 105])
n.add('Line', 'ab61', bus0='a', bus1='b', carrier='AC', x=0.002, s_nom=200)
n.add('Line', 'bc61', bus0='b', bus1='c', carrier='AC', x=0.002, s_nom=200)
n.add('Transformer', 'ca61', bus0='c', bus1='a', x=0.4, s_nom=200, phase_shift=2, v_ang_max=2.5)
n.add('Bus', ['d', 'e', 'f'])
n.add('Generator', 'hydro61_shift', bus='d', p_nom=300, marginal_cost=10)
n.add('Generator', 'diesel61_shift', bus='f', p_nom=300, marginal_cost=200)
n.add('Load', 'town61_shift', bus='f', p_set=[90, 75, 120, 105])
n.add('Line', 'de61', bus0='d', bus1='e', carrier='AC', x=0.002, s_nom=200)
n.add('Line', 'ef61', bus0='e', bus1='f', carrier='AC', x=0.002, s_nom=200)
n.add(
'Transformer',
'fd61',
bus0='f',
bus1='d',
x=0.4,
s_nom=80,
phase_shift_min=-30,
phase_shift_max=30,
v_ang_max=5,
)
return n
Rung 62 — a start-up cost per snapshot#
This rung runs n.optimize() with a committable unit whose start-up and
shut-down costs change over time. PyPSA declares the start_up_cost and
shut_down_cost of a Generator, a Link and a Process as static or series
(PyPSA/PyPSA#1909). It takes the cost of a start or a stop at its snapshot
(optimize.py:457-468). It still adds the cost without the weight of the
snapshot and without the weight of the period, as rung 48 states.
The file states the same, with Generator_start_up_cost,
Generator_shut_down_cost and the Link and Process pairs over snapshot. A
static cost is the same value at each snapshot.
The rung adds a committable peaker that the load needs in the third and
fourth snapshots. A start costs 100 in the second snapshot and 900 in the
third, and a stop is free only in the last snapshot. PyPSA solves to 6850.0:
the unit starts one snapshot early at minimum output and stops in the last
snapshot. Each series changes the objective (#807).
| cost changed | objective |
|---|---|
start-up cost held at its mean, 640.0 |
7090.0 |
shut-down cost held at its mean, 240.0 |
7050.0 |
| start-up cost times the snapshot weight | 6900.0 |
| PyPSA | status | note |
|---|---|---|
start_up_cost, shut_down_cost per snapshot |
done | Generator, Link and Process over [scenario, snapshot, …], with no snapshot weight, as rung 48 |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective6850.0, 60 rows.
The network, as PyPSA code
rung_62_startup_cost_per_snapshot.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 62: a start-up cost per snapshot — a committable unit starts in the snapshot where a start costs least, with no snapshot weight on that cost."""
from __future__ import annotations
from datetime import datetime
SNAPSHOTS = [datetime(2015, 1, 1, hour) for hour in range(5)]
WEIGHTINGS = [2.0, 1.5, 1.0, 2.0, 1.0]
START_UP_COST = [400.0, 100.0, 900.0, 900.0, 900.0]
SHUT_DOWN_COST = [300.0, 300.0, 300.0, 300.0, 0.0]
def build():
"""A whole network, not the spine: a committable peaker the load needs in the third and fourth snapshots, cheap to start one snapshot early and free to stop in the last."""
import pypsa
n = pypsa.Network()
n.set_snapshots(SNAPSHOTS)
n.snapshot_weightings['objective'] = WEIGHTINGS
n.add('Bus', 'grid')
n.add('Generator', 'base62', bus='grid', p_nom=60, marginal_cost=10)
n.add('Generator', 'dear62', bus='grid', p_nom=100, marginal_cost=90)
n.add(
'Generator',
'peak62',
bus='grid',
p_nom=50,
marginal_cost=20,
committable=True,
p_min_pu=0.4,
up_time_before=0,
)
n.generators_t.start_up_cost['peak62'] = START_UP_COST
n.generators_t.shut_down_cost['peak62'] = SHUT_DOWN_COST
n.add('Load', 'town62', bus='grid', p_set=[50, 50, 100, 100, 50])
return n
Rung 63 — modular branches and storage#
PyPSA builds {c}-n_mod and the {c}-{attr}_modularity row for every
component with a nominal build (variables.py:371-392,
constraints.py:1927-1987, called at optimize.py:871-874 and :901-905):
- Generator, Link and Process on
p_nom - Line and Transformer on
s_nom - StorageUnit on
p_nom - Store on
e_nom
PyPSA builds each only for a unit that is extendable, modular and among
c.active_assets (variables.py:387, constraints.py:1967). These are the
units whose static active flag is true
(components/descriptors.py:151-187). The file states the pairs of Line,
Transformer, StorageUnit and Store. In the file, a unit is among
active_assets where it is active in at least one snapshot.
PyPSA reads the static active flag, so it also builds both for a unit whose
build year and lifetime miss every period. No other PyPSA row contains
{c}-n_mod. PyPSA builds no {c}-{attr} column for a unit that is not active
either. The file declares {c}_{attr}_ext there, priced at zero by
{c}_capital_weight.
The rung feeds an east bus over a modular line and a modular transformer, with
a modular storage unit and store on it. Beside them are a modular extendable
generator, link and process with active = False, for which PyPSA builds no
n_mod. PyPSA solves to 13927.0, and each module size changes the
objective. With the module size of that component set to zero, PyPSA solves to
these objectives (#815).
| module size set to zero | objective |
|---|---|
line s_nom_mod |
13882.0 |
transformer s_nom_mod |
13921.0 |
storage unit p_nom_mod |
13877.0 |
store e_nom_mod |
13873.0 |
| PyPSA | status | note |
|---|---|---|
Line-n_mod, Line-s_nom_modularity |
done | |
Transformer-n_mod, Transformer-s_nom_modularity |
done | |
StorageUnit-n_mod, StorageUnit-p_nom_modularity |
done | |
Store-n_mod, Store-e_nom_modularity |
done | |
{c}-n_mod, {c}-{attr}_modularity for a unit that is not active |
done | no column and no row where count({c}_active, over=snapshot) > 0 fails |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective13927.0, 116 rows.
The network, as PyPSA code
rung_63_modular_branches_and_storage.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 63: modular branches and storage — a line, a transformer, a storage unit and a store each built in whole modules, beside a modular generator, link and process that are not active."""
from __future__ import annotations
import spine
def build():
"""The spine plus an east bus fed over a modular line and a modular transformer, with a modular storage unit and store, and three inactive modular builds."""
n = spine.build()
n.add('Bus', 'east')
n.add(
'Line', 'ne63', bus0='north', bus1='east', x=0.01, r=0.001, s_nom_extendable=True, s_nom_mod=25, capital_cost=3
)
n.add('Transformer', 'se63', bus0='south', bus1='east', x=0.01, s_nom_extendable=True, s_nom_mod=15, capital_cost=2)
n.add(
'StorageUnit',
'battery63',
bus='east',
p_nom_extendable=True,
p_nom_mod=15,
max_hours=2,
capital_cost=5,
cyclic_state_of_charge=True,
)
n.add('Store', 'tank63', bus='east', e_nom_extendable=True, e_nom_mod=40, capital_cost=2, e_cyclic=True)
n.add('Generator', 'east_backup63', bus='east', p_nom=200, marginal_cost=200)
n.add('Load', 'east_load63', bus='east', p_set=[20, 70, 110, 40])
idle = {'p_nom_extendable': True, 'p_nom_mod': 20, 'capital_cost': 1, 'marginal_cost': 1, 'active': False}
n.add('Generator', 'idle_gen63', bus='east', **idle)
n.add('Link', 'idle_link63', bus0='north', bus1='east', **idle)
n.add('Process', 'idle_proc63', bus0='south', bus1='east', rate0=-1.25, **idle)
return n
Rung 64 — commitment edges#
This rung runs n.optimize(multi_investment_periods=True) with a committable
generator, link and process built in the later period, and a committable unit
with no build cap. PyPSA counts a start against the previous snapshot's status
and reads a status it did not build as zero (constraints.py:391). So a unit
that opens in a later period starts from off at its first snapshot, and pays
its start-up cost there. The file states this zero status in
Generator_previous_status and its Link and Process siblings. Without the
zero, the previous status at that first snapshot is absent. The start-up row is
then not built, and the unit starts for free.
A committed extendable unit has a big-M bound in its com-ext-p-* rows. Where
p_nom_max is finite and positive, M is that cap times the highest
p_max_pu. Where the cap is infinite, PyPSA takes the committable_big_m
keyword (components.py:1082-1152). Without the keyword, M is ten times the
largest of the peak total load and the component's largest finite p_nom and
p_nom_max. Generator_big_m is M by the same rule. An M below the output the
solver chooses caps that output.
PyPSA solves to 9140.0. The three units start once each, at the first
snapshot of the 2030 period, and their start-up costs take that period's weight
of 0.5. With their start-up rows dropped there, PyPSA solves to 9035.0, the
three weighted start-up costs less. The uncapped unit builds 70, below its M
of 3000. With committable_big_m=50, PyPSA solves to 11640.0, and the unit
builds 50 (#817, #975).
| PyPSA | status | note |
|---|---|---|
{c}-com-transition-start-up for a unit built in a later period |
done | Generator_previous_status is zero at the first snapshot a unit is active in past the first of the horizon |
{c}-com-ext-p-* (big-M) with p_nom_max = inf |
done | Generator_big_m is M by PyPSA's rule |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective9140.0, 205 rows.
The network, as PyPSA code
rung_64_commitment_edges.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 64: commitment edges — a committable generator, link and process built in the later period start up when they open, and a committable unit without a build cap is released by an inferred big M."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def build():
"""A whole network, not the spine: eight snapshots over two periods, three committable units built in 2030, an uncapped committable build, and a dear peaker."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(4)] + [(2030, datetime(2030, 1, 1, t)) for t in range(4)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [10.0, 10.0]
n.snapshot_weightings['objective'] = [2.0, 1.5, 2.5, 2.0, 2.0, 1.5, 2.5, 2.0]
n.add('Bus', ['hub', 'fuel'])
n.add('Generator', 'well64', bus='fuel', p_nom=200, marginal_cost=1)
n.add('Generator', 'peak64', bus='hub', p_nom=300, marginal_cost=50)
late = {'committable': True, 'p_nom': 30, 'p_min_pu': 0.25, 'build_year': 2030, 'lifetime': 30}
n.add('Generator', 'late_gen64', bus='hub', marginal_cost=2, start_up_cost=100, **late)
n.add('Link', 'late_link64', bus0='fuel', bus1='hub', marginal_cost=1, start_up_cost=70, **late)
n.add('Process', 'late_proc64', bus0='fuel', bus1='hub', rate0=-1.25, start_up_cost=40, **late)
n.add(
'Generator',
'uncapped64',
bus='hub',
committable=True,
p_nom_extendable=True,
capital_cost=20,
marginal_cost=10,
p_min_pu=0.2,
up_time_before=0,
start_up_cost=30,
)
n.add('Load', 'hub_load64', bus='hub', p_set=[40, 60, 70, 50, 120, 140, 150, 130])
return n
Rung 66 — cycles per period#
This rung runs
n.optimize.optimize_security_constrained(multi_investment_periods=True) over
a network whose meshes change between periods. PyPSA builds the cycle basis of
each period from the branches that are active in it. It writes the cycle rows of
each period over that period's snapshots only (constraints.py:1710-1729). A
cycle label is a position in its period's basis, so label 0 may name a
different cycle in each period.
The file states the cycle weights Line_cycle_weight,
Transformer_cycle_weight and Transformer_phase_shift_cycle_weight over
period. Kirchhoff_Voltage_Law contains, at each snapshot, the weights of
that snapshot's period. A cycle label that the basis of a period does not reach
has no weight there, and its row is not built. On a plain run, the data has
one period, and the rows are the standard ones.
The rung joins a and c over ab66 and bc66, which form a path in the
2020 period. The line ca66 comes in with 2030 and closes a triangle, so 2030
has one cycle and 2020 has none. PyPSA builds 2 cycle rows and solves to
16000.0. A single basis over all branches would apply the triangle's row to
the 2020 snapshots, with ca66 absent. That row and the balance at b force
the flow on ab66 and bc66 to zero. Without those two lines in 2020, PyPSA
solves to 43900.0 (#814).
The run outages ca66, which is active in 2030 only. In 2020 ca66 is not
active, so the file states no copy in 2020 and PyPSA builds none
(abstract.py:653-664). PyPSA builds the 6 copies of 2030 for the lower and
the upper limit, and the file states the same 6.
| PyPSA | status | note |
|---|---|---|
Kirchhoff-Voltage-Law per investment period |
done | the cycle weights span period, read at each snapshot's period |
Line-fix-s-*-security-… for a branch not active in a period |
done | no factor and no copy |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective16000.0, 62 rows.
The network, as PyPSA code
rung_66_cycles_per_period.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 66: cycles per period — a line that comes in with the second period closes a triangle, under a security-constrained run that outages it."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
BRANCH_OUTAGES = ['ca66']
def build():
"""A whole network, not the spine: a path from `a` to `c` in 2020, and a triangle in 2030 once `ca66` stands."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(2)] + [(2030, datetime(2030, 1, 1, t)) for t in range(2)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [10.0, 10.0]
n.snapshot_weightings['objective'] = [2.0, 1.5, 2.5, 3.0]
n.add('Bus', ['a', 'b', 'c'])
n.add('Generator', 'hydro66', bus='a', p_nom=300, marginal_cost=10)
n.add('Generator', 'diesel66', bus='c', p_nom=300, marginal_cost=100)
n.add('Load', 'town66', bus='c', p_set=[80, 100, 90, 110])
n.add('Line', 'ab66', bus0='a', bus1='b', x=0.1, s_nom=150)
n.add('Line', 'bc66', bus0='b', bus1='c', x=0.1, s_nom=150)
n.add('Line', 'ca66', bus0='c', bus1='a', x=0.1, s_nom=40, build_year=2030, lifetime=50)
return n
Rung 68 — outage factors per period#
This rung runs
n.optimize.optimize_security_constrained(multi_investment_periods=True) with
an outage of a line that retires before the last period. Line_BODF and
Transformer_BODF are the outage factors of each period, from the branches
active in it. The file states the copies of a period over its snapshots only.
PyPSA does the same. It finds the sub-networks of each period's active
branches and names the copies of the period …-period-{p}
(abstract.py:653-664). It refuses an outage named in branch_outages that
is active in no period (abstract.py:634-643).
The rung joins a and b over two lines rated 60. ab_old68 is active in
2020 and retires in 2025, and ab68 is active in both periods. The run outages
ab_old68. In 2020 its factor on ab68 is 1, so ab68 carries the whole
import alone after the outage, and the import is at most 60. In 2030
ab_old68 is not active and has no copy.
PyPSA solves to 13000.0. Without the 2020 copies, the import is up to 120
in 2020, and PyPSA solves to 8050.0.
| PyPSA | status | note |
|---|---|---|
Line-fix-s-*-security-… for an outage of a branch not active in the last period |
done | Line_BODF and Transformer_BODF span period |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective13000.0, 46 rows.
The network, as PyPSA code
rung_68_outage_factors_per_period.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 68: outage factors per period — a security-constrained run outages a line that retires before the last period.
Each period takes the outage factors of the branches active in it, so the
outage of the retired line binds in 2020 and builds no rows in 2030.
"""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
BRANCH_OUTAGES = ['ab_old68']
PERIODS = {2020: (1.0, [2.0, 1.5], [80, 70]), 2030: (0.5, [2.5, 3.0], [90, 60])}
def network(periods: list[int]):
"""Cheap hydro at `a`, a town with diesel at `b`, and two parallel lines; `ab_old68` retires in 2025."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples([(p, datetime(p, 1, 1, t)) for p in periods for t in range(2)])
n.investment_periods = periods
n.investment_period_weightings['objective'] = [PERIODS[p][0] for p in periods]
n.investment_period_weightings['years'] = 10.0
n.snapshot_weightings['objective'] = [w for p in periods for w in PERIODS[p][1]]
n.add('Bus', ['a', 'b'])
n.add('Generator', 'hydro68', bus='a', p_nom=300, marginal_cost=10)
n.add('Generator', 'diesel68', bus='b', p_nom=300, marginal_cost=100)
n.add('Load', 'town68', bus='b', p_set=[load for p in periods for load in PERIODS[p][2]])
n.add('Line', 'ab68', bus0='a', bus1='b', x=0.1, s_nom=60)
n.add('Line', 'ab_old68', bus0='a', bus1='b', x=0.1, s_nom=60, build_year=2000, lifetime=25)
return n
def build():
"""Both periods: after the outage of `ab_old68`, `ab68` carries the whole import of 2020 alone."""
return network(list(PERIODS))
Rung 69 — storage that retires before the close#
This rung runs n.optimize(multi_investment_periods=True) with a
primary_energy row over storage that carries its level across periods and
retires before the last counted snapshot. PyPSA fills the level forward over
the counted snapshots and takes the last value
(global_constraints.py:476-479, :530-533, and :686-689, :736-739 for
operational_limit). The level exists only where the storage is active.
PyPSA's row therefore contains the level at the last counted snapshot where the
storage is active.
The file states that snapshot as
StorageUnit_last_counted_active and
Store_last_counted_active. The counted
snapshots are the whole horizon or one period, and a storage is active in one
continuous stretch of periods. So the snapshot is the last one where the two
overlap.
The rung adds a store and a storage unit of an emitting carrier. Each has an
initial level of 20 and retires after 2020. A CO2 cap of 5 applies over the
horizon. The row bounds the two levels at the last snapshot of 2020 to at least
35. A row at the last snapshot of 2030, where neither storage is active,
would have no variable and would be 0 <= -35.
| PyPSA | status | note |
|---|---|---|
primary_energy, operational_limit |
done | for storage that retires before the last counted snapshot, StorageUnit_last_counted_active and Store_last_counted_active, the last counted snapshot where the storage is active |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective350.0, 37 rows.
The network, as PyPSA code
rung_69_storage_retires_before_close.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 69: storage that retires before the last counted snapshot closes a `primary_energy` row at the last level it holds."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def build():
"""A whole network, not the spine: four snapshots over two periods, a store and a storage unit of an emitting carrier that retire after 2020, and a CO2 cap over the horizon."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(2)] + [(2030, datetime(2030, 1, 1, t)) for t in range(2)]
)
n.investment_periods = [2020, 2030]
n.add('Carrier', 'gas', co2_emissions=1.0)
n.add('Bus', 'hub')
n.add('Generator', 'base69', bus='hub', p_nom=100, marginal_cost=10)
n.add('Load', 'hub_load', bus='hub', p_set=10)
n.add('Store', 'e_retire', bus='hub', carrier='gas', e_nom=30, e_initial=20, build_year=2020, lifetime=10)
n.add(
'StorageUnit',
'su_retire',
bus='hub',
carrier='gas',
p_nom=10,
max_hours=3,
state_of_charge_initial=20,
build_year=2020,
lifetime=10,
)
n.add(
'GlobalConstraint', 'co2_cap', type='primary_energy', carrier_attribute='co2_emissions', sense='<=', constant=5
)
return n
Rung 70 — a risk preference of weight zero#
This rung is rung 14 with n.set_risk_preference(alpha=0.5, omega=0.0).
Under any risk preference, PyPSA builds CVaR-a, CVaR-theta, CVaR,
CVaR-excess-{s} and CVaR-def (optimize.py:497, variables.py:508), and
refuses a quadratic cost (optimize.py:506-513). Without a risk preference,
it builds none of them. With and without the risk preference, PyPSA solves to
7946.733333333334, and the risk preference adds 3 rows (#849).
In the file, the columns, the rows and the assumption against a quadratic cost
exist where CVaR_omega has a value. A value of 0.0 counts. CVaR_omega
has a value only where a risk preference is set. The columns take
missing: 0, so the objective's cost CVaR_omega * CVaR exists without them.
| PyPSA | status | note |
|---|---|---|
CVaR-a, CVaR-theta, CVaR at omega = 0 |
done | where CVaR_omega has a value |
CVaR-excess-{s}, CVaR-def at omega = 0 |
done | where CVaR_omega has a value |
objective at omega = 0 |
refused | a quadratic cost, Refusals |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective7946.733333333334, 87 rows.
The network, as PyPSA code
rung_70_risk_preference_zero.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 70: rung 14 with a risk preference of weight zero — PyPSA builds the CVaR columns and rows all the same."""
from __future__ import annotations
import spine
def build():
"""Rung 14's calm and stormy futures, with `omega = 0`: the optimum is the risk-neutral one."""
n = spine.build()
n.add('Generator', 'wind70', bus='south', p_nom_extendable=True, p_nom_max=100, marginal_cost=1, capital_cost=20)
n.add('Load', 'port70', bus='south')
n.set_scenarios({'calm': 0.6, 'stormy': 0.4})
n.c.loads.dynamic.p_set[('calm', 'port70')] = [10, 20, 15, 10]
n.c.loads.dynamic.p_set[('stormy', 'port70')] = [40, 60, 50, 30]
n.c.generators.dynamic.p_max_pu[('calm', 'wind70')] = [0.9, 0.7, 0.8, 0.6]
n.c.generators.dynamic.p_max_pu[('stormy', 'wind70')] = [0.3, 0.2, 0.4, 0.1]
n.set_risk_preference(alpha=0.5, omega=0.0)
return n
Rung 71 — a unit that is not committable comes in running#
This rung runs n.optimize() with a unit that is not committable and has
up_time_before = 0, a p_init and a ramp limit, beside a committable unit.
PyPSA reads the status brought in as one for every unit that is not
committable (constraints.py:323). The unit ramps from its p_init into the
first snapshot, and its ramp row gets no start-up term
(constraints.py:1141-1147). Before PyPSA/PyPSA#1973, the same unit came in
off, and its first-snapshot ramp depended on whether another unit was
committable (PyPSA/PyPSA#1943).
The file states Generator_status_initial as one for a unit that is not
committable, so Generator_previous_p opens on its p_init.
The cheap unit ramps from 10 by 10 a snapshot, and its ramp row is tight at
the optimum. PyPSA solves to 4836.1111. With up_time_before = 1, or without
the committable unit, it solves to the same objective. Without the p_init,
it solves to 4500.0 (#977).
| PyPSA | status | note |
|---|---|---|
Generator-p-ramp_limit_up at the first snapshot, for a unit that is not committable |
done | Generator_status_initial is one, whatever up_time_before |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective4836.111111111111, 72 rows.
The network, as PyPSA code
rung_71_running_unless_committable.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 71: a unit that is not committable comes in running — its first-snapshot ramp starts from its `p_init`, whatever its `up_time_before` and whether another unit is committable."""
from __future__ import annotations
import spine
def build():
"""The spine plus a cheap ramp-limited unit that is not committable, with `up_time_before=0` and a `p_init`, beside a committable one."""
n = spine.build()
n.add(
'Generator',
'flex71',
bus='south',
p_nom=40,
marginal_cost=5,
ramp_limit_up=0.25,
ramp_limit_start_up=0.5,
up_time_before=0,
p_init=10,
)
n.add('Generator', 'unit71', bus='south', committable=True, p_nom=20, p_min_pu=0.5, marginal_cost=20)
return n
Rung 72 — modular down time#
This rung runs n.optimize() with a fixed and an extendable committable
modular generator, each with a minimum down time. The status of a modular unit
counts the modules that are on. PyPSA bounds its down-time row by the module
count: p_nom / p_nom_mod for a fixed build, and n_mod for an extendable one
(constraints.py:656-683). Before PyPSA/PyPSA#2051, the bound was one, so a
modular unit with a minimum down time ran one module at most.
The file states the bound as Generator_modules, which is Generator_n_mod
for an extendable modular build and Generator_modules_installed otherwise.
Both units come in off, and the load dips below what one module can run at its
minimum part load. A fixed modular unit that came in running is the open case
PyPSA/PyPSA#2114: the start-up transition and this row count the status it
brought in as one module. Modules that stop at the dip stay off for two snapshots.
PyPSA solves to 17190.0. Without the minimum down times, it solves to
9390.0. At 51986084, before PyPSA/PyPSA#2051, it solved to 43495.0
(#977).
| PyPSA | status | note |
|---|---|---|
Generator-com-down-time, modular |
done | bound by Generator_modules, not by one |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective17190.0, 111 rows.
The network, as PyPSA code
rung_72_modular_down_time.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 72: modular down time — a committable modular unit that stops some modules keeps the others running, because its down-time row is bound by its module count, not by one."""
from __future__ import annotations
import spine
def build():
"""The spine plus a dip bus served by a fixed and an extendable committable modular unit, each with a minimum down time, and a dear backup."""
n = spine.build()
n.add('Bus', 'dip')
common = {
'committable': True,
'p_nom_mod': 10,
'p_min_pu': 0.8,
'min_down_time': 2,
'up_time_before': 0,
'start_up_cost': 5,
}
n.add('Generator', 'fixed72', bus='dip', p_nom=40, marginal_cost=2, **common)
n.add(
'Generator',
'ext72',
bus='dip',
p_nom_extendable=True,
p_nom_max=30,
capital_cost=1,
marginal_cost=3,
**common,
)
n.add('Generator', 'dip_backup', bus='dip', p_nom=200, marginal_cost=100)
n.add('Load', 'dip_load', bus='dip', p_set=[60, 15, 60, 70])
return n
Rung 73 — the ramp of a fixed modular unit#
This rung runs n.optimize() with a fixed committable modular generator that
has ramp limits and came in running at a p_init. Each module that is on
ramps by its own share, ramp_limit_up * p_nom_mod
(constraints.py:1131-1134). At the first snapshot, PyPSA counts every
installed module as on, since the status brought in is on or off
(constraints.py:1152-1157). The start-up transition and the down-time row
count the same status as one module (PyPSA/PyPSA#2114). This rung has no
start-up cost and no minimum down time, so neither of them binds. Before PyPSA/PyPSA#2055, the share was
ramp_limit_up * p_nom, so a fixed modular unit ramped as many times faster
as it had modules (PyPSA/PyPSA#1985).
The file states the share against Generator_p_nom_committed, one module for
any modular build. Generator_ramp_previous_status counts every installed
module at the first snapshot.
The unit has three modules of 20 and ramps by 10 per module that is on.
Its first-snapshot ramp from 15 to 45 and the ramps that follow are tight
at the optimum. PyPSA solves to 12300.0. Without the p_init, it solves to
11320.0, and without the ramp limits, to 8135.0. At 51986084, before
PyPSA/PyPSA#2055, it solved to 8135.0 (#977).
| PyPSA | status | note |
|---|---|---|
Generator-p-ramp_limit_up, -down, fixed modular committable |
done | per module, and every installed module on at the first snapshot after a running start |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective12300.0, 80 rows.
The network, as PyPSA code
rung_73_fixed_modular_ramp.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 73: the ramp of a fixed modular committable unit — each running module ramps by its own share, and a unit that came in running counts all its modules on at the first snapshot."""
from __future__ import annotations
import spine
def build():
"""The spine plus a step bus served by a fixed committable modular unit that came in running at a `p_init`, with a dear backup."""
n = spine.build()
n.add('Bus', 'step')
n.add(
'Generator',
'modramp73',
bus='step',
committable=True,
p_nom=60,
p_nom_mod=20,
p_min_pu=0.2,
marginal_cost=2,
ramp_limit_up=0.5,
ramp_limit_down=0.5,
ramp_limit_start_up=0.5,
ramp_limit_shut_down=0.5,
p_init=15,
)
n.add('Generator', 'step_backup', bus='step', p_nom=200, marginal_cost=100)
n.add('Load', 'step_load', bus='step', p_set=[60, 20, 55, 30])
return n
Rung 75 — a primary-energy cap for one period over storage that reopens per period#
This rung runs n.optimize(multi_investment_periods=True) with a
primary_energy row for 2030 alone, over a gas storage unit and a gas store
that each reopen per period. PyPSA counts the storage change of the stated
period only: its opening level less its closing level at the period's last
snapshot, weighted by that period's years (global_constraints.py:390-394,
:489-495, :543-549). Before PyPSA/PyPSA#2099, PyPSA took the last snapshot
of every period and raised KeyError (PyPSA/PyPSA#2093).
The file states the closing level as StorageUnit_closing_weight and
Store_closing_weight, which are the period's years at the last snapshot of
each counted period.
The cap of 600 binds, and the row counts 10 years of each storage's
opening level against it. PyPSA solves to 12900.0. Without the cap, it solves
to 9000.0, and with the cap over the horizon, to 16500.0. With the storage
on a carrier that emits nothing, it solves to 11400.0 (#977).
| PyPSA | status | note |
|---|---|---|
primary_energy for one period over storage that reopens per period, no scenarios |
done | StorageUnit_closing_weight and Store_closing_weight in the stated period |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective12900.0, 137 rows.
The network, as PyPSA code
rung_75_period_primary_energy_storage.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 75: a primary-energy cap for one period over storage that reopens per period — the cap counts the storage change of its own period alone."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def build():
"""A whole network, not the spine: eight snapshots over two periods of unequal years, a gas unit, and a gas storage unit and store that each reopen per period."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(4)] + [(2030, datetime(2030, 1, 1, t)) for t in range(4)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [5.0, 10.0]
n.add('Carrier', 'gas75', co2_emissions=0.5)
n.add('Bus', 'hub')
n.add('Generator', 'gas75', bus='hub', carrier='gas75', p_nom=100, marginal_cost=30)
n.add('Generator', 'clean75', bus='hub', p_nom=200, marginal_cost=60)
n.add(
'StorageUnit',
'tank75',
bus='hub',
carrier='gas75',
p_nom=20,
max_hours=4,
state_of_charge_initial=60,
state_of_charge_initial_per_period=True,
)
n.add('Store', 'pile75', bus='hub', carrier='gas75', e_nom=60, e_initial=40, e_initial_per_period=True)
n.add('Load', 'town75', bus='hub', p_set=[60, 80, 70, 50, 90, 110, 100, 80])
n.add(
'GlobalConstraint',
'co2_2030',
type='primary_energy',
carrier_attribute='co2_emissions',
sense='<=',
constant=600,
investment_period=2030,
)
return n
Rung 76 — a store that covers a storage unit#
This rung runs n.optimize() with four stores that use what PyPSA/PyPSA#1934
gave the Store. PyPSA deprecates the StorageUnit in favour of the Store, and
still builds every StorageUnit row.
- A store has a charging variable
Store-p_storebeside its net dispatchStore-p, sop + p_storeis the discharge. Every store has the variable. Where a store does not split its dispatch, PyPSA holdsp_storeat zero by its bound rows, and the charging bound applies top(stores.py:79-116). A store splits where an efficiency differs from one, a cost applies to one direction only, orcycles_maxis finite. - A finite
max_hoursgives the store a power ratinge_nom / max_hours. PyPSA bounds the discharge byp_max_puand the charge by-p_min_putimes that rating (constraints.py:142-144,:239-241). The bound rows areStore-fix-p-*,Store-fix-p_store-*and theirextsiblings. Withoutmax_hours, the rating is infinite and PyPSA builds no row on that side. - The energy balance draws the discharge down before
efficiency_dispatch, adds the charge afterefficiency_store, and addsinflowlessspill(constraints.py:2472-2476,:2560-2567). marginal_cost_dispatchprices the discharge,marginal_cost_storethe charge, andspill_costthe spill.marginal_costkeeps pricing the net dispatch (optimize.py:108-120).
The file states the bounds per unit of e_nom as Store_discharge_floor_pu
and Store_charge_cap_pu, and whether a store splits as
Store_split_dispatch, from data prep. With every new attribute at its
default, the new rows hold p_store at zero and the balance is the old one.
The rung adds to the south bus a rated battery with an efficiency each way and
a discharge cost, a reservoir with inflow that spills at its rating, and an
extendable lossy store with a charging cost. An unrated extendable store sits
on the north bus. PyPSA solves to 11362.9444. The variants solve to these
objectives (#977):
| network | objective |
|---|---|
the battery without max_hours |
11343.8333 |
the battery's efficiencies at 1 |
11174.2333 |
the battery's discharge cost at 0 |
11317.3444 |
| the reservoir without inflow | 12523.7796 |
the reservoir's spill cost at 0 |
11329.1944 |
the reservoir without max_hours |
8538.7333 |
the lossy store's charging cost at 0 |
11332.0833 |
the lossy store's efficiency at 1 |
11043.0167 |
| PyPSA | status | note |
|---|---|---|
Store-p_store |
done | every store, held at zero where it does not split its dispatch |
Store-spill |
done | where: inflow > 0, missing: 0, with bounds on the variable, as StorageUnit-spill |
Store-fix-p-*, Store-ext-p-* |
done | on p + p_store, where max_hours has a value, and the floor of zero for a store that splits |
Store-fix-p_store-*, Store-ext-p_store-* |
done | the cap of zero for a store that does not split |
Store-energy_balance |
done | both efficiencies, inflow and spill |
marginal_cost_dispatch, marginal_cost_store, spill_cost of a Store |
done | in Store_opex |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective11362.944444444443, 139 rows.
The network, as PyPSA code
rung_76_store_power.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 76: stores that cover a storage unit — a power rating through `max_hours`, an efficiency each way, inflow and spill, and a cost each way."""
from __future__ import annotations
import spine
def build():
"""The spine plus four stores: a rated battery with two efficiencies, a reservoir with inflow, an extendable lossy store and an unrated extendable one, under a swinging load."""
n = spine.build()
n.add(
'Store',
'battery76',
bus='south',
e_nom=40,
max_hours=2,
efficiency_store=0.9,
efficiency_dispatch=0.95,
marginal_cost_dispatch=1,
e_cyclic=True,
)
n.add(
'Store',
'reservoir76',
bus='south',
e_nom=60,
max_hours=4,
p_min_pu=0,
inflow=[20, 20, 20, 20],
spill_cost=2,
e_initial=50,
)
n.add(
'Store',
'ext76',
bus='south',
e_nom_extendable=True,
e_nom_max=80,
capital_cost=1,
max_hours=4,
efficiency_store=0.8,
marginal_cost_store=0.5,
e_cyclic=True,
)
n.add('Store', 'plain76', bus='north', e_nom_extendable=True, capital_cost=3, e_cyclic=True)
n.add('Load', 'swing76', bus='south', p_set=[0, 60, 0, 70])
return n
Rung 77 — energy limits on every dispatching component#
This rung runs n.optimize(multi_investment_periods=True) with e_sum_min or
e_sum_max on a Generator, a Link, a Process and a Store. Since
PyPSA/PyPSA#1987, PyPSA builds {c}-e_sum_min and {c}-e_sum_max for each of
them (constraints.py:2938-2954). The energy is the sum of p weighted by
the generator weighting. For a Link, p is the withdrawal at bus0, and for
a Store, the net dispatch. Under multi_investment_periods, each snapshot also
weighs its period's years, as in an operational_limit row
(constraints.py:2935-2937).
The file states the four pairs of rows over energy_limit_weight, the
generator weighting times period_weight_years. On a plain run the years are
one, so the rows collapse to the standard ones.
The generator's, the process's and the store's caps and the link's floor are
tight at the optimum. The three other limits are not. PyPSA solves to
11618.75.
With the years at 1, it solves to 16428.125. Without the generator's
cap, it solves to 7900.625. Without the link's floor, it solves to
11508.75, without the process's cap to 8677.5, and without the store's
cap to 6225.0 (#977).
| PyPSA | status | note |
|---|---|---|
Generator-e_sum_min/max under multi_investment_periods |
done | each snapshot weighs its period's years |
Link-e_sum_min/max, Process-e_sum_min/max, Store-e_sum_min/max |
done | the same row over each component's p |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective11618.75, 143 rows.
The network, as PyPSA code
rung_77_energy_limits.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 77: energy limits on every dispatching component — `e_sum_min` and `e_sum_max` on a generator, link, process and store, with each period weighted by its years."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def build():
"""A whole network, not the spine: eight snapshots over two periods of unequal years, a cheap supply on a fuel bus, and four units under energy limits feeding a hub."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(4)] + [(2030, datetime(2030, 1, 1, t)) for t in range(4)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [5.0, 10.0]
n.snapshot_weightings['generators'] = [1.0, 2.0, 1.0, 2.0, 1.0, 2.0, 1.0, 2.0]
n.add('Bus', ['fuel', 'hub'])
n.add('Generator', 'well77', bus='fuel', p_nom=200, marginal_cost=1)
n.add('Generator', 'cheap77', bus='hub', p_nom=50, marginal_cost=5, e_sum_max=2500)
n.add('Generator', 'dear77', bus='hub', p_nom=200, marginal_cost=80)
n.add('Link', 'pipe77', bus0='fuel', bus1='hub', p_nom=40, marginal_cost=90, e_sum_min=400, e_sum_max=5000)
n.add(
'Process',
'plant77',
bus0='fuel',
bus1='hub',
rate0=-1.25,
p_nom=40,
marginal_cost=6,
e_sum_min=100,
e_sum_max=1800,
)
n.add('Store', 'pile77', bus='hub', e_nom=200, e_initial=200, marginal_cost=3, e_sum_min=-50, e_sum_max=150)
n.add('Load', 'hub_load77', bus='hub', p_set=[90, 110, 100, 80, 100, 120, 110, 90])
return n
Rung 78 — a cycle limit on a store#
This rung runs n.optimize() with two stores that have cycles_max, the most
equivalent full cycles over the horizon (PyPSA/PyPSA#2006). PyPSA builds one
Store-cycles_max row per store: the discharge p + p_store, weighted by the
stores weighting and summed over the horizon, is at most cycles_max times
e_nom, or the chosen e_nom of an extendable store
(constraints.py:2607-2633). A finite cycles_max makes the store split its
dispatch, so p_store carries the charge and the discharge is at least zero
(rung 76).
The file states the row over Store_e_nom_effective, the chosen build where
the store is extendable and the given one otherwise.
Both rows are tight at the optimum. PyPSA solves to 15004.0, and the
extendable store builds 114. Without the fixed store's limit, PyPSA solves
to 14974.0, and without the extendable store's limit, to 14947.0 (#977).
| PyPSA | status | note |
|---|---|---|
Store-cycles_max |
done | over Store_e_nom_effective |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective15004.0, 75 rows.
The network, as PyPSA code
rung_78_store_cycles.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 78: a cycle limit on a store — the discharge over the horizon is at most `cycles_max` times the build, fixed or chosen."""
from __future__ import annotations
import spine
def build():
"""The spine plus a fixed and an extendable store on the south bus, each with a cycle limit, under a swinging load."""
n = spine.build()
n.add('Store', 'cell78', bus='south', e_nom=30, e_cyclic=True, cycles_max=0.5)
n.add('Store', 'pack78', bus='south', e_nom_extendable=True, capital_cost=1, e_cyclic=True, cycles_max=0.5)
n.add('Load', 'swing78', bus='south', p_set=[0, 60, 0, 70])
return n
Rung 79 — flow-based market coupling#
This rung runs n.optimize() with a flow-based domain (PyPSA/PyPSA#2069).
A GlobalConstraint of type flow_based is one critical network element with
contingency (CNEC). It lists its zonal power transfer distribution factors
(PTDF) as ptdf_<name> attributes. A name is a zone bus, or a corridor link
whose flow enters the row. PyPSA builds:
FlowBased-net_position, an unbounded variable per zone bus and snapshot, which leaves the zone's nodal balance (flow_based.py:207-234,constraints.py:1578)GlobalConstraint-ram, a row per CNEC and snapshot: the net positions and corridor flows, each times its factor, are at most the CNEC's remaining available marginram. Aramof NaN drops the row in that snapshot, and aninvestment_periodlimits the row to that period. PyPSA reads noconstant(flow_based.py:262-288)FlowBased-balance, the net positions of each nameddomainsumming to zero per snapshot (flow_based.py:289-293)
The file reads the PTDF columns long, as GlobalConstraint_zonal_ptdf and
GlobalConstraint_link_ptdf, and the domains as the relation
Bus_flow_based_domain, from data prep. FlowBased_injection is the zone's
term in Bus_injection. On a run without flow-based rows, no bus is a zone and
the balance is the standard one.
The rung has three zone buses that trade only through their net positions,
and a corridor link from the north to the south zone. The north-south CNEC is
tight in every snapshot, and the centre-south one where it has a ram.
PyPSA solves to 17005.3191. With the corridor link's factor at 0, it
solves to 22596.8085. With the centre-south ram at 30 in every
snapshot, it solves to 16872.3404. Without the domain, the zones cannot
trade and PyPSA solves to 45000.0 (#977).
| PyPSA | status | note |
|---|---|---|
FlowBased-net_position |
done | where Bus_flow_based_zone, and in Bus_injection as FlowBased_injection |
GlobalConstraint-ram |
done | per CNEC and snapshot, where GlobalConstraint_ram has a value |
FlowBased-balance |
done | per domain, through Bus_flow_based_domain |
flow_based with a sense other than <=, and the domain checks PyPSA raises on |
refused | Refusals |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective17005.31914893617, 55 rows.
The network, as PyPSA code
rung_79_flow_based.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 79: flow-based market coupling — three zones trade through their net positions under two critical network elements, one of them with a corridor link."""
from __future__ import annotations
import math
import pandas as pd
def build():
"""A whole network, not the spine: three zone buses with their own supply and load, a corridor link from one to another, and a two-row flow-based domain."""
import pypsa
n = pypsa.Network()
n.set_snapshots(pd.date_range('2015-01-01', periods=4, freq='h'))
n.snapshot_weightings['objective'] = [2.0, 1.5, 2.5, 3.0]
n.add('Bus', ['north79', 'centre79', 'south79'])
n.add('Generator', 'hydro79', bus='north79', p_nom=200, marginal_cost=5)
n.add('Generator', 'coal79', bus='centre79', p_nom=200, marginal_cost=30)
n.add('Generator', 'gas79', bus='south79', p_nom=200, marginal_cost=60)
n.add('Load', 'north_load79', bus='north79', p_set=20)
n.add('Load', 'centre_load79', bus='centre79', p_set=[40, 50, 60, 50])
n.add('Load', 'south_load79', bus='south79', p_set=[80, 90, 100, 70])
n.add('Link', 'hvdc79', bus0='north79', bus1='south79', p_nom=30, p_min_pu=-1)
ptdf = pd.DataFrame(
{'north79': [0.6, 0.2], 'centre79': [0.1, 0.5], 'south79': [-0.3, -0.2], 'hvdc79': [-0.4, 0.0]},
index=['ns79', 'cs79'],
)
ram = pd.DataFrame({'ns79': [40.0, 40.0, 40.0, 40.0], 'cs79': [30.0, math.nan, 25.0, 30.0]}, index=n.snapshots)
n.c.global_constraints.add_flow_based(ptdf, ram=ram, domain='core79')
return n
Rung 80 — maintenance of fixed modular committable units#
This rung runs n.optimize() with a fixed modular committable generator, link
and process, each of four modules and maintainable. In maintenance, PyPSA
bounds the modules on by p_nom / p_nom_mod for a fixed build and by
p_nom_max / p_nom_mod for an extendable one (constraints.py:588-594). Before
PyPSA/PyPSA#2100, a fixed build took p_nom_max too, which is infinite by
default, and n.optimize() returned ('ok', 'unknown') without a solution
(PyPSA/PyPSA#2094).
The file states the bound as Generator_maintenance_modules, and the Link
and Process ones. They are Generator_modules_installed for a fixed build.
Each unit takes all four modules off for one snapshot, and a dear backup
covers the gap. PyPSA solves to 16026.25. Without maintenance, it solves to
13656.25. At 061974e3, PyPSA returns ('ok', 'unknown').
| PyPSA | status | note |
|---|---|---|
Generator-maint-modstatus-*, fixed modular, and the Link and Process ones |
done | bound by Generator_maintenance_modules, p_nom based |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective16026.25, 179 rows.
The network, as PyPSA code
rung_80_fixed_modular_maintenance.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 80: maintenance of fixed modular committable units — a generator, a link and a process of four modules each take all their modules off for one event."""
from __future__ import annotations
import spine
def build():
"""The spine plus a shop bus served by a fixed modular committable generator, link and process, each maintainable, with a dear backup."""
n = spine.build()
n.add('Bus', 'shop')
unit = {
'p_nom': 40,
'p_nom_mod': 10,
'committable': True,
'maintainable': True,
'maintenance_duration': 1,
}
n.add('Generator', 'gen80', bus='shop', marginal_cost=5, **unit)
n.add('Link', 'link80', bus0='north', bus1='shop', marginal_cost=1, **unit)
n.add('Process', 'process80', bus0='north', bus1='shop', rate0=-1.25, marginal_cost=1, **unit)
n.add('Generator', 'shop_backup', bus='shop', p_nom=200, marginal_cost=60)
n.add('Load', 'shop_load', bus='shop', p_set=[70, 90, 60, 80])
return n
Rung 81 — a piecewise marginal cost curve#
This rung runs n.optimize() with a generator whose marginal_cost is a
curve. The curve is a dictionary from p_pu to the marginal cost of the
segment that ends there. PyPSA builds the column Generator-marginal_cost_piecewise, holds it
above the curve at Generator-p, and puts it in the objective in place of
p times marginal_cost (optimize.py:308-348). The curve runs over
p_pu times p_nom, and its cost sums each segment's width times its
marginal cost (piecewise.py:305-312). The model-building library linopy
chooses how to state the curve. Where every curve is convex and no generator
with a curve is committable, it states each segment's line (_chord) and the
curve's ends (_domain_lo, _domain_hi).
The file states the curve as piecewise: entries along the breakpoint
dimension, and you compute the breakpoints in data prep.
Generator_marginal_cost_lp states the segment lines where
Generator_marginal_cost_as_lp, linopy's choice for the whole call, is true.
The curve of curve81 costs 12, then 20, then 45 per MWh. A south load walks
it from 24 MW in the first snapshot to the end of the second segment, 40 MW, in
the next two. In the last, gas is at its limit, so curve81 runs at 74 MW,
34 MW into its dear third segment. Its output costs 13.33, 16, 16 and 29.32 per MWh on average,
where a linear marginal_cost is one number. PyPSA solves to 30890.0.
| PyPSA | status | note |
|---|---|---|
Generator-marginal_cost_piecewise_chord, _domain_lo, _domain_hi |
split | Generator_marginal_cost_lp, where Generator_marginal_cost_as_lp is true |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective30890.0, 60 rows.
The network, as PyPSA code
rung_81_piecewise_marginal_cost.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 81: a piecewise `marginal_cost` curve — a generator that is not committable pays a convex cost curve, which PyPSA states as the line of each segment."""
from __future__ import annotations
import spine
def build():
"""The spine plus a generator on the south bus whose marginal cost climbs from 12 to 20 to 45 over three segments, and a south load that walks it up every segment."""
n = spine.build()
n.add('Generator', 'curve81', bus='south', p_nom=80, marginal_cost={0.0: 0.0, 0.25: 12.0, 0.5: 20.0, 1.0: 45.0})
n.add('Load', 'south_peak81', bus='south', p_set=[20, 60, 100, 170])
return n
Rung 82 — a piecewise marginal cost curve on a committable generator#
This rung adds a committable generator with a curve to the network of rung 81.
PyPSA then gates every curve of the call: by Generator-status for a
committable generator, and by 1 for any other (optimize.py:314-318,
piecewise.py:200-210). linopy states no segment lines under a gate. It
states each curve by increments instead: a share of each segment
(_delta), filled in order (_order_binary, _delta_bound, _fill_order,
_binary_order), at most the gate (_active_bound). Two rows tie the output
(_link) and the cost (_output_link) to the shares.
The file states the increments as Generator_marginal_cost_incremental, with
activity: Generator_status, where Generator_marginal_cost_as_lp is false.
A generator that is not committable has no status, so its curve is ungated,
which is a gate of 1. Where a curve's output or cost is not strictly
increasing, linopy states the curve by weights on its breakpoints instead, a
special ordered set of type
2 (SOS2). The file keeps the increments there, which reach the same
optimum. Under linearized_unit_commitment, the status is a share between 0
and 1, and a curve gate is binary or integer. So
the patch assumes that no generator with a curve is
committable.
The curve of block82 is not convex: its first 20 MW cost 40 per MWh and the
rest 15. It stays off in the first two snapshots, where its status holds its
curve at zero. In the third snapshot it runs at 84 MW: it pays for its dear
first 20 MW to reach the cheap rest. curve82 drops to 20 MW there. In the last
snapshot it runs at 100 MW. PyPSA solves to 26135.0.
| PyPSA | status | note |
|---|---|---|
Generator-marginal_cost_piecewise_delta, _order_binary, _delta_bound, _fill_order, _binary_order, _link, _output_link |
split | Generator_marginal_cost_incremental, where Generator_marginal_cost_as_lp is false, and in place of SOS2 rows |
Generator-marginal_cost_piecewise_active_bound |
split | where Generator_status exists, and no row at gate 1 |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective26135.0, 164 rows.
The network, as PyPSA code
rung_82_piecewise_marginal_cost_committable.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 82: a piecewise `marginal_cost` curve on a committable generator — its status gates the curve, and PyPSA states every curve of the call by increments filled in order."""
from __future__ import annotations
import spine
def build():
"""Rung 81's network plus a committable generator on the south bus whose curve is not convex: its first 20 MW cost 40 and the rest 15, and every snapshot on costs 50."""
n = spine.build()
n.add('Generator', 'curve82', bus='south', p_nom=80, marginal_cost={0.0: 0.0, 0.25: 12.0, 0.5: 20.0, 1.0: 45.0})
n.add(
'Generator',
'block82',
bus='south',
p_nom=100,
committable=True,
stand_by_cost=50,
marginal_cost={0.0: 0.0, 0.2: 40.0, 1.0: 15.0},
)
n.add('Load', 'south_peak82', bus='south', p_set=[20, 60, 100, 180])
return n
Rung 85 — a global constraint for one period over storage that reopens per period, on a network with scenarios#
This rung takes the network of rung 75 over two futures, calm and stormy,
and adds an operational_limit row on the gas carrier for 2030. PyPSA builds
each row once per scenario, and closes the storage at that scenario's level
(global_constraints.py:490-495, :544-549, :700-705, :747-752). Before
PyPSA/PyPSA#2115, PyPSA read the closing level by snapshot alone and raised
KeyError (PyPSA/PyPSA#2113).
The file already states each row per scenario, over that scenario's storage level, so it needs no change.
The clean unit costs 60 in the calm future and 90 in the stormy one.
The gas limit of 1100 binds in both futures and leaves the CO2 cap slack.
PyPSA solves to 14670.0. Without the gas limit, the CO2 cap binds and PyPSA
solves to 14460.0. Without either row, it solves to 9000.0.
| PyPSA | status | note |
|---|---|---|
primary_energy, operational_limit over storage that reopens per period, with scenarios |
done | the row per scenario, with StorageUnit_closing_weight and Store_closing_weight in each |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective14670.0, 276 rows.
The network, as PyPSA code
rung_85_scenario_period_storage.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 85: a primary-energy cap and an operational limit for one period over storage that reopens per period, on a network with scenarios — each scenario closes its storage at its own level."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
#: the clean unit's marginal cost in each scenario
CLEAN_COST = {'calm': 60, 'stormy': 90}
def build():
"""Rung 75's network over two futures in which the clean unit costs 60 and 90, with a 2030 limit on the gas carrier as well."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(4)] + [(2030, datetime(2030, 1, 1, t)) for t in range(4)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [5.0, 10.0]
n.add('Carrier', 'gas85', co2_emissions=0.5)
n.add('Bus', 'hub')
n.add('Generator', 'gas85', bus='hub', carrier='gas85', p_nom=100, marginal_cost=30)
n.add('Generator', 'clean85', bus='hub', p_nom=200, marginal_cost=60)
n.add(
'StorageUnit',
'tank85',
bus='hub',
carrier='gas85',
p_nom=20,
max_hours=4,
state_of_charge_initial=60,
state_of_charge_initial_per_period=True,
)
n.add('Store', 'pile85', bus='hub', carrier='gas85', e_nom=60, e_initial=40, e_initial_per_period=True)
n.add('Load', 'town85', bus='hub', p_set=[60, 80, 70, 50, 90, 110, 100, 80])
n.add(
'GlobalConstraint',
'co2_2030',
type='primary_energy',
carrier_attribute='co2_emissions',
sense='<=',
constant=600,
investment_period=2030,
)
n.add(
'GlobalConstraint',
'gas_2030',
type='operational_limit',
carrier_attribute='gas85',
sense='<=',
constant=1100,
investment_period=2030,
)
n.set_scenarios({'calm': 0.6, 'stormy': 0.4})
for scenario, cost in CLEAN_COST.items():
n.c.generators.static.loc[(scenario, 'clean85'), 'marginal_cost'] = cost
return n
Rung 86 — a piecewise marginal cost curve over investment periods#
This rung runs n.optimize(multi_investment_periods=True) with the curve of
rung 81 on a generator. PyPSA states the curve at every snapshot of every
period, and weights its cost as it weights p times marginal_cost. Before
PyPSA/PyPSA#2117, PyPSA handed linopy the snapshots as a two-level index, which
linopy could not expand the breakpoints over, and raised ValueError
(PyPSA/PyPSA#2116). PyPSA now flattens the index for the curve
(piecewise.py:173-177).
The file states the curve per snapshot, and the period weight comes in
through the objective, so it needs no change.
The load walks curve86 into its dear third segment in both periods. In the
last snapshot, curve86 runs at its limit of 80 MW, and the backup unit covers
the rest. The 2020 period costs 1980 and the 2030 period 4530, which weighs
half, so PyPSA solves to 4245.0.
| PyPSA | status | note |
|---|---|---|
Generator-marginal_cost_piecewise_chord, _domain_lo, _domain_hi, over periods |
done | Generator_marginal_cost_lp at every snapshot, weighted by the objective |
✔
pypsa 1.4.0.post1.dev62+g3625c8168solves this rung's network at objective4245.0, 40 rows.
The network, as PyPSA code
rung_86_piecewise_marginal_cost_periods.py
# SPDX-FileCopyrightText: mathspec Contributors
#
# SPDX-License-Identifier: MIT
"""Rung 86: a piecewise `marginal_cost` curve under `multi_investment_periods` — the curve holds at every snapshot of every period, and each period weighs its cost."""
from __future__ import annotations
from datetime import datetime
import pandas as pd
OPTIMIZE = {'multi_investment_periods': True}
def build():
"""A whole network, not the spine: four snapshots over two periods, a generator with rung 81's curve, and a dear backup unit."""
import pypsa
n = pypsa.Network()
n.snapshots = pd.MultiIndex.from_tuples(
[(2020, datetime(2020, 1, 1, t)) for t in range(2)] + [(2030, datetime(2030, 1, 1, t)) for t in range(2)]
)
n.investment_periods = [2020, 2030]
n.investment_period_weightings['objective'] = [1.0, 0.5]
n.investment_period_weightings['years'] = [10.0, 10.0]
n.add('Bus', 'hub')
n.add('Generator', 'curve86', bus='hub', p_nom=80, marginal_cost={0.0: 0.0, 0.25: 12.0, 0.5: 20.0, 1.0: 45.0})
n.add('Generator', 'backup86', bus='hub', p_nom=200, marginal_cost=50)
n.add('Load', 'town86', bus='hub', p_set=[30, 60, 50, 100])
return n
Refusals#
PyPSA raises on the data in the table below, warns on it, or handles it outside its general rows. None of these refusals is a limit of the language. The file states seven of the checks as assumptions, which a tool runs when it attaches the data.
| PyPSA raises | on | here |
|---|---|---|
ValueError, optimize.py:506-513 |
a nonzero marginal_cost_quadratic on any Generator, Link, Process, StorageUnit or Store under a risk preference, omega = 0 included |
Where CVaR_omega has a row, the file states an assumption: Generator_marginal_cost_quadratic_without_risk_preference, and the Link, Process, StorageUnit and Store ones. |
ValueError, constraints.py:1953 |
fixed modular p_nom not a multiple of p_nom_mod |
The file states a fractional module cap. |
ValueError, constraints.py:1605 |
load on a bus with nothing attached | The file states no row for the bus, so its load goes unserved. |
ValueError, optimize.py:472 |
no component carries a cost | The file states a feasibility problem. |
NotImplementedError, global_constraints.py:473, :527, :683, :733 |
storage that carries its level across periods in a primary_energy or operational_limit row, with period years != 1 |
The file states an assumption: StorageUnit_primary_energy_carried_over_has_unit_years, StorageUnit_operational_limit_carried_over_has_unit_years, and the Store ones. |
ValueError, global_constraints.py:380-385, :623-628 |
a primary_energy or operational_limit row that names an investment_period without multi_investment_periods |
You compute GlobalConstraint_counts_snapshot in data prep. |
ValueError, constraints.py:2513, :2691 |
an extendable lossy branch with s_nom_max = inf, either mode |
You compute Line_loss_max and Transformer_loss_max in data prep. |
RuntimeError, constraints.py:2739 |
the secant loop passing max_segments |
You compute the segment axis in data prep. |
ValueError, abstract.py:634-643 |
an outage named in branch_outages that is active in no period of the run |
Line_BODF and Transformer_BODF have no value for the outage, so the file states no copy. |
ConsistencyError, consistency.py:1250 |
a bus port, *_nom_set, delay, cyclic_delay, v_nom, x_pu_eff or r_pu_eff that differs by scenario (constants.py:26-58) |
The file states no such difference, because the ports, Generator_p_nom_set and the others, and Link_output_delay and the others span no scenario. Line_cycle_weight, Line_BODF and the Transformer ones span no scenario either. |
NotImplementedError, global_constraints.py:68-70 |
a tech_capacity_expansion_limit row on a network with scenarios |
Where there is more than one scenario, the file states an assumption: GlobalConstraint_tech_capacity_expansion_limit_without_scenarios. One scenario and none look alike in the file, and PyPSA refuses one scenario too. |
ConsistencyError, consistency.py:1531-1585 |
a maintainable component whose maintenance_duration or maintenance_events is not positive, whose events do not fit the weighted horizon, or that is extendable with p_nom_max = inf |
The file states an assumption: Generator_maintenance_events_positive, -duration_positive, -duration_fits_the_horizon, -events_fit_the_horizon, -build_cap_is_finite, and the Link and Process ones. |
ValueError, piecewise.py:284-304 |
a piecewise marginal_cost curve on an extendable generator, or on one whose p_nom is zero, missing or infinite |
The file states an assumption: Generator_marginal_cost_curve_on_a_fixed_build. |
KeyError, constraints.py:1916 |
a *_nom_set on a unit that is not extendable, or that has active = False |
The file states no row: Generator-p_nom_set and the others exist only where the build is extendable, and an inactive unit is fixed (Data prep). |
KeyError, constraints.py:2943 |
e_sum_min or e_sum_max on a Generator, Link, Process or Store with active = False |
The file states the row with no variable: Generator-e_sum_min and its siblings are 0 >= e_sum_min, which fails where the floor is positive. |
ValueError, constraints.py:1606-1611 |
a loaded bus where every unit and branch has retired in some snapshot | The file states the row with no variable: Bus-nodal_balance requires the load alone to be zero, and no solution meets it. |
ValueError, components/descriptors.py:141-144 |
a transmission_volume_expansion_limit, transmission_expansion_cost_limit or tech_capacity_expansion_limit row that names an investment_period outside n.investment_periods |
You compute Line_volume_weight, Line_expansion_cost_weight, Line_tech_capacity_weight and the other components' ones in data prep. |
ValueError, flow_based.py:109-114 |
a flow_based row whose sense is not <= |
The file states an assumption: GlobalConstraint_flow_based_sense_is_le. |
ValueError, flow_based.py:71-93, :116-166, :196-200 |
a flow-based domain where a line, transformer or link that is no corridor joins two zones, a zone has nothing attached, a ptdf_<name> names no bus or link, a corridor is inactive or touches no zone, a zone is in two domains, only some rows name a domain, or a row has no ram |
You compute GlobalConstraint_zonal_ptdf, GlobalConstraint_link_ptdf, Bus_flow_based_zone and Bus_flow_based_domain in data prep. |
a warning, or ConsistencyError under strict, consistency.py:235-287 |
a transformer with phase_shift_min > phase_shift_max, or a decided phase shift with an infinite bound |
The file states the rows as PyPSA builds them. In the first case the transformer keeps its fixed phase_shift, and in the second Transformer-phase_shift is unbounded on that side. |
nothing, and PyPSA lags by elapsed time, multiports.py:100-133 |
a Link or Process port with delay > 0 under a generators weighting that is not one value over the horizon |
The file states an assumption: Link_output_delay_under_uniform_weighting, and the Process one. The file states the delay as a shift by whole snapshots, the same number in every period (#299). |
nothing, and PyPSA takes 0 * inf as 0, constraints.py:129-134 |
a fixed build of inf where its least or most per-unit output is zero |
The file states no assumption. Generator-fix-p-lower and the other fixed bounds contain the product as written, so the row is 0 * inf, which has no value. |
Data prep#
These parameters take their values from the PyPSA network, as PyPSA sets them before it builds a row.
- Every parameter computed from
r,xors_nomtakes the values after PyPSA applies the line and transformer types (network/power_flow.py:448-571). A line with atypehasrandxfrom the type, itslengthandnum_parallel. A transformer with atypealso hass_nom,phase_shiftandtap_ratiofrom it. - The carrier of a Line, a Link or a Store with a blank
carrieris the carrier of its bus. PyPSA (network/power_flow.py:810-812,:861-867) fills a Line's or a Link's frombus0, and a Store's frombus. Growth limits and global constraints use the filled carrier. Line_cycle_weightisx_pu_effon an AC sub-network andr_pu_effon a DC one (networks.py:1434-1439). PyPSA weighs each branch inKirchhoff-Voltage-Lawby the carrier of its sub-network.- In a
tech_capacity_expansion_limitrow with a bus,Line_tech_capacity_weightand the other components' ones count a branch, a Link and a Process at itsbus0(global_constraints.py:95). They count a one-port component at itsbus. - On a plain run,
period_weight_objectiveandperiod_weight_yearsare one. PyPSA (optimize.py:231-233,:293-297) readsinvestment_period_weightingsonly undermulti_investment_periods. StorageUnit_primary_energy_weightandStore_primary_energy_weighthave a value only for storage that is not cyclic. PyPSA (global_constraints.py:452,:508) readscyclic_state_of_chargeande_cyclicalone, so a unit that cycles per period withcyclic_state_of_charge = Falsestill counts. Anoperational_limitrow also skips storage withactive = False(:661,:714).Generator_p_nom_extendableand the others are false for a unit withactive = False. PyPSA (variables.py:360-368) builds{c}-p_nomonly for extendable units that are active, so an inactive unit has no capacity variable in PyPSA or in the file.Line_BODFandTransformer_BODFare the outage factors of the sub-networks of each period's active branches (abstract.py:652-664, rung 68). As in PyPSA, a branch not active in a period has no value there.Link_output_delayandProcess_output_delayare numbers of snapshots: the PyPSAdelaydivided by the one value of thegeneratorsweighting, rounded up. The PyPSAdelayis elapsed time in that weighting, and PyPSA (multiports.py:100-133) rounds the source down to a snapshot start.Store_max_hoursandStore_cycles_maxhave a value only where PyPSA'smax_hoursorcycles_maxis finite.Store_split_dispatchis PyPSA'ssplit_dispatch()(stores.py:88-116), true where an efficiency differs from one,marginal_cost_dispatchormarginal_cost_storeis nonzero, orcycles_maxis finite, in any snapshot.GlobalConstraint_zonal_ptdfandGlobalConstraint_link_ptdfare theptdf_<name>columns of theflow_basedrows, read long by bus or link.Bus_flow_based_zoneis true for a bus that names a column.flow_based_domainhas one label perdomainthe rows name, or one label where they name none, andBus_flow_based_domainmaps each zone to the domain whose rows name it (flow_based.py:65-93).Generator_marginal_cost_curve_pis a curve'sp_putimesp_nom, andGenerator_marginal_cost_curve_costis zero at the first breakpoint and adds each segment's width times themarginal_costat its upper end (piecewise.py:305-312). A generator with no curve has no rows. Themarginal_costat the first breakpoint prices nothing.global_constrainthas no label for a global constraint with nothing to count. PyPSA (global_constraints.py:100-101,:558-559,:761-762,:847-848) skips a row whose set is empty. In the file, that row would be0against its constant, which fails where the constant is on the wrong side of zero.
The file#
The file states a plain n.optimize(), its multi-period class under the multi_investment_periods keyword, and its stochastic class, which n.set_scenarios(...) sets. Every quantity chosen once the future is known (second-stage) spans the futures dispatch is chosen in (scenario). Capacity is chosen once, before the future is known, and paid at its expected cost in each investment period its build year and lifetime span. Operation is the expectation over the scenarios' weights, with a share priced at the worst-outcome mean cost (CVaR) only where a risk preference is set. A plain run's data has one scenario, one period, unit weights, no outage and every active flag true, so the problem is the standard one. On a security-constrained run, PyPSA copies each branch flow limit once per outage, and the outage factors, active flags and scenario weights come from data prep.
Sets#
| Symbol | Meaning |
|---|---|
| \(\mathcal{B}\) | index \(b\) — breakpoint — the corners of a piecewise-linear curve, in order. A curve runs over the breakpoints its data has rows for, so curves may differ in length |
| \(\mathcal{N}\) | index \(n\) — bus with \(\mathrm{Bus\_flow\_based\_domain}: \mathcal{N} \to \mathcal{L},\ \mathrm{Generator\_bus}: \mathcal{G} \to \mathcal{N},\ \mathrm{Line\_bus0}: \mathcal{K} \to \mathcal{N},\ \mathrm{Line\_bus1}: \mathcal{K} \to \mathcal{N},\ \mathrm{Link\_bus0}: \mathcal{L} \to \mathcal{N},\ \mathrm{Link\_output\_bus}: \mathcal{O} \to \mathcal{N},\ \mathrm{Load\_bus}: \mathcal{D} \to \mathcal{N},\ \mathrm{Process\_output\_bus}: \mathcal{R} \to \mathcal{N},\ \mathrm{StorageUnit\_bus}: \mathcal{S} \to \mathcal{N},\ \mathrm{Store\_bus}: \mathcal{V} \to \mathcal{N},\ \mathrm{Transformer\_bus0}: \mathcal{M} \to \mathcal{N},\ \mathrm{Transformer\_bus1}: \mathcal{M} \to \mathcal{N}\), Bus_flow_based_domain is absent where the data has no row — network nodes |
| \(\mathcal{I}\) | index \(i\) — carrier with \(\mathrm{Generator\_carrier}: \mathcal{G} \to \mathcal{I},\ \mathrm{Line\_carrier}: \mathcal{K} \to \mathcal{I},\ \mathrm{Link\_carrier}: \mathcal{L} \to \mathcal{I},\ \mathrm{Process\_carrier}: \mathcal{J} \to \mathcal{I},\ \mathrm{StorageUnit\_carrier}: \mathcal{S} \to \mathcal{I},\ \mathrm{Store\_carrier}: \mathcal{V} \to \mathcal{I}\) — energy carriers, what a growth limit is set per |
| \(\mathcal{C}\) | index \(c\) — cycle — independent cycles of the passive network graph, the cycle basis, data prep. Each period has its own basis, of the branches that are active in it. A label is a position in that period's basis, so one label is a different cycle in another period |
| \(\mathcal{L}\) | index \(l\) — flow_based_domain with \(\mathrm{Bus\_flow\_based\_domain}: \mathcal{N} \to \mathcal{L}\), Bus_flow_based_domain is absent where the data has no row — the flow-based domains, each with its own zero-sum balance of net positions (PyPSA's domain of a flow_based row), data prep. A run whose flow-based rows name no domain has one. A run without them has none |
| \(\mathcal{G}\) | index \(g\) — generator with \(\mathrm{Generator\_bus}: \mathcal{G} \to \mathcal{N},\ \mathrm{Generator\_carrier}: \mathcal{G} \to \mathcal{I},\ \mathrm{Generator\_maintenance\_cover} \subseteq \mathcal{T} \times \mathcal{G} \times \Xi \times \mathcal{T}\), Generator_maintenance_cover is absent where the data has no row — generating units, each on one bus |
| \(\mathcal{O}\) | index \(o\) — global_constraint — PyPSA's GlobalConstraint rows, one label per declared limit |
| \(\mathcal{K}\) | index \(k\) — line with \(\mathrm{Line\_bus0}: \mathcal{K} \to \mathcal{N},\ \mathrm{Line\_bus1}: \mathcal{K} \to \mathcal{N},\ \mathrm{Line\_carrier}: \mathcal{K} \to \mathcal{I},\ \mathrm{Outage\_line}: \mathcal{K}^{\mathrm{out}} \to \mathcal{K}\), Outage_line is absent where the data has no row — passive branches, each between two buses, their flow set by impedance |
| \(\mathcal{L}\) | index \(l\) — link with \(\mathrm{Link\_bus0}: \mathcal{L} \to \mathcal{N},\ \mathrm{Link\_carrier}: \mathcal{L} \to \mathcal{I},\ \mathrm{Link\_maintenance\_cover} \subseteq \mathcal{T} \times \mathcal{L} \times \Xi \times \mathcal{T},\ \mathrm{Link\_output\_link}: \mathcal{O} \to \mathcal{L}\), Link_maintenance_cover is absent where the data has no row — controllable connections, each from one bus to the buses it delivers to |
| \(\mathcal{O}\) | index \(o\) — link_output with \(\mathrm{Link\_output\_bus}: \mathcal{O} \to \mathcal{N},\ \mathrm{Link\_output\_link}: \mathcal{O} \to \mathcal{L}\) — a link's output ports, one label per port a link declares. The labels are PyPSA's bus1, bus2, … columns read long, data prep. A link of any number of output ports is then one term in the balance |
| \(\mathcal{D}\) | index \(d\) — load with \(\mathrm{Load\_bus}: \mathcal{D} \to \mathcal{N}\) — demands, each on one bus |
| \(\mathcal{K}^{\mathrm{out}}\) | index \(\kappa\) — outage with \(\mathrm{Outage\_line}: \mathcal{K}^{\mathrm{out}} \to \mathcal{K},\ \mathrm{Outage\_transformer}: \mathcal{K}^{\mathrm{out}} \to \mathcal{M}\), Outage_line is absent where the data has no row, Outage_transformer is absent where the data has no row — the passive branches a security-constrained run takes out one at a time (PyPSA's branch_outages), each a line or a transformer. A plain run has none |
| \(\mathcal{Y}\) | index \(y\) — period with \(\mathrm{snapshot\_period}: \mathcal{T} \to \mathcal{Y}\) — investment periods. It is PyPSA's investment_periods |
| \(\mathcal{J}\) | index \(j\) — process with \(\mathrm{Process\_carrier}: \mathcal{J} \to \mathcal{I},\ \mathrm{Process\_maintenance\_cover} \subseteq \mathcal{T} \times \mathcal{J} \times \Xi \times \mathcal{T},\ \mathrm{Process\_output\_process}: \mathcal{R} \to \mathcal{J}\), Process_maintenance_cover is absent where the data has no row — generalised multi-port converters, each with an internal power that every port draws or delivers at its own rate |
| \(\mathcal{R}\) | index \(r\) — process_output with \(\mathrm{Process\_output\_bus}: \mathcal{R} \to \mathcal{N},\ \mathrm{Process\_output\_process}: \mathcal{R} \to \mathcal{J}\) — a process's ports, one label per port a process declares. The labels are PyPSA's bus0, bus1, …, each with a signed rate, data prep. A process of any number of ports is then one term in the balance |
| \(\Xi\) | index \(\xi\) — scenario with \(\mathrm{Generator\_maintenance\_cover} \subseteq \mathcal{T} \times \mathcal{G} \times \Xi \times \mathcal{T},\ \mathrm{Link\_maintenance\_cover} \subseteq \mathcal{T} \times \mathcal{L} \times \Xi \times \mathcal{T},\ \mathrm{Process\_maintenance\_cover} \subseteq \mathcal{T} \times \mathcal{J} \times \Xi \times \mathcal{T}\), Generator_maintenance_cover is absent where the data has no row, Link_maintenance_cover is absent where the data has no row, Process_maintenance_cover is absent where the data has no row — the futures dispatch is chosen in, each with a weight |
| \(\mathcal{M}\) | index \(m\) — segment — the cuts a passive branch's loss curve is held above. They are PyPSA's tangents, as many as its segments count, or its secants, as many as its tolerance loop places. A lossless run has none |
| \(\mathcal{T}\) | index \(t\) — snapshot with \(\mathrm{Generator\_maintenance\_cover} \subseteq \mathcal{T} \times \mathcal{G} \times \Xi \times \mathcal{T},\ \mathrm{Link\_maintenance\_cover} \subseteq \mathcal{T} \times \mathcal{L} \times \Xi \times \mathcal{T},\ \mathrm{Process\_maintenance\_cover} \subseteq \mathcal{T} \times \mathcal{J} \times \Xi \times \mathcal{T},\ \mathrm{snapshot\_period}: \mathcal{T} \to \mathcal{Y}\), Generator_maintenance_cover is absent where the data has no row, Link_maintenance_cover is absent where the data has no row, Process_maintenance_cover is absent where the data has no row — dispatch periods |
| \(\mathcal{S}\) | index \(s\) — storage_unit with \(\mathrm{StorageUnit\_bus}: \mathcal{S} \to \mathcal{N},\ \mathrm{StorageUnit\_carrier}: \mathcal{S} \to \mathcal{I}\) — storage units, dispatch and store behind one bus connection |
| \(\mathcal{V}\) | index \(v\) — store with \(\mathrm{Store\_bus}: \mathcal{V} \to \mathcal{N},\ \mathrm{Store\_carrier}: \mathcal{V} \to \mathcal{I}\) — pure energy stores, each on one bus |
| \(\mathcal{M}\) | index \(m\) — transformer with \(\mathrm{Outage\_transformer}: \mathcal{K}^{\mathrm{out}} \to \mathcal{M},\ \mathrm{Transformer\_bus0}: \mathcal{M} \to \mathcal{N},\ \mathrm{Transformer\_bus1}: \mathcal{M} \to \mathcal{N}\), Outage_transformer is absent where the data has no row — passive branches between two buses, their flow set by impedance and tap ratio, with a phase shift fixed or optimised |
Parameters#
| Symbol | Meaning |
|---|---|
| \(\mathrm{Bus\_attached}\) | Bus_attached over \(\mathcal{N} \times \mathcal{T}\) — whether a load, a unit, a port or a branch that is active in the snapshot's period is attached to the bus, data prep. A bus nothing is attached to has no balance row |
| \(\mathrm{zone}\) | Bus_flow_based_zone over \(\mathcal{N}\) — whether a bus is a zone of a flow-based domain: some flow_based row has its ptdf_<bus> column, data prep |
| \(\alpha\) | CVaR_alpha (scalar), absent where the data has no row — PyPSA's risk_preference['alpha'] — the confidence level. The tail is the other 1 - alpha of the probability. CVaR_alpha has a value exactly where CVaR_omega has one |
| \(\omega\) | CVaR_omega (scalar), \(0\) where the data has no row — PyPSA's risk_preference['omega'] — the share of operating cost priced at the tail rather than in expectation. CVaR_omega has a value only where a risk preference is set. Without a risk preference, the file has no CVaR variables or rows. With omega = 0, PyPSA builds them, and the optimum is the risk-neutral one |
| \(\overline{\Delta}\) | Carrier_max_growth over \(\mathcal{I}\), absent where the data has no row — most capacity of a carrier that may be added in a period. No value means no limit. It is the least over the scenarios, as PyPSA takes it (global_constraints.py:239-243), data prep. PyPSA reads it only under multi_investment_periods (global_constraints.py:232-233). Otherwise you feed no value in data prep |
| \(\mathrm{r}\) | Carrier_max_relative_growth over \(\mathcal{I}\) — share of the previous period's additions that may be added on top: the least over the scenarios, as PyPSA takes it, data prep |
| \(\mathrm{Cycle\_in\_basis}\) | Cycle_in_basis over \(\mathcal{C} \times \mathcal{Y}\) — whether a period's cycle basis reaches the cycle label, data prep. A label beyond the basis of the period has no Kirchhoff voltage law row |
| \(\mathrm{on}\) | Generator_active over \(\mathcal{G} \times \mathcal{T}\) — whether a generator is active in a snapshot's period. It is PyPSA's active, from build year and lifetime, data prep |
| \(\mathrm{act}\) | Generator_active_asset over \(\mathcal{G}\) — whether a generator is one of PyPSA's active_assets, by its active flag alone, without build year and lifetime. A unit that is active in no snapshot of the horizon still counts. PyPSA refuses one that differs by scenario (constants.py:54) |
| \(\mathrm{M}\) | Generator_big_m over \(\mathcal{G} \times \Xi\), absent where the data has no row — the bound a committed extendable generator's big-M rows release it by, data prep. It is the build cap p_nom_max times the highest p_max_pu, where the cap is finite and positive. Elsewhere it is committable_big_m times the highest p_max_pu (components.py:1082-1152). Without that keyword, it is ten times the largest of the peak total load and the component's largest finite p_nom and p_nom_max. Where there is none of them, it is 1e6. Below the output a solve wants, it caps that output |
| \(\mathrm{c}^{\mathrm{cap}}\) | Generator_capital_cost over \(\mathcal{G} \times \Xi\) — cost of one unit of nominal power for the modelled horizon (PyPSA's periodized_cost), data prep. It is overnight_cost as an annuity over lifetime at discount_rate, times nyears, where overnight_cost is given, and capital_cost where it is not. Both add fom_cost (components.py:1157-1178, costs.py:102-203) |
| \(\mathrm{W}\) | Generator_capital_weight over \(\mathcal{G}\) — the sum of period weights a generator is active in. It is PyPSA's active * period_weighting, summed, data prep |
| \(\mathrm{com}\) | Generator_committable over \(\mathcal{G}\) — whether output is gated by an on/off status decision |
| \(\mathrm{rest}\) | Generator_down_time_brought_in over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — one while the down time a unit brought into the horizon is still running, zero after. It is PyPSA's stop down_time_before snapshots before the first, counted while min_down_time has not passed since (constraints.py:302-323, :677-678), data prep. Under a rolling horizon with linearised unit commitment, it is the fraction of the stop the previous window solved |
| \(\overline{\mathrm{E}}\) | Generator_e_sum_max over \(\mathcal{G} \times \Xi\) — most energy over the horizon: a fuel or emission budget in energy terms. It is infinity where no cap is meant |
| \(\underline{\mathrm{E}}\) | Generator_e_sum_min over \(\mathcal{G} \times \Xi\) — least energy over the horizon. It is minus infinity where no floor is meant |
| \(\mathrm{new}\) | Generator_first_active over \(\mathcal{G} \times \mathcal{Y}\) — one in the first period a generator is active in, zero elsewhere, data prep. PyPSA takes active & (active.cumsum() == 1) (global_constraints.py:278) |
| \(\mathrm{mnt}\) | Generator_maintainable over \(\mathcal{G}\) — whether a generator must be taken off for maintenance within the horizon: in any scenario, as PyPSA takes the union over them (components.py:1000-1003) |
| \(\tau^{\mathrm{mnt}}\) | Generator_maintenance_duration over \(\mathcal{G} \times \Xi\), absent where the data has no row — the hours of generator weightings one maintenance event covers, PyPSA's maintenance_duration. It has no value where the generator is not maintainable. No row contains it. In data prep you turn it into Generator_maintenance_cover and Generator_maintenance_start_blocked. The assumptions state that it fits the horizon |
| \(\mathrm{n}^{\mathrm{mnt}}\) | Generator_maintenance_events over \(\mathcal{G} \times \Xi\) — how many maintenance events occur within the horizon |
| \(\gamma\) | Generator_maintenance_pu over \(\mathcal{G} \times \Xi\) — the share of the build a maintenance event takes off |
| \(\mathrm{blk}\) | Generator_maintenance_start_blocked over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — true where no maintenance event may start, data prep. The snapshots it would cover there run past the end of the horizon or into one the generator is not active in. It is PyPSA's active & ~valid, from maintenance_duration and the generator weightings |
| \(\mathrm{c}\) | Generator_marginal_cost over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — cost of one unit of output |
| \(\mathrm{C}^{\mathrm{curve}}\) | Generator_marginal_cost_curve_cost over \(\mathcal{B} \times \mathcal{G}\), absent where the data has no row — the running cost at each breakpoint of a generator's piecewise marginal_cost curve: zero at the first, then each segment's width in output times the marginal_cost at its upper breakpoint, summed, data prep. It is linopy's Slopes(...).to_breakpoints (piecewise.py:310-312) |
| \(\mathrm{p}^{\mathrm{curve}}\) | Generator_marginal_cost_curve_p over \(\mathcal{B} \times \mathcal{G}\), absent where the data has no row — the output at each breakpoint of a generator's piecewise marginal_cost curve: PyPSA's p_pu breakpoints times p_nom, data prep. A generator with no curve has no rows |
| \(\mathrm{c}^{(2)}\) | Generator_marginal_cost_quadratic over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — cost of the square of one unit of output |
| \(\mathrm{DT}\) | Generator_min_down_time over \(\mathcal{G} \times \Xi\) — least snapshots a unit stays off once stopped |
| \(\mathrm{UT}\) | Generator_min_up_time over \(\mathcal{G} \times \Xi\) — least snapshots a unit stays on once started |
| \(\mathrm{N}^{\mathrm{fix}}\) | Generator_modules_installed over \(\mathcal{G} \times \Xi\) — how many whole modules a committable build has in place, data prep. It is Generator_p_nom / Generator_p_nom_mod where a fixed build is modular, and one where it is not. PyPSA refuses a fixed modular build whose nominal power is not a whole number of modules |
| \(\mathrm{b}\) | Generator_operational_limit_weight over \(\mathcal{G} \times \mathcal{O} \times \Xi\), absent where the data has no row — one where the generator is in the row's set, data prep. One outside it has no row |
| \(\mathrm{p}^{0}\) | Generator_p_init over \(\mathcal{G} \times \Xi\), \(0\) where the data has no row — the output a unit brought into the horizon, data prep. It is PyPSA's p_init where the unit came in running, and zero where it came in off. Under a rolling horizon, it is the output the previous window solved. No value means it is unknown, and the unit then has no ramp row at the first snapshot |
| \(\overline{\mathrm{p}}\) | Generator_p_max_pu over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — most output, per unit of nominal power: an availability profile |
| \(\underline{\mathrm{p}}\) | Generator_p_min_pu over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — least output, per unit of nominal power |
| \(\mathrm{nonneg}\) | Generator_p_min_pu_nonneg over \(\mathcal{G}\) — true where none of the generator's own minimums-per-unit is negative. It is PyPSA's per-unit (p_min_pu >= 0).all() over every snapshot and scenario, data prep |
| \(\mathrm{p}^{\mathrm{nom}}\) | Generator_p_nom over \(\mathcal{G} \times \Xi\) — nominal power |
| \(\mathrm{ext}\) | Generator_p_nom_extendable over \(\mathcal{G}\) — whether the nominal power is a decision |
| \(\overline{\mathrm{p}}^{\mathrm{nom}}\) | Generator_p_nom_max over \(\mathcal{G} \times \Xi\) — most nominal power an extendable generator may be built at |
| \(\underline{\mathrm{p}}^{\mathrm{nom}}\) | Generator_p_nom_min over \(\mathcal{G} \times \Xi\) — least nominal power an extendable generator may be built at |
| \(\mathrm{p}^{\mathrm{mod}}\) | Generator_p_nom_mod over \(\mathcal{G}\), absent where the data has no row — the module size a build comes in whole numbers of. No value means the build is continuous |
| \(\mathrm{p}^{\mathrm{nom,set}}\) | Generator_p_nom_set over \(\mathcal{G}\), absent where the data has no row — a given nominal power for an extendable generator. One without a value has no row here |
| \(\mathrm{p}^{\mathrm{set}}\) | Generator_p_set over \(\mathcal{G} \times \Xi \times \mathcal{T}\), absent where the data has no row — a given output schedule. A generator without one has no row here |
| \(\mathrm{a}\) | Generator_primary_energy_weight over \(\mathcal{G} \times \mathcal{O} \times \Xi \times \mathcal{T}\), absent where the data has no row — the constrained attribute per unit of energy at the bus: the carrier's co2_emissions over the generator's efficiency at the snapshot, data prep. A generator of an unweighted carrier has no row |
| \(\mathrm{rd}\) | Generator_ramp_limit_down over \(\mathcal{G} \times \Xi \times \mathcal{T}\), absent where the data has no row — most a generator may lower its output between snapshots, per unit of nominal power. No value means no limit. The value at the later of the two snapshots applies, so the limit may change over time |
| \(\mathrm{rd}^{\mathrm{dn}}\) | Generator_ramp_limit_shut_down over \(\mathcal{G} \times \Xi\), absent where the data has no row — most output in the snapshot before a unit stops, per unit of nominal power. No value means no limit |
| \(\mathrm{ru}^{\mathrm{up}}\) | Generator_ramp_limit_start_up over \(\mathcal{G} \times \Xi\), absent where the data has no row — most output in the snapshot a unit starts, per unit of nominal power. No value means no limit |
| \(\mathrm{ru}\) | Generator_ramp_limit_up over \(\mathcal{G} \times \Xi \times \mathcal{T}\), absent where the data has no row — most a generator may raise its output between snapshots, per unit of nominal power. No value means no limit. The value at the later of the two snapshots applies, so the limit may change over time |
| \(\mathrm{c}^{\mathrm{dn}}\) | Generator_shut_down_cost over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — cost of one stop in this snapshot |
| \(\mathrm{sgn}\) | Generator_sign over \(\mathcal{G}\) — the sign output enters its bus's balance with. It is PyPSA's sign, 1 unless given, -1 for a unit that draws power. PyPSA refuses one that differs by scenario (constants.py:46) |
| \(\mathrm{c}^{\mathrm{on}}\) | Generator_stand_by_cost over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — cost of one snapshot spent on |
| \(\mathrm{c}^{\mathrm{up}}\) | Generator_start_up_cost over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — cost of one start in this snapshot |
| \(\mathrm{u}^{0}\) | Generator_status_initial over \(\mathcal{G} \times \Xi\) — one where the unit was on before the first snapshot, zero where off. It is PyPSA's up_time_before > 0 for a committable unit, and one for a unit that is not, whose ramp rows PyPSA starts from on (constraints.py:323), data prep. Under a rolling horizon, it is the status the previous window solved, a fraction where unit commitment is linearised |
| \(\mathrm{m}\) | Generator_tech_capacity_weight over \(\mathcal{G} \times \mathcal{O}\), absent where the data has no row — one where the generator is in the row's carrier-and-bus set, data prep. One outside it, or one that is not active in the row's investment_period, has no row |
| \(\mathrm{hold}\) | Generator_up_time_brought_in over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — one while the up time a unit brought into the horizon is still running, zero after. It is PyPSA's start up_time_before snapshots before the first, counted while min_up_time has not passed since (constraints.py:302-323, :677-678), data prep. Under a rolling horizon with linearised unit commitment, it is the fraction of the start the previous window solved |
| \(\mathrm{K}\) | GlobalConstraint_constant over \(\mathcal{O} \times \Xi\) — the constant the total is held against. What a variable cannot represent is folded in here in data prep. That is an initial charge, times its period's years for each counted period where the storage reopens per period, or a non-extendable build. PyPSA reads it per scenario (global_constraints.py:577, :780, :859-864) |
| \(\mathrm{in}\) | GlobalConstraint_counts_snapshot over \(\mathcal{O} \times \Xi \times \mathcal{T}\), false where the data has no row — whether a row counts a snapshot in a scenario (PyPSA's investment_period), data prep. It is every snapshot where the row states no period, and only that period's where it states one. A row that states a period the run does not model has no label here, as PyPSA skips it (global_constraints.py:388). PyPSA reads the column only under multi_investment_periods. It fails on a row that states a period without it (global_constraints.py:380-385) |
| \(\mathrm{PTDF}^{f}\) | GlobalConstraint_link_ptdf over \(\mathcal{O} \times \mathcal{L} \times \Xi \times \mathcal{T}\), \(0\) where the data has no row — the factor of a flow_based row for a corridor link, whose flow enters the row: PyPSA's ptdf_<link> column read long, data prep. No value is zero |
| \(\mathrm{RAM}\) | GlobalConstraint_ram over \(\mathcal{O} \times \Xi \times \mathcal{T}\), absent where the data has no row — the remaining available margin of a flow_based row, one critical network element (CNEC). It may differ by snapshot. No value drops the row in that snapshot (flow_based.py:201) |
| \(\mathrm{sense}\) | GlobalConstraint_sense over \(\mathcal{O} \times \Xi\) — the row's sense in each scenario: <=, >= or ==. PyPSA reads a row's sense per scenario (global_constraints.py:576, :779, :859-864) |
| \(\mathrm{type}\) | GlobalConstraint_type over \(\mathcal{O}\) — which formula the row takes: primary_energy, operational_limit, transmission_volume_expansion_limit, transmission_expansion_cost_limit, tech_capacity_expansion_limit or flow_based |
| \(\mathrm{PTDF}^{z}\) | GlobalConstraint_zonal_ptdf over \(\mathcal{N} \times \mathcal{O} \times \Xi \times \mathcal{T}\), \(0\) where the data has no row — the zonal power transfer distribution factor of a flow_based row for a zone bus: PyPSA's ptdf_<bus> column read long, data prep. It may differ by snapshot. No value is zero, as PyPSA fills it |
| \(\beta\) | Line_BODF over \(\mathcal{K} \times \mathcal{K}^{\mathrm{out}} \times \mathcal{Y}\), absent where the data has no row — the share of an outaged branch's flow a line takes on when that branch goes out (PyPSA's BODF), data prep. It comes from the PTDF of the sub-network the period's active branches form. A row exists only where the line and the outage are active in the period and share a sub-network. It is -1 at the outaged line itself. PyPSA takes the factors of the first scenario (abstract.py:652), and refuses impedances that differ by scenario (constants.py:33-35). So the factors are the same in every scenario |
| \(\mathrm{on}^{s}\) | Line_active over \(\mathcal{K} \times \mathcal{T}\) — whether a line is active in a snapshot's period. It is PyPSA's active, data prep |
| \(\mathrm{act}^{s}\) | Line_active_asset over \(\mathcal{K}\) — whether a line is one of PyPSA's active_assets, by its active flag alone, without build year and lifetime. A unit that is active in no snapshot of the horizon still counts. PyPSA refuses one that differs by scenario (constants.py:54) |
| \(\mathrm{c}^{\mathrm{cap},s}\) | Line_capital_cost over \(\mathcal{K} \times \Xi\) — cost of one unit of nominal apparent power for the modelled horizon (PyPSA's periodized_cost), data prep. It is overnight_cost as an annuity over lifetime at discount_rate, times nyears, where overnight_cost is given, and capital_cost where it is not. Both add fom_cost (components.py:1157-1178, costs.py:102-203) |
| \(\mathrm{W}^{s}\) | Line_capital_weight over \(\mathcal{K}\) — the sum of period weights a line is active in. It is PyPSA's active * period_weighting, summed, data prep |
| \(\mathrm{x}\) | Line_cycle_weight over \(\mathcal{C} \times \mathcal{K} \times \mathcal{Y}\), absent where the data has no row — the line's series impedance, signed by its orientation in the cycle: the cycle basis, data prep. A line in no cycle has no row. PyPSA builds the cycle basis from the first scenario (networks.py:1430-1437), and refuses a v_nom, x_pu_eff or r_pu_eff that differs by scenario (constants.py:33-35). So the weight is the same in every scenario |
| \(\mathrm{cc}\) | Line_expansion_cost_weight over \(\mathcal{O} \times \mathcal{K} \times \Xi\), absent where the data has no row — the line's capital cost where its carrier is in the row's set, data prep. Where the row states no investment_period under multi_investment_periods, it is times the objective weights of the periods the line is active in. A line outside the set, or one that is not active in the row's period, has no row. The capital cost is PyPSA's capital_cost property, which is Line_capital_cost without fom_cost (components.py:1182-1200, global_constraints.py:905-914) |
| \(\mathrm{new}^{s}\) | Line_first_active over \(\mathcal{K} \times \mathcal{Y}\) — one in the first period a line is active in, zero elsewhere, data prep. PyPSA takes active & (active.cumsum() == 1) (global_constraints.py:278) |
| \(\overline{\ell}\) | Line_loss_max over \(\mathcal{K} \times \Xi \times \mathcal{T}\) — the loss at a line's rating. It is PyPSA's r_pu_eff * (s_max_pu * s_nom_max)**2, data prep |
| \(\mathrm{b}\) | Line_loss_offset over \(\mathcal{K} \times \Xi \times \mathcal{M} \times \mathcal{T}\) — where that cut meets the loss axis: a tangent's loss_k - slope_k * p_k, a secant's -r_pu_eff * p_k * p_k+1, negative, data prep |
| \(\mathrm{a}\) | Line_loss_slope over \(\mathcal{K} \times \Xi \times \mathcal{M} \times \mathcal{T}\) — the slope of a cut to the loss curve, data prep. It is a tangent's 2 * r_pu_eff * p_k at its segment's flow, or a secant's r_pu_eff * (p_k + p_k+1) between consecutive breakpoints |
| \(\overline{\mathrm{s}}\) | Line_s_max_pu over \(\mathcal{K} \times \Xi \times \mathcal{T}\) — most flow either way, per unit of nominal apparent power |
| \(\mathrm{s}^{\mathrm{nom}}\) | Line_s_nom over \(\mathcal{K} \times \Xi\) — nominal apparent power |
| \(\mathrm{ext}^{s}\) | Line_s_nom_extendable over \(\mathcal{K}\) — whether the nominal apparent power is a decision |
| \(\overline{\mathrm{s}}^{\mathrm{nom}}\) | Line_s_nom_max over \(\mathcal{K} \times \Xi\) — most nominal apparent power an extendable line may be built at |
| \(\underline{\mathrm{s}}^{\mathrm{nom}}\) | Line_s_nom_min over \(\mathcal{K} \times \Xi\) — least nominal apparent power an extendable line may be built at |
| \(\mathrm{s}^{\mathrm{mod}}\) | Line_s_nom_mod over \(\mathcal{K}\), absent where the data has no row — the module size a build comes in whole numbers of. No value means the build is continuous |
| \(\mathrm{s}^{\mathrm{nom,set}}\) | Line_s_nom_set over \(\mathcal{K}\), absent where the data has no row — a given nominal apparent power for an extendable line. One without a value has no row here |
| \(\mathrm{s}^{\mathrm{set}}\) | Line_s_set over \(\mathcal{K} \times \Xi \times \mathcal{T}\), absent where the data has no row — a given flow schedule. A line without one has no row here |
| \(\mathrm{m}^{l}\) | Line_tech_capacity_weight over \(\mathcal{O} \times \mathcal{K}\), absent where the data has no row — one where the line is in the row's carrier-and-bus set, data prep. One outside it, or one that is not active in the row's investment_period, has no row |
| \(\overline{\delta}\) | Line_v_ang_max over \(\mathcal{K} \times \Xi\), absent where the data has no row — the most the voltage angle difference across a line may be either way, in degrees. It is PyPSA's v_ang_max. Infinite, and so no row, by default. A line whose carrier is not AC has no row either. The deprecated v_ang_min is ignored, as PyPSA ignores it with a DeprecationWarning (constraints.py:1797-1804) |
| \(\mathrm{len}\) | Line_volume_weight over \(\mathcal{O} \times \mathcal{K} \times \Xi\), absent where the data has no row — the line's length in each scenario where its carrier is in the row's set (global_constraints.py:882-883), data prep. A line outside it, or one that is not active in the row's investment_period, has no row |
| \(\mathrm{x}^{\mathrm{eff}}\) | Line_x_pu_eff over \(\mathcal{K} \times \Xi\) — the line's effective series reactance. It is PyPSA's x_pu_eff, x over the square of its bus's nominal voltage, data prep |
| \(\mathrm{on}^{f}\) | Link_active over \(\mathcal{L} \times \mathcal{T}\) — whether a link is active in a snapshot's period. It is PyPSA's active, data prep |
| \(\mathrm{act}^{f}\) | Link_active_asset over \(\mathcal{L}\) — whether a link is one of PyPSA's active_assets, by its active flag alone, without build year and lifetime. A unit that is active in no snapshot of the horizon still counts. PyPSA refuses one that differs by scenario (constants.py:54) |
| \(\mathrm{M}^{f}\) | Link_big_m over \(\mathcal{L} \times \Xi\), absent where the data has no row — the bound a committed extendable link's big-M rows release it by, data prep. It is the build cap p_nom_max times the highest p_max_pu, where the cap is finite and positive. Elsewhere it is committable_big_m times the highest p_max_pu (components.py:1082-1152). Without that keyword, it is ten times the largest of the peak total load and the component's largest finite p_nom and p_nom_max. Where there is none of them, it is 1e6. Below the flow a solve wants, it caps that flow |
| \(\mathrm{c}^{\mathrm{cap},f}\) | Link_capital_cost over \(\mathcal{L} \times \Xi\) — cost of one unit of nominal power for the modelled horizon (PyPSA's periodized_cost), data prep. It is overnight_cost as an annuity over lifetime at discount_rate, times nyears, where overnight_cost is given, and capital_cost where it is not. Both add fom_cost (components.py:1157-1178, costs.py:102-203) |
| \(\mathrm{W}^{f}\) | Link_capital_weight over \(\mathcal{L}\) — the sum of period weights a link is active in. It is PyPSA's active * period_weighting, summed, data prep |
| \(\mathrm{com}^{f}\) | Link_committable over \(\mathcal{L}\) — whether flow is gated by an on/off status decision |
| \(\mathrm{rest}^{f}\) | Link_down_time_brought_in over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — one while the down time a link brought into the horizon is still running, zero after. It is PyPSA's stop down_time_before snapshots before the first, counted while min_down_time has not passed since (constraints.py:302-323, :677-678), data prep. Under a rolling horizon with linearised unit commitment, it is the fraction of the stop the previous window solved |
| \(\overline{\mathrm{E}}^{f}\) | Link_e_sum_max over \(\mathcal{L} \times \Xi\) — most energy of its flow at bus0, the withdrawal there, over the horizon. It is infinity where no cap is meant |
| \(\underline{\mathrm{E}}^{f}\) | Link_e_sum_min over \(\mathcal{L} \times \Xi\) — least energy of its flow at bus0, the withdrawal there, over the horizon. It is minus infinity where no floor is meant |
| \(\eta\) | Link_efficiency over \(\mathcal{O} \times \Xi \times \mathcal{T}\) — share of the flow that arrives at an output port, PyPSA's efficiency, efficiency2, … read long: negative where that port consumes rather than delivers. Read at the snapshot the flow arrives, so a delayed port delivers at its arrival snapshot's efficiency (constraints.py:1575) |
| \(\mathrm{cc}^{f}\) | Link_expansion_cost_weight over \(\mathcal{O} \times \mathcal{L} \times \Xi\), absent where the data has no row — the link's capital cost where its carrier is in the row's set, data prep. Where the row states no investment_period under multi_investment_periods, it is times the objective weights of the periods the link is active in. A link outside the set, or one that is not active in the row's period, has no row. The capital cost is PyPSA's capital_cost property, which is Link_capital_cost without fom_cost (components.py:1182-1200, global_constraints.py:905-914) |
| \(\mathrm{new}^{f}\) | Link_first_active over \(\mathcal{L} \times \mathcal{Y}\) — one in the first period a link is active in, zero elsewhere, data prep. PyPSA takes active & (active.cumsum() == 1) (global_constraints.py:278) |
| \(\mathrm{mnt}^{f}\) | Link_maintainable over \(\mathcal{L}\) — whether a link must be taken off for maintenance within the horizon: in any scenario, as PyPSA takes the union over them (components.py:1000-1003) |
| \(\tau^{f,\mathrm{mnt}}\) | Link_maintenance_duration over \(\mathcal{L} \times \Xi\), absent where the data has no row — the hours of generator weightings one maintenance event covers, PyPSA's maintenance_duration. It has no value where the link is not maintainable. No row contains it. In data prep you turn it into Link_maintenance_cover and Link_maintenance_start_blocked. The assumptions state that it fits the horizon |
| \(\mathrm{n}^{f,\mathrm{mnt}}\) | Link_maintenance_events over \(\mathcal{L} \times \Xi\) — how many maintenance events occur within the horizon |
| \(\gamma^{f}\) | Link_maintenance_pu over \(\mathcal{L} \times \Xi\) — the share of the build a maintenance event takes off |
| \(\mathrm{blk}^{f}\) | Link_maintenance_start_blocked over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — true where no maintenance event may start, data prep. The snapshots it would cover there run past the end of the horizon or into one the link is not active in. It is PyPSA's active & ~valid, from maintenance_duration and the generator weightings |
| \(\mathrm{c}^{f}\) | Link_marginal_cost over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — cost of one unit of flow |
| \(\mathrm{c}^{f,(2)}\) | Link_marginal_cost_quadratic over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — cost of the square of one unit of flow |
| \(\mathrm{DT}^{f}\) | Link_min_down_time over \(\mathcal{L} \times \Xi\) — least snapshots a link stays off once stopped |
| \(\mathrm{UT}^{f}\) | Link_min_up_time over \(\mathcal{L} \times \Xi\) — least snapshots a link stays on once started |
| \(\mathrm{N}^{f,\mathrm{fix}}\) | Link_modules_installed over \(\mathcal{L} \times \Xi\) — how many whole modules a committable build has in place, data prep. It is Link_p_nom / Link_p_nom_mod where a fixed build is modular, and one where it is not. PyPSA refuses a fixed modular build whose nominal power is not a whole number of modules |
| \(\mathrm{cyc}^{f}\) | Link_output_cyclic_delay over \(\mathcal{O}\) — whether a delayed port's flow wraps from the end of its investment period (PyPSA's cyclic_delay, cyclic_delay2, …). Where it does not, the flow still in transit at each period's first snapshots is lost. The flag is the same in every scenario, as the delay is |
| \(\mathrm{d}^{f}\) | Link_output_delay over \(\mathcal{O}\) — snapshots a port's delivery lags its link's flow (PyPSA's delay, delay2, … read long). In data prep you divide it by the snapshot_weightings.generators value and round it up. It is zero for a port that delivers at once. A port has the same delay in every scenario, because PyPSA refuses a delay that differs by scenario (constants.py:55) |
| \(\mathrm{f}^{0}\) | Link_p_init over \(\mathcal{L} \times \Xi\), \(0\) where the data has no row — the flow a link brought into the horizon, data prep. It is PyPSA's p_init where the link came in running, and zero where it came in off. Under a rolling horizon, it is the flow the previous window solved. No value means it is unknown, and the link then has no ramp row at the first snapshot |
| \(\overline{\mathrm{f}}\) | Link_p_max_pu over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — most flow, per unit of nominal power |
| \(\underline{\mathrm{f}}\) | Link_p_min_pu over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — least flow, per unit of nominal power: negative for a link that carries both ways |
| \(\mathrm{nonneg}^{f}\) | Link_p_min_pu_nonneg over \(\mathcal{L}\) — true where none of the link's own minimums-per-unit is negative. It is PyPSA's per-unit (p_min_pu >= 0).all() over every snapshot and scenario, data prep |
| \(\mathrm{f}^{\mathrm{nom}}\) | Link_p_nom over \(\mathcal{L} \times \Xi\) — nominal power |
| \(\mathrm{ext}^{f}\) | Link_p_nom_extendable over \(\mathcal{L}\) — whether the nominal power is a decision |
| \(\overline{\mathrm{f}}^{\mathrm{nom}}\) | Link_p_nom_max over \(\mathcal{L} \times \Xi\) — most nominal power an extendable link may be built at |
| \(\underline{\mathrm{f}}^{\mathrm{nom}}\) | Link_p_nom_min over \(\mathcal{L} \times \Xi\) — least nominal power an extendable link may be built at |
| \(\mathrm{f}^{\mathrm{mod}}\) | Link_p_nom_mod over \(\mathcal{L}\), absent where the data has no row — the module size a build comes in whole numbers of. No value means the build is continuous |
| \(\mathrm{f}^{\mathrm{nom,set}}\) | Link_p_nom_set over \(\mathcal{L}\), absent where the data has no row — a given nominal power for an extendable link. One without a value has no row here |
| \(\mathrm{f}^{\mathrm{set}}\) | Link_p_set over \(\mathcal{L} \times \Xi \times \mathcal{T}\), absent where the data has no row — a given flow schedule. A link without one has no row here |
| \(\mathrm{rd}^{f}\) | Link_ramp_limit_down over \(\mathcal{L} \times \Xi \times \mathcal{T}\), absent where the data has no row — most a link may lower its flow between snapshots, per unit of nominal power. No value means no limit. The value at the later of the two snapshots applies, so the limit may change over time |
| \(\mathrm{rd}^{f,\mathrm{dn}}\) | Link_ramp_limit_shut_down over \(\mathcal{L} \times \Xi\), absent where the data has no row — most flow in the snapshot before a link stops, per unit of nominal power. No value means no limit |
| \(\mathrm{ru}^{f,\mathrm{up}}\) | Link_ramp_limit_start_up over \(\mathcal{L} \times \Xi\), absent where the data has no row — most flow in the snapshot a link starts, per unit of nominal power. No value means no limit |
| \(\mathrm{ru}^{f}\) | Link_ramp_limit_up over \(\mathcal{L} \times \Xi \times \mathcal{T}\), absent where the data has no row — most a link may raise its flow between snapshots, per unit of nominal power. No value means no limit. The value at the later of the two snapshots applies, so the limit may change over time |
| \(\mathrm{c}^{f,\mathrm{dn}}\) | Link_shut_down_cost over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — cost of one stop in this snapshot |
| \(\mathrm{c}^{f,\mathrm{on}}\) | Link_stand_by_cost over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — cost of one snapshot spent on |
| \(\mathrm{c}^{f,\mathrm{up}}\) | Link_start_up_cost over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — cost of one start in this snapshot |
| \(\mathrm{u}^{f,0}\) | Link_status_initial over \(\mathcal{L} \times \Xi\) — one where the link was on before the first snapshot, zero where off. It is PyPSA's up_time_before > 0 for a committable link, and one for a link that is not, whose ramp rows PyPSA starts from on (constraints.py:323), data prep. Under a rolling horizon, it is the status the previous window solved, a fraction where unit commitment is linearised |
| \(\mathrm{m}^{f}\) | Link_tech_capacity_weight over \(\mathcal{O} \times \mathcal{L}\), absent where the data has no row — one where the link is in the row's carrier-and-bus set, data prep. One outside it, or one that is not active in the row's investment_period, has no row |
| \(\mathrm{hold}^{f}\) | Link_up_time_brought_in over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — one while the up time a link brought into the horizon is still running, zero after. It is PyPSA's start up_time_before snapshots before the first, counted while min_up_time has not passed since (constraints.py:302-323, :677-678), data prep. Under a rolling horizon with linearised unit commitment, it is the fraction of the start the previous window solved |
| \(\mathrm{len}^{f}\) | Link_volume_weight over \(\mathcal{O} \times \mathcal{L} \times \Xi\), absent where the data has no row — the link's length in each scenario where its carrier is in the row's set (global_constraints.py:882-883), data prep. A link outside it, or one that is not active in the row's investment_period, has no row |
| \(\mathrm{on}^{\mathrm{load}}\) | Load_active over \(\mathcal{D}\) — whether a load is active in the network, PyPSA's active. A load has no build year and no lifetime, so the flag is the same in every snapshot. PyPSA refuses one that differs by scenario (constants.py:54) |
| \(\mathrm{load}\) | Load_p_set over \(\mathcal{D} \times \Xi \times \mathcal{T}\) — demand |
| \(\mathrm{sgn}^{\mathrm{load}}\) | Load_sign over \(\mathcal{D}\) — the sign a load's demand enters its bus's balance with. It is PyPSA's sign, -1 unless given, 1 for a load that feeds its bus. PyPSA refuses one that differs by scenario (constants.py:46) |
| \(\mathrm{on}^{z}\) | Process_active over \(\mathcal{J} \times \mathcal{T}\) — whether a process is active in a snapshot's period. It is PyPSA's active, data prep |
| \(\mathrm{act}^{z}\) | Process_active_asset over \(\mathcal{J}\) — whether a process is one of PyPSA's active_assets, by its active flag alone, without build year and lifetime. A unit that is active in no snapshot of the horizon still counts. PyPSA refuses one that differs by scenario (constants.py:54) |
| \(\mathrm{M}^{z}\) | Process_big_m over \(\mathcal{J} \times \Xi\), absent where the data has no row — the bound a committed extendable process's big-M rows release it by, data prep. It is the build cap p_nom_max times the highest p_max_pu, where the cap is finite and positive. Elsewhere it is committable_big_m times the highest p_max_pu (components.py:1082-1152). Without that keyword, it is ten times the largest of the peak total load and the component's largest finite p_nom and p_nom_max. Where there is none of them, it is 1e6. Below the internal power a solve wants, it caps that internal power |
| \(\mathrm{c}^{\mathrm{cap},z}\) | Process_capital_cost over \(\mathcal{J} \times \Xi\) — cost of one unit of nominal power for the modelled horizon (PyPSA's periodized_cost), data prep. It is overnight_cost as an annuity over lifetime at discount_rate, times nyears, where overnight_cost is given, and capital_cost where it is not. Both add fom_cost (components.py:1157-1178, costs.py:102-203) |
| \(\mathrm{W}^{z}\) | Process_capital_weight over \(\mathcal{J}\) — the sum of period weights a process is active in. It is PyPSA's active * period_weighting, summed, data prep |
| \(\mathrm{com}^{z}\) | Process_committable over \(\mathcal{J}\) — whether internal power is gated by an on/off status decision |
| \(\mathrm{rest}^{z}\) | Process_down_time_brought_in over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — one while the down time a process brought into the horizon is still running, zero after. It is PyPSA's stop down_time_before snapshots before the first, counted while min_down_time has not passed since (constraints.py:302-323, :677-678), data prep. Under a rolling horizon with linearised unit commitment, it is the fraction of the stop the previous window solved |
| \(\overline{\mathrm{E}}^{z}\) | Process_e_sum_max over \(\mathcal{J} \times \Xi\) — most energy of its internal power over the horizon. It is infinity where no cap is meant |
| \(\underline{\mathrm{E}}^{z}\) | Process_e_sum_min over \(\mathcal{J} \times \Xi\) — least energy of its internal power over the horizon. It is minus infinity where no floor is meant |
| \(\mathrm{new}^{z}\) | Process_first_active over \(\mathcal{Y} \times \mathcal{J}\) — one in the first period a process is active in, zero elsewhere, data prep. PyPSA takes active & (active.cumsum() == 1) (global_constraints.py:278) |
| \(\mathrm{mnt}^{z}\) | Process_maintainable over \(\mathcal{J}\) — whether a process must be taken off for maintenance within the horizon: in any scenario, as PyPSA takes the union over them (components.py:1000-1003) |
| \(\tau^{z,\mathrm{mnt}}\) | Process_maintenance_duration over \(\mathcal{J} \times \Xi\), absent where the data has no row — the hours of generator weightings one maintenance event covers, PyPSA's maintenance_duration. It has no value where the process is not maintainable. No row contains it. In data prep you turn it into Process_maintenance_cover and Process_maintenance_start_blocked. The assumptions state that it fits the horizon |
| \(\mathrm{n}^{z,\mathrm{mnt}}\) | Process_maintenance_events over \(\mathcal{J} \times \Xi\) — how many maintenance events occur within the horizon |
| \(\gamma^{z}\) | Process_maintenance_pu over \(\mathcal{J} \times \Xi\) — the share of the build a maintenance event takes off |
| \(\mathrm{blk}^{z}\) | Process_maintenance_start_blocked over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — true where no maintenance event may start, data prep. The snapshots it would cover there run past the end of the horizon or into one the process is not active in. It is PyPSA's active & ~valid, from maintenance_duration and the generator weightings |
| \(\mathrm{c}^{z}\) | Process_marginal_cost over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — cost of one unit of internal power |
| \(\mathrm{c}^{z,(2)}\) | Process_marginal_cost_quadratic over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — cost of the square of one unit of internal power |
| \(\mathrm{DT}^{z}\) | Process_min_down_time over \(\mathcal{J} \times \Xi\) — least snapshots a process stays off once stopped |
| \(\mathrm{UT}^{z}\) | Process_min_up_time over \(\mathcal{J} \times \Xi\) — least snapshots a process stays on once started |
| \(\mathrm{N}^{z,\mathrm{fix}}\) | Process_modules_installed over \(\mathcal{J} \times \Xi\) — how many whole modules a committable build has in place, data prep. It is Process_p_nom / Process_p_nom_mod where a fixed build is modular, and one where it is not. PyPSA refuses a fixed modular build whose nominal power is not a whole number of modules |
| \(\mathrm{cyc}^{z}\) | Process_output_cyclic_delay over \(\mathcal{R}\) — whether a delayed port's transfer wraps from the end of its investment period (PyPSA's cyclic_delay0, cyclic_delay1, …). Where it does not, the energy still in transit at each period's first snapshots is lost. The flag is the same in every scenario, as the delay is |
| \(\mathrm{d}^{z}\) | Process_output_delay over \(\mathcal{R}\) — snapshots a port's transfer lags its process's internal power (PyPSA's delay0, delay1, … read long). In data prep you divide it by the snapshot_weightings.generators value and round it up. It is zero for a port that transfers at once. A port has the same delay in every scenario, as a link's port has |
| \(\mathrm{z}^{0}\) | Process_p_init over \(\mathcal{J} \times \Xi\), \(0\) where the data has no row — the internal power a process brought into the horizon, data prep. It is PyPSA's p_init where the process came in running, and zero where it came in off. Under a rolling horizon, it is the internal power the previous window solved. No value means it is unknown, and the process then has no ramp row at the first snapshot |
| \(\overline{\mathrm{z}}\) | Process_p_max_pu over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — most internal power, per unit of nominal power |
| \(\underline{\mathrm{z}}\) | Process_p_min_pu over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — least internal power, per unit of nominal power: negative for a process that runs both ways |
| \(\mathrm{nonneg}^{z}\) | Process_p_min_pu_nonneg over \(\mathcal{J}\) — true where none of the process's own minimums-per-unit is negative. It is PyPSA's per-unit (p_min_pu >= 0).all() over every snapshot and scenario, data prep |
| \(\mathrm{z}^{\mathrm{nom}}\) | Process_p_nom over \(\mathcal{J} \times \Xi\) — nominal internal power |
| \(\mathrm{ext}^{z}\) | Process_p_nom_extendable over \(\mathcal{J}\) — whether the nominal internal power is a decision |
| \(\overline{\mathrm{z}}^{\mathrm{nom}}\) | Process_p_nom_max over \(\mathcal{J} \times \Xi\) — most nominal power an extendable process may be built at |
| \(\underline{\mathrm{z}}^{\mathrm{nom}}\) | Process_p_nom_min over \(\mathcal{J} \times \Xi\) — least nominal power an extendable process may be built at |
| \(\mathrm{z}^{\mathrm{mod}}\) | Process_p_nom_mod over \(\mathcal{J}\), absent where the data has no row — the module size a build comes in whole numbers of. No value means the build is continuous |
| \(\mathrm{z}^{\mathrm{nom,set}}\) | Process_p_nom_set over \(\mathcal{J}\), absent where the data has no row — a given nominal power for an extendable process. One without a value has no row here |
| \(\mathrm{z}^{\mathrm{set}}\) | Process_p_set over \(\mathcal{J} \times \Xi \times \mathcal{T}\), absent where the data has no row — a given internal power schedule. A process without one has no row here |
| \(\mathrm{rd}^{z}\) | Process_ramp_limit_down over \(\mathcal{J} \times \Xi \times \mathcal{T}\), absent where the data has no row — most a process may lower its internal power between snapshots, per unit of nominal power. No value means no limit. The value at the later of the two snapshots applies, so the limit may change over time |
| \(\mathrm{rd}^{z,\mathrm{dn}}\) | Process_ramp_limit_shut_down over \(\mathcal{J} \times \Xi\), absent where the data has no row — most internal power in the snapshot before a process stops, per unit of nominal power. No value means no limit |
| \(\mathrm{ru}^{z,\mathrm{up}}\) | Process_ramp_limit_start_up over \(\mathcal{J} \times \Xi\), absent where the data has no row — most internal power in the snapshot a process starts, per unit of nominal power. No value means no limit |
| \(\mathrm{ru}^{z}\) | Process_ramp_limit_up over \(\mathcal{J} \times \Xi \times \mathcal{T}\), absent where the data has no row — most a process may raise its internal power between snapshots, per unit of nominal power. No value means no limit. The value at the later of the two snapshots applies, so the limit may change over time |
| \(\alpha\) | Process_rate over \(\mathcal{R} \times \Xi \times \mathcal{T}\) — the energy a port draws or delivers per unit of internal power (PyPSA's rate0, rate1, … read long). It is negative where the port withdraws and positive where it injects. A link is a process whose bus0 rate is minus one and whose output rates are its efficiencies. The rate at the snapshot the transfer arrives applies, so a delayed port transfers at its arrival snapshot's rate (constraints.py:1575) |
| \(\mathrm{c}^{z,\mathrm{dn}}\) | Process_shut_down_cost over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — cost of one stop in this snapshot |
| \(\mathrm{c}^{z,\mathrm{on}}\) | Process_stand_by_cost over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — cost of one snapshot spent on |
| \(\mathrm{c}^{z,\mathrm{up}}\) | Process_start_up_cost over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — cost of one start in this snapshot |
| \(\mathrm{u}^{z,0}\) | Process_status_initial over \(\mathcal{J} \times \Xi\) — one where the process was on before the first snapshot, zero where off. It is PyPSA's up_time_before > 0 for a committable process, and one for a process that is not, whose ramp rows PyPSA starts from on (constraints.py:323), data prep. Under a rolling horizon, it is the status the previous window solved, a fraction where unit commitment is linearised |
| \(\mathrm{m}^{z}\) | Process_tech_capacity_weight over \(\mathcal{O} \times \mathcal{J}\), absent where the data has no row — one where the process is in the row's carrier-and-bus set, data prep. One outside it, or one that is not active in the row's investment_period, has no row |
| \(\mathrm{hold}^{z}\) | Process_up_time_brought_in over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — one while the up time a process brought into the horizon is still running, zero after. It is PyPSA's start up_time_before snapshots before the first, counted while min_up_time has not passed since (constraints.py:302-323, :677-678), data prep. Under a rolling horizon with linearised unit commitment, it is the fraction of the start the previous window solved |
| \(\mathrm{on}^{h}\) | StorageUnit_active over \(\mathcal{T} \times \mathcal{S}\) — whether a storage unit is active in a snapshot's period. It is PyPSA's active, data prep |
| \(\mathrm{act}^{h}\) | StorageUnit_active_asset over \(\mathcal{S}\) — whether a storage unit is one of PyPSA's active_assets, by its active flag alone, without build year and lifetime. A unit that is active in no snapshot of the horizon still counts. PyPSA refuses one that differs by scenario (constants.py:54) |
| \(\mathrm{c}^{\mathrm{cap},h}\) | StorageUnit_capital_cost over \(\Xi \times \mathcal{S}\) — cost of one unit of nominal power for the modelled horizon (PyPSA's periodized_cost), data prep. It is overnight_cost as an annuity over lifetime at discount_rate, times nyears, where overnight_cost is given, and capital_cost where it is not. Both add fom_cost (components.py:1157-1178, costs.py:102-203) |
| \(\mathrm{W}^{h}\) | StorageUnit_capital_weight over \(\mathcal{S}\) — the sum of period weights a storage unit is active in. It is PyPSA's active * period_weighting, summed, data prep |
| \(\mathrm{cyc}\) | StorageUnit_cyclic_state_of_charge over \(\Xi \times \mathcal{S}\) — whether the horizon closes on itself instead of opening on the initial charge |
| \(\mathrm{cyc}^{y}\) | StorageUnit_cyclic_state_of_charge_per_period over \(\Xi \times \mathcal{S}\) — whether each investment period closes on itself instead of carrying its charge on to the next. It overrides cyclic_state_of_charge and state_of_charge_initial_per_period. PyPSA reads it only under multi_investment_periods. Otherwise you feed false in data prep |
| \(\eta^{+}\) | StorageUnit_efficiency_dispatch over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — share of the charge drawn down that reaches the bus |
| \(\eta^{-}\) | StorageUnit_efficiency_store over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — share of the power drawn from the bus that becomes charge |
| \(\mathrm{new}^{h}\) | StorageUnit_first_active over \(\mathcal{Y} \times \mathcal{S}\) — one in the first period a storage unit is active in, zero elsewhere, data prep. PyPSA takes active & (active.cumsum() == 1) (global_constraints.py:278) |
| \(\mathrm{idle}\) | StorageUnit_inactive_snapshots over \(\mathcal{S}\) — how many snapshots a storage unit is not active in. It is PyPSA's (~active).sum(), data prep. A cyclic unit reaches back this many snapshots further, so it closes on the last snapshot it is active in |
| \(\mathrm{inflow}\) | StorageUnit_inflow over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — energy arriving per hour, a river into a reservoir |
| \(\mathrm{c}^{h}\) | StorageUnit_marginal_cost over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — cost of one unit of dispatch |
| \(\mathrm{c}^{h,(2)}\) | StorageUnit_marginal_cost_quadratic over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — cost of the square of one unit of dispatch. Storing is not charged |
| \(\mathrm{c}^{\mathrm{soc}}\) | StorageUnit_marginal_cost_storage over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — cost of one unit of charge held over one snapshot |
| \(\mathrm{T}^{h}\) | StorageUnit_max_hours over \(\Xi \times \mathcal{S}\) — energy capacity, as hours of dispatch at nominal power |
| \(\mathrm{open}\) | StorageUnit_opens_late over \(\mathcal{T} \times \mathcal{S}\) — whether a snapshot is the first a storage unit is active in, where that is not the first of the horizon, data prep. It is PyPSA's active.cumsum() == 1 over the snapshots it is active in, past the first snapshot. It is false in a run where every unit is active throughout |
| \(\mathrm{b}^{h}\) | StorageUnit_operational_limit_weight over \(\mathcal{O} \times \Xi \times \mathcal{S}\), absent where the data has no row — one where the storage unit is in the row's set, data prep. One outside it has no row |
| \(\mathrm{h}^{+,\mathrm{set}}\) | StorageUnit_p_dispatch_set over \(\Xi \times \mathcal{T} \times \mathcal{S}\), absent where the data has no row — a given dispatch schedule. A unit without one has no row here |
| \(\overline{\mathrm{h}}\) | StorageUnit_p_max_pu over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — most dispatch, per unit of nominal power |
| \(\underline{\mathrm{h}}\) | StorageUnit_p_min_pu over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — most storing, per unit of nominal power and negated |
| \(\mathrm{h}^{\mathrm{nom}}\) | StorageUnit_p_nom over \(\Xi \times \mathcal{S}\) — nominal power |
| \(\mathrm{ext}^{h}\) | StorageUnit_p_nom_extendable over \(\mathcal{S}\) — whether the nominal power is a decision |
| \(\overline{\mathrm{h}}^{\mathrm{nom}}\) | StorageUnit_p_nom_max over \(\Xi \times \mathcal{S}\) — most nominal power an extendable storage unit may be built at |
| \(\underline{\mathrm{h}}^{\mathrm{nom}}\) | StorageUnit_p_nom_min over \(\Xi \times \mathcal{S}\) — least nominal power an extendable storage unit may be built at |
| \(\mathrm{h}^{\mathrm{mod}}\) | StorageUnit_p_nom_mod over \(\mathcal{S}\), absent where the data has no row — the module size a build comes in whole numbers of. No value means the build is continuous |
| \(\mathrm{h}^{\mathrm{nom,set}}\) | StorageUnit_p_nom_set over \(\mathcal{S}\), absent where the data has no row — a given nominal power for an extendable storage unit. One without a value has no row here |
| \(\mathrm{h}^{\mathrm{set}}\) | StorageUnit_p_set over \(\Xi \times \mathcal{T} \times \mathcal{S}\), absent where the data has no row — a given net dispatch schedule. A unit without one has no row here |
| \(\mathrm{h}^{-,\mathrm{set}}\) | StorageUnit_p_store_set over \(\Xi \times \mathcal{T} \times \mathcal{S}\), absent where the data has no row — a given charging schedule. A unit without one has no row here |
| \(\mathrm{a}^{h}\) | StorageUnit_primary_energy_weight over \(\mathcal{O} \times \Xi \times \mathcal{S}\), absent where the data has no row — the constrained attribute per unit of charge depleted, data prep. An unweighted unit has no row |
| \(\rho\) | StorageUnit_retention over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — share of charge kept over a snapshot. It is PyPSA's (1 - standing_loss) ** elapsed hours, data prep |
| \(\mathrm{sgn}^{h}\) | StorageUnit_sign over \(\mathcal{S}\) — the sign net dispatch enters its bus's balance with. It is PyPSA's sign, 1 unless given. PyPSA refuses one that differs by scenario (constants.py:46) |
| \(\mathrm{c}^{\mathrm{spill}}\) | StorageUnit_spill_cost over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — cost of one unit of inflow passed on unused |
| \(\mathrm{soc}^{0}\) | StorageUnit_state_of_charge_initial over \(\Xi \times \mathcal{S}\) — charge held before the first snapshot |
| \(\mathrm{reset}\) | StorageUnit_state_of_charge_initial_per_period over \(\Xi \times \mathcal{S}\) — whether each investment period opens on the initial charge instead of carrying the previous period's. PyPSA reads it only under multi_investment_periods. Otherwise you feed false in data prep |
| \(\mathrm{soc}^{\mathrm{set}}\) | StorageUnit_state_of_charge_set over \(\Xi \times \mathcal{T} \times \mathcal{S}\), absent where the data has no row — a given charge schedule. A unit without one has no row here |
| \(\mathrm{m}^{h}\) | StorageUnit_tech_capacity_weight over \(\mathcal{O} \times \mathcal{S}\), absent where the data has no row — one where the storage unit is in the row's carrier-and-bus set, data prep. One outside it, or one that is not active in the row's investment_period, has no row |
| \(\mathrm{on}^{e}\) | Store_active over \(\mathcal{T} \times \mathcal{V}\) — whether a store is active in a snapshot's period. It is PyPSA's active, data prep |
| \(\mathrm{act}^{e}\) | Store_active_asset over \(\mathcal{V}\) — whether a store is one of PyPSA's active_assets, by its active flag alone, without build year and lifetime. A unit that is active in no snapshot of the horizon still counts. PyPSA refuses one that differs by scenario (constants.py:54) |
| \(\mathrm{c}^{\mathrm{cap},e}\) | Store_capital_cost over \(\Xi \times \mathcal{V}\) — cost of one unit of nominal capacity for the modelled horizon (PyPSA's periodized_cost), data prep. It is overnight_cost as an annuity over lifetime at discount_rate, times nyears, where overnight_cost is given, and capital_cost where it is not. Both add fom_cost (components.py:1157-1178, costs.py:102-203) |
| \(\mathrm{W}^{e}\) | Store_capital_weight over \(\mathcal{V}\) — the sum of period weights a store is active in. It is PyPSA's active * period_weighting, summed, data prep |
| \(\mathrm{cyc}^{\max}\) | Store_cycles_max over \(\Xi \times \mathcal{V}\), absent where the data has no row — most equivalent full cycles over the horizon: discharge divided by e_nom. No value means PyPSA's default of infinity, and no limit |
| \(\mathrm{cyc}^{e}\) | Store_e_cyclic over \(\Xi \times \mathcal{V}\) — whether the horizon closes on itself instead of opening on the initial energy |
| \(\mathrm{cyc}^{e,y}\) | Store_e_cyclic_per_period over \(\Xi \times \mathcal{V}\) — whether each investment period closes on itself instead of carrying its energy on to the next. It overrides e_cyclic and e_initial_per_period. PyPSA reads it only under multi_investment_periods. Otherwise you feed false in data prep |
| \(\mathrm{e}^{0}\) | Store_e_initial over \(\Xi \times \mathcal{V}\) — energy held before the first snapshot |
| \(\mathrm{reset}^{e}\) | Store_e_initial_per_period over \(\Xi \times \mathcal{V}\) — whether each investment period opens on the initial energy instead of carrying the previous period's. PyPSA reads it only under multi_investment_periods. Otherwise you feed false in data prep |
| \(\overline{\mathrm{e}}\) | Store_e_max_pu over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — most energy held, per unit of nominal capacity |
| \(\underline{\mathrm{e}}\) | Store_e_min_pu over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — least energy held, per unit of nominal capacity: negative for a store that may go short |
| \(\mathrm{e}^{\mathrm{nom}}\) | Store_e_nom over \(\Xi \times \mathcal{V}\) — nominal energy capacity |
| \(\mathrm{ext}^{e}\) | Store_e_nom_extendable over \(\mathcal{V}\) — whether the nominal energy capacity is a decision |
| \(\overline{\mathrm{e}}^{\mathrm{nom}}\) | Store_e_nom_max over \(\Xi \times \mathcal{V}\) — most nominal capacity an extendable store may be built at |
| \(\underline{\mathrm{e}}^{\mathrm{nom}}\) | Store_e_nom_min over \(\Xi \times \mathcal{V}\) — least nominal capacity an extendable store may be built at |
| \(\mathrm{e}^{\mathrm{mod}}\) | Store_e_nom_mod over \(\mathcal{V}\), absent where the data has no row — the module size a build comes in whole numbers of. No value means the build is continuous |
| \(\mathrm{e}^{\mathrm{nom,set}}\) | Store_e_nom_set over \(\mathcal{V}\), absent where the data has no row — a given nominal capacity for an extendable store. One without a value has no row here |
| \(\mathrm{e}^{\mathrm{set}}\) | Store_e_set over \(\Xi \times \mathcal{T} \times \mathcal{V}\), absent where the data has no row — a given energy schedule. A store without one has no row here |
| \(\overline{\mathrm{E}}^{q}\) | Store_e_sum_max over \(\Xi \times \mathcal{V}\) — most energy of its net dispatch over the horizon. It is infinity where no cap is meant |
| \(\underline{\mathrm{E}}^{q}\) | Store_e_sum_min over \(\Xi \times \mathcal{V}\) — least energy of its net dispatch over the horizon. It is minus infinity where no floor is meant |
| \(\eta^{e,+}\) | Store_efficiency_dispatch over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — share of the energy drawn down that reaches the bus |
| \(\eta^{e,-}\) | Store_efficiency_store over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — share of the power drawn from the bus that becomes energy held |
| \(\mathrm{new}^{e}\) | Store_first_active over \(\mathcal{Y} \times \mathcal{V}\) — one in the first period a store is active in, zero elsewhere, data prep. PyPSA takes active & (active.cumsum() == 1) (global_constraints.py:278) |
| \(\mathrm{idle}^{e}\) | Store_inactive_snapshots over \(\mathcal{V}\) — how many snapshots a store is not active in. It is PyPSA's (~active).sum(), data prep. A cyclic store reaches back this many snapshots further, so it closes on the last snapshot it is active in |
| \(\mathrm{inflow}^{e}\) | Store_inflow over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — energy arriving per hour, a river into a reservoir |
| \(\mathrm{c}^{q}\) | Store_marginal_cost over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — cost of one unit of power delivered |
| \(\mathrm{c}^{q,+}\) | Store_marginal_cost_dispatch over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — cost of one unit of discharge, p + p_store |
| \(\mathrm{c}^{q,(2)}\) | Store_marginal_cost_quadratic over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — cost of the square of the net power delivered, so charging costs as much as delivering |
| \(\mathrm{c}^{e}\) | Store_marginal_cost_storage over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — cost of one unit of energy held over one snapshot |
| \(\mathrm{c}^{q,-}\) | Store_marginal_cost_store over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — cost of one unit of charging, p_store |
| \(\mathrm{T}^{e}\) | Store_max_hours over \(\Xi \times \mathcal{V}\), absent where the data has no row — energy capacity, as hours of discharge at the power rating. The rating is e_nom / max_hours. No value means PyPSA's default of infinity, and the store's power is then unbounded |
| \(\mathrm{open}^{e}\) | Store_opens_late over \(\mathcal{T} \times \mathcal{V}\) — whether a snapshot is the first a store is active in, where that is not the first of the horizon, data prep. It is PyPSA's active.cumsum() == 1 over the snapshots it is active in, past the first snapshot. It is false in a run where every store is active throughout |
| \(\mathrm{b}^{e}\) | Store_operational_limit_weight over \(\mathcal{O} \times \Xi \times \mathcal{V}\), absent where the data has no row — one where the store is in the row's set, data prep. One outside it has no row |
| \(\overline{\mathrm{q}}\) | Store_p_max_pu over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — most discharge, per unit of the power rating. It is PyPSA's p_max_pu, 1 unless given, read only where max_hours has a value |
| \(\underline{\mathrm{q}}\) | Store_p_min_pu over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — most charging, per unit of the power rating, negated. It is PyPSA's p_min_pu, -1 unless given, read only where max_hours has a value |
| \(\mathrm{q}^{\mathrm{set}}\) | Store_p_set over \(\Xi \times \mathcal{T} \times \mathcal{V}\), absent where the data has no row — a given schedule of power delivered. A store without one has no row here |
| \(\mathrm{a}^{e}\) | Store_primary_energy_weight over \(\mathcal{O} \times \Xi \times \mathcal{V}\), absent where the data has no row — the constrained attribute per unit of energy depleted, data prep. An unweighted store has no row |
| \(\rho^{e}\) | Store_retention over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — share of energy kept over a snapshot. It is PyPSA's (1 - standing_loss) ** elapsed hours, data prep |
| \(\mathrm{sgn}^{q}\) | Store_sign over \(\mathcal{V}\) — the sign the power a store delivers enters its bus's balance with. It is PyPSA's sign, 1 unless given. PyPSA refuses one that differs by scenario (constants.py:46) |
| \(\mathrm{c}^{\mathrm{spill},e}\) | Store_spill_cost over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — cost of one unit of inflow passed on unused |
| \(\mathrm{split}^{e}\) | Store_split_dispatch over \(\Xi \times \mathcal{V}\) — whether a store charges through p_store apart from its net dispatch p. It is PyPSA's split_dispatch(): true where an efficiency differs from one, a cost applies to one direction only, or cycles_max is finite, in any snapshot (stores.py:88-116), data prep. Where it is false, p_store is held at zero and the charging bound applies to p |
| \(\mathrm{m}^{e}\) | Store_tech_capacity_weight over \(\mathcal{O} \times \mathcal{V}\), absent where the data has no row — one where the store is in the row's carrier-and-bus set, data prep. One outside it, or one that is not active in the row's investment_period, has no row |
| \(\beta^{\sigma}\) | Transformer_BODF over \(\mathcal{K}^{\mathrm{out}} \times \mathcal{Y} \times \mathcal{M}\), absent where the data has no row — the share of an outaged branch's flow a transformer takes on when that branch goes out, as a line's. A row only where the transformer and the outage are active in the period and share a sub-network. It is the same in every scenario, as a line's |
| \(\mathrm{on}^{\sigma}\) | Transformer_active over \(\mathcal{T} \times \mathcal{M}\) — whether a transformer is active in a snapshot's period. It is PyPSA's active, data prep |
| \(\mathrm{act}^{\sigma}\) | Transformer_active_asset over \(\mathcal{M}\) — whether a transformer is one of PyPSA's active_assets, by its active flag alone, without build year and lifetime. A unit that is active in no snapshot of the horizon still counts. PyPSA refuses one that differs by scenario (constants.py:54) |
| \(\mathrm{c}^{\mathrm{cap},\sigma}\) | Transformer_capital_cost over \(\Xi \times \mathcal{M}\) — cost of one unit of nominal apparent power for the modelled horizon (PyPSA's periodized_cost), data prep. It is overnight_cost as an annuity over lifetime at discount_rate, times nyears, where overnight_cost is given, and capital_cost where it is not. Both add fom_cost (components.py:1157-1178, costs.py:102-203) |
| \(\mathrm{W}^{\sigma}\) | Transformer_capital_weight over \(\mathcal{M}\) — the sum of period weights a transformer is active in. It is PyPSA's active * period_weighting, summed, data prep |
| \(\mathrm{x}^{\sigma}\) | Transformer_cycle_weight over \(\mathcal{C} \times \mathcal{Y} \times \mathcal{M}\), absent where the data has no row — the transformer's effective series reactance, x times its tap ratio, signed by its orientation in the cycle. It is PyPSA's x_pu_eff, the cycle basis, data prep. A transformer in no cycle has no row. It is the same in every scenario, as a line's |
| \(\overline{\ell}^{\sigma}\) | Transformer_loss_max over \(\Xi \times \mathcal{T} \times \mathcal{M}\) — the loss at a transformer's rating (PyPSA's r_pu_eff * (s_max_pu * s_nom_max)**2), data prep. Its r_pu_eff is the resistance over the given s_nom times the tap ratio |
| \(\mathrm{b}^{\sigma}\) | Transformer_loss_offset over \(\Xi \times \mathcal{M} \times \mathcal{T} \times \mathcal{M}\) — where that cut meets the loss axis: a tangent's loss_k - slope_k * p_k, a secant's -r_pu_eff * p_k * p_k+1, negative, data prep |
| \(\mathrm{a}^{\sigma}\) | Transformer_loss_slope over \(\Xi \times \mathcal{M} \times \mathcal{T} \times \mathcal{M}\) — the slope of a cut to a transformer's loss curve, data prep. It is a tangent's 2 * r_pu_eff * p_k or a secant's r_pu_eff * (p_k + p_k+1), as a line's. It uses the transformer's own r_pu_eff and rating |
| \(\mathrm{Transformer\_phase\_shift\_cycle\_weight}\) | Transformer_phase_shift_cycle_weight over \(\mathcal{C} \times \mathcal{Y} \times \Xi \times \mathcal{M}\) — the cycle sign for a varying transformer's phase shift, times π/180 so a shift in degrees enters the cycle sum in radians, data prep. It is zero for a fixed transformer or one in no cycle |
| \(\varphi^{\sigma}\) | Transformer_phase_shift_fixed over \(\Xi \times \mathcal{T} \times \mathcal{M}\) — a fixed transformer's phase shift at each snapshot, in degrees. It is PyPSA's phase_shift, zero by default. A varying transformer's shift is a decision instead |
| \(\mathrm{Transformer\_phase\_shift\_max}\) | Transformer_phase_shift_max over \(\Xi \times \mathcal{M}\) — the most a varying transformer's phase shift may take, in degrees. It is PyPSA's phase_shift_max. Equal to phase_shift_min for a fixed transformer |
| \(\mathrm{Transformer\_phase\_shift\_min}\) | Transformer_phase_shift_min over \(\Xi \times \mathcal{M}\) — the least a varying transformer's phase shift may take, in degrees. It is PyPSA's phase_shift_min. Where it is below phase_shift_max the shift is a decision, otherwise the transformer keeps its fixed phase_shift |
| \(\mathrm{Transformer\_phase\_shift\_varying}\) | Transformer_phase_shift_varying over \(\Xi \times \mathcal{M}\) — whether a transformer's phase shift is a decision (PyPSA's phase_shift_min < phase_shift_max, read as a flag in data prep). False is a fixed shift, set by phase_shift. It may differ by scenario, as PyPSA reads it per scenario (variables.py:446-456) |
| \(\vartheta\) | Transformer_phase_shift_weight over \(\mathcal{C} \times \Xi \times \mathcal{T} \times \mathcal{M}\), absent where the data has no row — a fixed transformer's phase shift in radians at each snapshot, signed by its orientation in the cycle, data prep. It is a constant added to the cycle sum. It is zero for a varying transformer, whose shift is a decision instead. The constant and the variable term thus never both count a shift. A transformer with no shift or in no cycle of its snapshot's period has no row |
| \(\overline{\sigma}\) | Transformer_s_max_pu over \(\Xi \times \mathcal{T} \times \mathcal{M}\) — most flow either way, per unit of nominal apparent power |
| \(\sigma^{\mathrm{nom}}\) | Transformer_s_nom over \(\Xi \times \mathcal{M}\) — nominal apparent power |
| \(\mathrm{ext}^{\sigma}\) | Transformer_s_nom_extendable over \(\mathcal{M}\) — whether the nominal apparent power is a decision |
| \(\overline{\sigma}^{\mathrm{nom}}\) | Transformer_s_nom_max over \(\Xi \times \mathcal{M}\) — most nominal apparent power an extendable transformer may be built at |
| \(\underline{\sigma}^{\mathrm{nom}}\) | Transformer_s_nom_min over \(\Xi \times \mathcal{M}\) — least nominal apparent power an extendable transformer may be built at |
| \(\sigma^{\mathrm{mod}}\) | Transformer_s_nom_mod over \(\mathcal{M}\), absent where the data has no row — the module size a build comes in whole numbers of. No value means the build is continuous |
| \(\sigma^{\mathrm{nom,set}}\) | Transformer_s_nom_set over \(\mathcal{M}\), absent where the data has no row — a given nominal apparent power for an extendable transformer. One without a value has no row here |
| \(\sigma^{\mathrm{set}}\) | Transformer_s_set over \(\Xi \times \mathcal{T} \times \mathcal{M}\), absent where the data has no row — a given flow schedule. A transformer without one has no row here |
| \(\overline{\delta}^{\sigma}\) | Transformer_v_ang_max over \(\Xi \times \mathcal{M}\), absent where the data has no row — the most the voltage angle difference across a transformer, its phase shift included, may be either way, in degrees. It is PyPSA's v_ang_max. Infinite, and so no row, by default. The deprecated v_ang_min is ignored, as a line's |
| \(\mathrm{x}^{\mathrm{eff},\sigma}\) | Transformer_x_pu_eff over \(\Xi \times \mathcal{M}\) — the transformer's effective series reactance. It is PyPSA's x_pu_eff, x over its s_nom times its tap ratio, data prep |
| \(\mathrm{w}^{y}\) | period_weight_objective over \(\mathcal{Y}\) — PyPSA's investment_period_weightings.objective — what a period's cost weighs. PyPSA reads it only under multi_investment_periods. Otherwise you feed one in data prep, whatever the column contains (optimize.py:231-233, :293-297) |
| \(\mathrm{w}^{\mathrm{yr}}\) | period_weight_years over \(\mathcal{Y}\) — PyPSA's investment_period_weightings.years — what a period's energy weighs in a primary_energy, operational_limit or e_sum row. PyPSA reads it only under multi_investment_periods. Otherwise you feed one in data prep |
| \(\pi\) | scenario_weight over \(\Xi\) — PyPSA's scenario_weightings.weight — the probability of a future |
| \(\mathrm{w}^{\mathrm{gen}}\) | snapshot_weightings_generators over \(\mathcal{T}\) — PyPSA's snapshot_weightings.generators — hours a snapshot weighs in an energy total |
| \(\mathrm{w}\) | snapshot_weightings_objective over \(\mathcal{T}\) — PyPSA's snapshot_weightings.objective — hours a snapshot weighs in the cost |
| \(\mathrm{w}^{\mathrm{sto}}\) | snapshot_weightings_stores over \(\mathcal{T}\) — PyPSA's snapshot_weightings.stores — hours a snapshot weighs in a storage balance |
| \(\mathrm{lossy}\) | transmission_losses (scalar) — whether the network dissipates transmission losses (PyPSA's transmission_losses read as a flag). Its mode, tangents or secants, only decides how you fill the segment axis in data prep. The rows are the same. False with no segments is a lossless run. A security-constrained run over a network with passive branches builds no loss. PyPSA hands the keyword not to create_model (abstract.py:645-650) but to the solver (:549), so you feed false there in data prep |
Conditions#
| Symbol | Meaning |
|---|---|
| \(\mathrm{on}^{\mathrm{com,ext}}\) | Generator_com_ext over \(\mathcal{G} \times \mathcal{T}\) — a committable generator with an extendable, non-modular build that is active in the snapshot's period. These are PyPSA's com-ext rows, whose status is relaxed against the chosen build |
| \(\mathrm{on}^{\mathrm{com}}\) | Generator_committed over \(\mathcal{G} \times \mathcal{T}\) — a committable generator that is active in the snapshot's period: every unit-commitment row's set |
| \(\mathrm{on}^{\mathrm{mnt,ext}}\) | Generator_maint_ext over \(\mathcal{G} \times \mathcal{T}\) — a maintainable generator with an extendable build that is active in the snapshot's period, unless it is committable and modular. These are the maintenance rows against the chosen build |
| \(\mathrm{lp}\) | Generator_marginal_cost_as_lp (scalar) — linopy's choice of method for all generator cost curves at once (method="auto"): the segment lines where every curve is convex and no generator with a curve is committable, the incremental rows otherwise. PyPSA gates the curves by Generator-status where one is committable, and linopy states no segment lines under a gate (optimize.py:314-318, piecewise.py:200-224) |
| \(\mathrm{curve}\) | Generator_marginal_cost_curve over \(\mathcal{G}\) — a generator that pays along a piecewise marginal_cost curve: one of PyPSA's active_assets with curve data (optimize.py:307, piecewise.py:230-257) |
| \(\mathrm{bend}\) | Generator_marginal_cost_curve_bends_down over \(\mathcal{B} \times \mathcal{G}\) — a breakpoint where a generator's cost curve gets less steep: the segment after it rises less per unit of output than the one before |
| \(\mathrm{prev}\) | Generator_ramps_from_previous over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — a snapshot whose ramp uses a previous output. It is any snapshot but its period's first, and the horizon's first where the generator comes in off or carries an initial output |
| \(\mathrm{on}^{s,\mathrm{lossy}}\) | Line_lossy over \(\mathcal{K} \times \mathcal{T}\) — a line that is active in the snapshot's period, where the run models transmission losses |
| \(\mathrm{on}^{f,\mathrm{com,ext}}\) | Link_com_ext over \(\mathcal{L} \times \mathcal{T}\) — a committable link with an extendable, non-modular build that is active in the snapshot's period. These are PyPSA's com-ext rows, whose status is relaxed against the chosen build |
| \(\mathrm{on}^{f,\mathrm{com}}\) | Link_committed over \(\mathcal{L} \times \mathcal{T}\) — a committable link that is active in the snapshot's period: every unit-commitment row's set |
| \(\mathrm{on}^{f,\mathrm{mnt,ext}}\) | Link_maint_ext over \(\mathcal{L} \times \mathcal{T}\) — a maintainable link with an extendable build that is active in the snapshot's period, unless it is committable and modular. These are the maintenance rows against the chosen build |
| \(\mathrm{prev}^{f}\) | Link_ramps_from_previous over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — a snapshot whose ramp uses a previous output. It is any snapshot but its period's first, and the horizon's first where the link comes in off or carries an initial output |
| \(\mathrm{on}^{z,\mathrm{com,ext}}\) | Process_com_ext over \(\mathcal{J} \times \mathcal{T}\) — a committable process with an extendable, non-modular build that is active in the snapshot's period. These are PyPSA's com-ext rows, whose status is relaxed against the chosen build |
| \(\mathrm{on}^{z,\mathrm{com}}\) | Process_committed over \(\mathcal{J} \times \mathcal{T}\) — a committable process that is active in the snapshot's period: every unit-commitment row's set |
| \(\mathrm{on}^{z,\mathrm{mnt,ext}}\) | Process_maint_ext over \(\mathcal{J} \times \mathcal{T}\) — a maintainable process with an extendable build that is active in the snapshot's period, unless it is committable and modular. These are the maintenance rows against the chosen build |
| \(\mathrm{prev}^{z}\) | Process_ramps_from_previous over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — a snapshot whose ramp uses a previous output. It is any snapshot but its period's first, and the horizon's first where the process comes in off or carries an initial output |
| \(\mathrm{on}^{h,\mathrm{ext}}\) | StorageUnit_ext over \(\mathcal{T} \times \mathcal{S}\) — a storage unit with an extendable build that is active in the snapshot's period. It is PyPSA's ext rows |
| \(\mathrm{on}^{h,\mathrm{fix}}\) | StorageUnit_fix over \(\mathcal{T} \times \mathcal{S}\) — a storage unit with a fixed build that is active in the snapshot's period. It is PyPSA's fix rows |
| \(\mathrm{on}^{\sigma,\mathrm{lossy}}\) | Transformer_lossy over \(\mathcal{T} \times \mathcal{M}\) — a transformer that is active in the snapshot's period, where the run models transmission losses |
Variables#
| Symbol | Meaning |
|---|---|
| \(CVaR\) | CVaR (scalar), \(0\) where the mask leaves it out — CVaR — the tail's average cost, what the objective prices at omega |
| \(a\) | CVaR_a over \(\Xi\), \(0\) where the mask leaves it out — CVaR-a — how far a scenario's operating cost exceeds the tail's start. It is nothing where it does not |
| \(\theta\) | CVaR_theta (scalar), \(0\) where the mask leaves it out — CVaR-theta — where the tail starts, the value at risk |
| \(\mathit{NP}\) | FlowBased_net_position over \(\mathcal{N} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — FlowBased-net_position — what a zone bus exports to the flow-based domain, unbounded. It leaves the zone's nodal balance (flow_based.py:207-234). A bus that is no zone has none |
| \(\mu\) | Generator_maintenance over \(\mathcal{G} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — Generator-maintenance — whether a maintainable generator is in maintenance: continuous, and one exactly where an event covers the snapshot |
| \(\mu^{\mathrm{nom}}\) | Generator_maintenance_capacity over \(\mathcal{G} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — Generator-maintenance_capacity — the chosen build while in maintenance, zero otherwise: the product the maintcap rows linearise |
| \(\mu^{\mathrm{up}}\) | Generator_maintenance_start over \(\mathcal{G} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — Generator-maintenance_start — whether a maintenance event starts in this snapshot |
| \(\mu^{u}\) | Generator_maintenance_status over \(\mathcal{G} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — Generator-maintenance_status — the status while in maintenance, zero otherwise: the product the maint-status rows linearise, so a unit in maintenance may also be off |
| \(C^{\mathrm{curve}}\) | Generator_marginal_cost_piecewise over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — Generator-marginal_cost_piecewise — the running cost of a generator that pays along a curve, held above the curve |
| \(N\) | Generator_n_mod over \(\mathcal{G}\) — Generator-n_mod — how many modules of an extendable modular build. A unit that is not active has none |
| \(p\) | Generator_p over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — Generator-p — output of a generator in a snapshot |
| \(P\) | Generator_p_nom_ext over \(\mathcal{G}\) — Generator-p_nom — nominal power where it is a decision, for a generator that is active. A fixed build is the parameter of the same PyPSA name |
| \(\mathit{dn}\) | Generator_shut_down over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — Generator-shut_down — how much of a committable unit turns off this snapshot, capped as the status is |
| \(\mathit{up}\) | Generator_start_up over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — Generator-start_up — how much of a committable unit turns on this snapshot, capped as the status is |
| \(u\) | Generator_status over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — Generator-status — how much of a committable unit is on. It is an integer the rows below cap at one, or at the module count where the build is modular |
| \(\ell\) | Line_loss over \(\mathcal{K} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — Line-loss — what a line dissipates carrying its flow, pushed down by the cost and held up by the cuts. Absent, and zero in the balance, where the network is lossless |
| \(N^{s}\) | Line_n_mod over \(\mathcal{K}\) — Line-n_mod — how many modules of an extendable modular build. A unit that is not active has none |
| \(s\) | Line_s over \(\mathcal{K} \times \Xi \times \mathcal{T}\) — Line-s — PyPSA's p0, the flow measured at the Line_bus0 end: a positive value withdraws there and injects at Line_bus1, lossless |
| \(S\) | Line_s_nom_ext over \(\mathcal{K}\) — Line-s_nom — nominal apparent power where it is a decision, for a line that is active. A fixed build is the parameter of the same PyPSA name |
| \(\mu^{f}\) | Link_maintenance over \(\mathcal{L} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — Link-maintenance — whether a maintainable link is in maintenance: continuous, and one exactly where an event covers the snapshot |
| \(\mu^{f,\mathrm{nom}}\) | Link_maintenance_capacity over \(\mathcal{L} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — Link-maintenance_capacity — the chosen build while in maintenance, zero otherwise: the product the maintcap rows linearise |
| \(\mu^{f,\mathrm{up}}\) | Link_maintenance_start over \(\mathcal{L} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — Link-maintenance_start — whether a maintenance event starts in this snapshot |
| \(\mu^{f,u}\) | Link_maintenance_status over \(\mathcal{L} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — Link-maintenance_status — the status while in maintenance, zero otherwise: the product the maint-status rows linearise, so a unit in maintenance may also be off |
| \(N^{f}\) | Link_n_mod over \(\mathcal{L}\) — Link-n_mod — how many modules of an extendable modular build. A unit that is not active has none |
| \(f\) | Link_p over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — Link-p — PyPSA's p0, the flow measured at the Link_bus0 end. A positive value withdraws there and injects at every bus the link's output ports deliver to |
| \(F\) | Link_p_nom_ext over \(\mathcal{L}\) — Link-p_nom — nominal power where it is a decision, for a link that is active. A fixed build is the parameter of the same PyPSA name |
| \(\mathit{dn}^{f}\) | Link_shut_down over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — Link-shut_down — how much of a committable link turns off this snapshot, capped as the status is |
| \(\mathit{up}^{f}\) | Link_start_up over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — Link-start_up — how much of a committable link turns on this snapshot, capped as the status is |
| \(u^{f}\) | Link_status over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — Link-status — how much of a committable link is on. It is an integer the rows below cap at one, or at the module count where the build is modular |
| \(\mu^{z}\) | Process_maintenance over \(\mathcal{J} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — Process-maintenance — whether a maintainable process is in maintenance: continuous, and one exactly where an event covers the snapshot |
| \(\mu^{z,\mathrm{nom}}\) | Process_maintenance_capacity over \(\mathcal{J} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — Process-maintenance_capacity — the chosen build while in maintenance, zero otherwise: the product the maintcap rows linearise |
| \(\mu^{z,\mathrm{up}}\) | Process_maintenance_start over \(\mathcal{J} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — Process-maintenance_start — whether a maintenance event starts in this snapshot |
| \(\mu^{z,u}\) | Process_maintenance_status over \(\mathcal{J} \times \Xi \times \mathcal{T}\), \(0\) where the mask leaves it out — Process-maintenance_status — the status while in maintenance, zero otherwise: the product the maint-status rows linearise, so a unit in maintenance may also be off |
| \(N^{z}\) | Process_n_mod over \(\mathcal{J}\) — Process-n_mod — how many modules of an extendable modular build. A unit that is not active has none |
| \(z\) | Process_p over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — Process-p — PyPSA's internal power p. A positive value drives every port at its own rate, withdrawing where the rate is negative and injecting where it is positive |
| \(Z\) | Process_p_nom_ext over \(\mathcal{J}\) — Process-p_nom — nominal internal power where it is a decision, for a process that is active. A fixed build is the parameter of the same PyPSA name |
| \(\mathit{dn}^{z}\) | Process_shut_down over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — Process-shut_down — how much of a committable process turns off this snapshot, capped as the status is |
| \(\mathit{up}^{z}\) | Process_start_up over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — Process-start_up — how much of a committable process turns on this snapshot, capped as the status is |
| \(u^{z}\) | Process_status over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — Process-status — how much of a committable process is on. It is an integer the rows below cap at one, or at the module count where the build is modular |
| \(N^{h}\) | StorageUnit_n_mod over \(\mathcal{S}\) — StorageUnit-n_mod — how many modules of an extendable modular build. A unit that is not active has none |
| \(h^{+}\) | StorageUnit_p_dispatch over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — StorageUnit-p_dispatch — power delivered to the bus |
| \(H\) | StorageUnit_p_nom_ext over \(\mathcal{S}\) — StorageUnit-p_nom — nominal power where it is a decision, for a storage unit that is active. A fixed build is the parameter of the same PyPSA name |
| \(h^{-}\) | StorageUnit_p_store over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — StorageUnit-p_store — power drawn from the bus into charge |
| \(\mathit{spill}\) | StorageUnit_spill over \(\Xi \times \mathcal{T} \times \mathcal{S}\), \(0\) where the mask leaves it out — StorageUnit-spill — inflow passed on unused. It is zero where there is no inflow, so the balance keeps its row there. The bounds are PyPSA's, on the variable rather than as rows |
| \(\mathit{soc}\) | StorageUnit_state_of_charge over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — StorageUnit-state_of_charge — energy held at the end of a snapshot |
| \(e\) | Store_e over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — Store-e — energy held at the end of a snapshot |
| \(E\) | Store_e_nom_ext over \(\mathcal{V}\) — Store-e_nom — nominal capacity where it is a decision, for a store that is active. A fixed build is the parameter of the same PyPSA name |
| \(N^{e}\) | Store_n_mod over \(\mathcal{V}\) — Store-n_mod — how many modules of an extendable modular build. A unit that is not active has none |
| \(q\) | Store_p over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — Store-p — net power delivered to the bus. Charging is negative |
| \(q^{-}\) | Store_p_store over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — Store-p_store — power drawn from the bus to charge, so that p + p_store is the discharge. It is held at zero by its rows where the store does not split its dispatch |
| \(\mathit{spill}^{e}\) | Store_spill over \(\Xi \times \mathcal{T} \times \mathcal{V}\), \(0\) where the mask leaves it out — Store-spill — inflow passed on unused. It is zero where there is no inflow, so the balance keeps its row there. The bounds are PyPSA's, on the variable rather than as rows |
| \(\ell^{\sigma}\) | Transformer_loss over \(\Xi \times \mathcal{T} \times \mathcal{M}\), \(0\) where the mask leaves it out — Transformer-loss — what a transformer dissipates carrying its flow, as a line does. Absent, and zero in the balance, where the network is lossless |
| \(N^{\sigma}\) | Transformer_n_mod over \(\mathcal{M}\) — Transformer-n_mod — how many modules of an extendable modular build. A unit that is not active has none |
| \(\mathit{Transformer\_phase\_shift}\) | Transformer_phase_shift over \(\Xi \times \mathcal{T} \times \mathcal{M}\), \(0\) where the mask leaves it out — Transformer-phase_shift — a phase-shifting transformer's voltage angle shift in degrees, chosen per snapshot to redistribute the flows around its cycles without moving active power. Absent, and zero in the cycle sum, where the shift is fixed |
| \(\sigma\) | Transformer_s over \(\Xi \times \mathcal{T} \times \mathcal{M}\) — Transformer-s — PyPSA's p0, the flow measured at the Transformer_bus0 end: a positive value withdraws there and injects at Transformer_bus1, lossless |
| \(\Sigma\) | Transformer_s_nom_ext over \(\mathcal{M}\) — Transformer-s_nom — nominal apparent power where it is a decision, for a transformer that is active. A fixed build is the parameter of the same PyPSA name |
Definitions#
| Symbol | Meaning |
|---|---|
| \(\mathit{Bus\_injection}\) | Bus_injection over \(\mathcal{N} \times \Xi \times \mathcal{T}\) — what every component puts into a bus, less what it takes out of it. PyPSA writes each term into the balance, and a load on its right-hand side |
| \(\mathit{Carrier\_additions}\) | Carrier_additions over \(\mathcal{I} \times \mathcal{Y}\) — what a carrier adds in a period: every extendable component of that carrier, counting each build in the first period it is active in. Like PyPSA, it sums only the components that carry a carrier attribute, so a transformer, which has none, counts in no carrier |
| \(\mathrm{r}^{+}\) | Carrier_relative_growth over \(\mathcal{I}\) — the share of the previous period's additions in a carrier's growth limit, PyPSA's max_relative_growth clipped at zero. A negative share thus adds nothing and never tightens the limit |
| \(\mathit{Cycle\_angle\_sum}\) | Cycle_angle_sum over \(\mathcal{C} \times \Xi \times \mathcal{T}\) — the voltage angle differences around a cycle: every branch flow times its cycle weight, and every transformer phase shift |
| \(\mathit{NP}^{\mathrm{inj}}\) | FlowBased_injection over \(\mathcal{N} \times \Xi \times \mathcal{T}\) — what a zone bus exports to its flow-based domain leaves the bus (constraints.py:1578) |
| \(\mathit{Generator\_additions}\) | Generator_additions over \(\mathcal{I} \times \mathcal{Y}\) |
| \(\mathit{Generator\_capex}\) | Generator_capex (scalar) |
| \(\mathit{Generator\_commitment\_opex}\) | Generator_commitment_opex over \(\Xi\) |
| \(\mathit{Generator\_injection}\) | Generator_injection over \(\mathcal{N} \times \Xi \times \mathcal{T}\) |
| \(\mathrm{Generator\_maintenance\_modules}\) | Generator_maintenance_modules over \(\mathcal{G} \times \Xi\) — how many modules of a modular committable generator can be on in maintenance: its build cap in modules where the build is extendable, and the modules installed where it is fixed (constraints.py:588-594) |
| \(\mathrm{M}\) | Generator_modules over \(\mathcal{G} \times \Xi\) — how many modules a committed unit's status counts: the chosen number where the build is extendable and modular, and the modules installed otherwise (constraints.py:656-661) |
| \(\mathit{Generator\_operational\_limit}\) | Generator_operational_limit over \(\mathcal{O} \times \Xi\) |
| \(\mathit{Generator\_opex}\) | Generator_opex over \(\Xi\) |
| \(\overleftarrow{p}^{\circ}\) | Generator_p_carried_over over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — the output a generator carries over into a snapshot. At the first, it is the p_init it brought in, which is zero where it came in off. After that, it is the previous snapshot's |
| \(\widehat{\mathrm{p}}^{\mathrm{nom}}\) | Generator_p_nom_committed over \(\mathcal{G} \times \Xi\) — the build a committed unit's ramp rows are taken against: one module where the build is modular, since its status counts modules, and the given build otherwise (constraints.py:1131-1134) |
| \(\widetilde{\mathrm{p}}^{\mathrm{nom}}\) | Generator_p_nom_effective over \(\mathcal{G} \times \Xi\) — the build a generator's limits are taken against: the chosen one where it is extendable, the given one otherwise |
| \(\overleftarrow{p}\) | Generator_previous_p over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — the output a generator carries into a snapshot. It is nothing at the first snapshot it is active in past the first of the horizon: PyPSA fills an output it did not build with zero (constraints.py:716). It is the output carried over otherwise |
| \(\overleftarrow{u}\) | Generator_previous_status over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — the commitment state a generator carries into a snapshot. It is off at the first snapshot it is active in past the first of the horizon. PyPSA reads a status it did not build as off there (constraints.py:391). It is the state carried over otherwise |
| \(\mathit{Generator\_primary\_energy}\) | Generator_primary_energy over \(\mathcal{O} \times \Xi\) |
| \(\Delta^{-}\) | Generator_ramp_down_allowance over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — how far a generator may lower output between two snapshots. It is its ramp limit of the build while it stays on, plus its shut-down ramp in the snapshot it turns off |
| \(\widetilde{\mathrm{rd}}\) | Generator_ramp_down_rate over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — the ramp limit in a unit's down row, PyPSA's ramp_limit_down. It is the full build where PyPSA has none, since a shut-down ramp alone builds the row |
| \(\overleftarrow{\bar u}\) | Generator_ramp_previous_status over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — the status a committed unit's ramp rows carry into a snapshot. At the first snapshot, a fixed modular unit that came in running has every module on, since the state it brought in is on or off (constraints.py:1152-1157). The transition and down-time rows count it as one module (PyPSA/PyPSA#2114). It is the state carried in otherwise |
| \(\Delta^{+}\) | Generator_ramp_up_allowance over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — how far a generator may raise output between two snapshots. It is its ramp limit of the build while it stays on, plus its start-up ramp in the snapshot it turns on |
| \(\widetilde{\mathrm{ru}}\) | Generator_ramp_up_rate over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — the ramp limit in a unit's up row, PyPSA's ramp_limit_up. It is the full build where PyPSA has none, since a start-up ramp alone builds the row |
| \(C^{\mathrm{run}}\) | Generator_running_cost over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — what a generator's output costs in a snapshot: the value read off its piecewise curve where it has one, and its linear marginal_cost otherwise. A curve replaces the linear cost, as PyPSA drops the unit from that term (optimize.py:333-348) |
| \(\widetilde{\mathrm{rd}}^{\mathrm{dn}}\) | Generator_shut_down_rate over \(\mathcal{G} \times \Xi\) — the shut-down ramp in a unit's down row (PyPSA's ramp_limit_shut_down), or the full build where it has none |
| \(\widetilde{\mathrm{ru}}^{\mathrm{up}}\) | Generator_start_up_rate over \(\mathcal{G} \times \Xi\) — the start-up ramp in a unit's up row (PyPSA's ramp_limit_start_up), or the full build where it has none |
| \(\overleftarrow{u}^{\circ}\) | Generator_status_carried_over over \(\mathcal{G} \times \Xi \times \mathcal{T}\) — the state a generator carries over into a snapshot: the state it brought into the horizon at the first, the previous snapshot's after that |
| \(\mathit{Generator\_tech\_capacity\_expansion}\) | Generator_tech_capacity_expansion over \(\mathcal{O}\) |
| \(\mathit{w}^{\mathrm{gc}}\) | GlobalConstraint_energy_weight over \(\mathcal{O} \times \Xi \times \mathcal{T}\) — what one unit of power at a snapshot counts for in a row. Where the row counts the snapshot, it is the generator weighting times the years of the snapshot's period. Where it does not, it is nothing |
| \(\mathit{Line\_additions}\) | Line_additions over \(\mathcal{I} \times \mathcal{Y}\) |
| \(\mathit{Line\_angle\_sum}\) | Line_angle_sum over \(\mathcal{C} \times \Xi \times \mathcal{T}\) |
| \(\mathit{Line\_capex}\) | Line_capex (scalar) |
| \(\mathit{Line\_injection}\) | Line_injection over \(\mathcal{N} \times \Xi \times \mathcal{T}\) |
| \(\check{s}\) | Line_s_monitored over \(\mathcal{K} \times \Xi \times \mathcal{T}\) — the flow in a line's post-contingency rows. It is its flow where it is active, and nothing where it does not. PyPSA builds those rows for every branch of the sub-network in every snapshot |
| \(\mathit{Line\_tech\_capacity\_expansion}\) | Line_tech_capacity_expansion over \(\mathcal{O}\) |
| \(\mathit{Line\_transmission\_expansion\_cost}\) | Line_transmission_expansion_cost over \(\mathcal{O} \times \Xi\) |
| \(\mathit{Line\_transmission\_volume\_expansion}\) | Line_transmission_volume_expansion over \(\mathcal{O} \times \Xi\) |
| \(\mathit{Link\_additions}\) | Link_additions over \(\mathcal{I} \times \mathcal{Y}\) |
| \(\mathit{Link\_capex}\) | Link_capex (scalar) |
| \(\mathit{Link\_commitment\_opex}\) | Link_commitment_opex over \(\Xi\) |
| \(\mathit{Link\_injection}\) | Link_injection over \(\mathcal{N} \times \Xi \times \mathcal{T}\) |
| \(\mathrm{Link\_maintenance\_modules}\) | Link_maintenance_modules over \(\mathcal{L} \times \Xi\) — how many modules of a modular committable link can be on in maintenance: its build cap in modules where the build is extendable, and the modules installed where it is fixed (constraints.py:588-594) |
| \(\mathrm{M}^{f}\) | Link_modules over \(\mathcal{L} \times \Xi\) — how many modules a committed link's status counts: the chosen number where the build is extendable and modular, and the modules installed otherwise (constraints.py:656-661) |
| \(\mathit{Link\_opex}\) | Link_opex over \(\Xi\) |
| \(\overrightarrow{f}\) | Link_output_arrival over \(\mathcal{O} \times \Xi \times \mathcal{T}\) — what a link delivers to an output port at a snapshot. It is its flow delayed by the port's delay within its investment period. It is multiplied by the port's efficiency at the snapshot the flow arrives. Where the port is cyclic_delay, the delayed flow wraps from the period's end. Where it is not, the flow still in transit at the period's first snapshots is lost. A port that does not delay (delay zero) delivers its flow unshifted, cyclic or not |
| \(\overleftarrow{f}^{\circ}\) | Link_p_carried_over over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — the output a link carries over into a snapshot. At the first, it is the p_init it brought in, which is zero where it came in off. After that, it is the previous snapshot's |
| \(\widehat{\mathrm{f}}^{\mathrm{nom}}\) | Link_p_nom_committed over \(\mathcal{L} \times \Xi\) — the build a committed link's ramp rows are taken against: one module where the build is modular, since its status counts modules, and the given build otherwise (constraints.py:1131-1134) |
| \(\widetilde{\mathrm{f}}^{\mathrm{nom}}\) | Link_p_nom_effective over \(\mathcal{L} \times \Xi\) — the build a link's limits are taken against: the chosen one where it is extendable, the given one otherwise |
| \(\overleftarrow{f}\) | Link_previous_p over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — the output a link carries into a snapshot. It is nothing at the first snapshot it is active in past the first of the horizon: PyPSA fills an output it did not build with zero (constraints.py:716). It is the output carried over otherwise |
| \(\overleftarrow{u}^{f}\) | Link_previous_status over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — the commitment state a link carries into a snapshot. It is off at the first snapshot it is active in past the first of the horizon. PyPSA reads a status it did not build as off there (constraints.py:391). It is the state carried over otherwise |
| \(\Delta^{f,-}\) | Link_ramp_down_allowance over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — how far a link may lower flow between two snapshots. It is its ramp limit of the build while it stays on, plus its shut-down ramp in the snapshot it turns off |
| \(\widetilde{\mathrm{rd}}^{f}\) | Link_ramp_down_rate over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — the ramp limit in a link's down row, PyPSA's ramp_limit_down. It is the full build where PyPSA has none, since a shut-down ramp alone builds the row |
| \(\overleftarrow{\bar u}^{f}\) | Link_ramp_previous_status over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — the status a committed link's ramp rows carry into a snapshot. At the first snapshot, a fixed modular link that came in running has every module on, since the state it brought in is on or off (constraints.py:1152-1157). The transition and down-time rows count it as one module (PyPSA/PyPSA#2114). It is the state carried in otherwise |
| \(\Delta^{f,+}\) | Link_ramp_up_allowance over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — how far a link may raise flow between two snapshots. It is its ramp limit of the build while it stays on, plus its start-up ramp in the snapshot it turns on |
| \(\widetilde{\mathrm{ru}}^{f}\) | Link_ramp_up_rate over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — the ramp limit in a link's up row, PyPSA's ramp_limit_up. It is the full build where PyPSA has none, since a start-up ramp alone builds the row |
| \(\widetilde{\mathrm{rd}}^{f,\mathrm{dn}}\) | Link_shut_down_rate over \(\mathcal{L} \times \Xi\) — the shut-down ramp in a link's down row (PyPSA's ramp_limit_shut_down), or the full build where it has none |
| \(\widetilde{\mathrm{ru}}^{f,\mathrm{up}}\) | Link_start_up_rate over \(\mathcal{L} \times \Xi\) — the start-up ramp in a link's up row (PyPSA's ramp_limit_start_up), or the full build where it has none |
| \(\overleftarrow{u}^{\circ f}\) | Link_status_carried_over over \(\mathcal{L} \times \Xi \times \mathcal{T}\) — the state a link carries over into a snapshot: the state it brought into the horizon at the first, the previous snapshot's after that |
| \(\mathit{Link\_tech\_capacity\_expansion}\) | Link_tech_capacity_expansion over \(\mathcal{O}\) |
| \(\mathit{Link\_transmission\_expansion\_cost}\) | Link_transmission_expansion_cost over \(\mathcal{O} \times \Xi\) |
| \(\mathit{Link\_transmission\_volume\_expansion}\) | Link_transmission_volume_expansion over \(\mathcal{O} \times \Xi\) |
| \(\check{\mathrm{load}}\) | Load_demand over \(\mathcal{D} \times \Xi \times \mathcal{T}\) — what a load draws from its bus's balance. It is its demand times its sign where it is active, and nothing where it is not. PyPSA drops an inactive load from the balance (constraints.py:1591-1592) |
| \(\mathrm{Load\_injection}\) | Load_injection over \(\mathcal{N} \times \Xi \times \mathcal{T}\) |
| \(\hat{s}\) | Outage_s over \(\mathcal{K}^{\mathrm{out}} \times \Xi \times \mathcal{T}\) — the flow an outage takes off its branch: the outaged line's or transformer's flow before it goes out |
| \(\mathit{Process\_additions}\) | Process_additions over \(\mathcal{I} \times \mathcal{Y}\) |
| \(\mathit{Process\_capex}\) | Process_capex (scalar) |
| \(\mathit{Process\_commitment\_opex}\) | Process_commitment_opex over \(\Xi\) |
| \(\mathit{Process\_injection}\) | Process_injection over \(\mathcal{N} \times \Xi \times \mathcal{T}\) |
| \(\mathrm{Process\_maintenance\_modules}\) | Process_maintenance_modules over \(\mathcal{J} \times \Xi\) — how many modules of a modular committable process can be on in maintenance: its build cap in modules where the build is extendable, and the modules installed where it is fixed (constraints.py:588-594) |
| \(\mathrm{M}^{z}\) | Process_modules over \(\mathcal{J} \times \Xi\) — how many modules a committed process's status counts: the chosen number where the build is extendable and modular, and the modules installed otherwise (constraints.py:656-661) |
| \(\mathit{Process\_opex}\) | Process_opex over \(\Xi\) |
| \(\overrightarrow{z}\) | Process_output_arrival over \(\mathcal{R} \times \Xi \times \mathcal{T}\) — what a process transfers at a port at a snapshot. It is its internal power delayed by the port's delay within its investment period. It is multiplied by the port's rate at the snapshot the transfer arrives. Where the port is cyclic_delay, the delayed transfer wraps from the period's end. Where it is not, the energy still in transit at the period's first snapshots is lost. A port that does not delay (delay zero) transfers at once, cyclic or not |
| \(\overleftarrow{z}^{\circ}\) | Process_p_carried_over over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — the output a process carries over into a snapshot. At the first, it is the p_init it brought in, which is zero where it came in off. After that, it is the previous snapshot's |
| \(\widehat{\mathrm{z}}^{\mathrm{nom}}\) | Process_p_nom_committed over \(\mathcal{J} \times \Xi\) — the build a committed process's ramp rows are taken against: one module where the build is modular, since its status counts modules, and the given build otherwise (constraints.py:1131-1134) |
| \(\widetilde{\mathrm{z}}^{\mathrm{nom}}\) | Process_p_nom_effective over \(\mathcal{J} \times \Xi\) — the build a process's limits are taken against: the chosen one where it is extendable, the given one otherwise |
| \(\overleftarrow{z}\) | Process_previous_p over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — the output a process carries into a snapshot. It is nothing at the first snapshot it is active in past the first of the horizon: PyPSA fills an output it did not build with zero (constraints.py:716). It is the output carried over otherwise |
| \(\overleftarrow{u}^{z}\) | Process_previous_status over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — the commitment state a process carries into a snapshot. It is off at the first snapshot it is active in past the first of the horizon. PyPSA reads a status it did not build as off there (constraints.py:391). It is the state carried over otherwise |
| \(\Delta^{z,-}\) | Process_ramp_down_allowance over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — how far a process may lower internal power between two snapshots. It is its ramp limit of the build while it stays on, plus its shut-down ramp in the snapshot it turns off |
| \(\widetilde{\mathrm{rd}}^{z}\) | Process_ramp_down_rate over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — the ramp limit in a process's down row, PyPSA's ramp_limit_down. It is the full build where PyPSA has none, since a shut-down ramp alone builds the row |
| \(\overleftarrow{\bar u}^{z}\) | Process_ramp_previous_status over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — the status a committed process's ramp rows carry into a snapshot. At the first snapshot, a fixed modular process that came in running has every module on, since the state it brought in is on or off (constraints.py:1152-1157). The transition and down-time rows count it as one module (PyPSA/PyPSA#2114). It is the state carried in otherwise |
| \(\Delta^{z,+}\) | Process_ramp_up_allowance over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — how far a process may raise internal power between two snapshots. It is its ramp limit of the build while it stays on, plus its start-up ramp in the snapshot it turns on |
| \(\widetilde{\mathrm{ru}}^{z}\) | Process_ramp_up_rate over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — the ramp limit in a process's up row, PyPSA's ramp_limit_up. It is the full build where PyPSA has none, since a start-up ramp alone builds the row |
| \(\widetilde{\mathrm{rd}}^{z,\mathrm{dn}}\) | Process_shut_down_rate over \(\mathcal{J} \times \Xi\) — the shut-down ramp in a process's down row (PyPSA's ramp_limit_shut_down), or the full build where it has none |
| \(\widetilde{\mathrm{ru}}^{z,\mathrm{up}}\) | Process_start_up_rate over \(\mathcal{J} \times \Xi\) — the start-up ramp in a process's up row (PyPSA's ramp_limit_start_up), or the full build where it has none |
| \(\overleftarrow{u}^{\circ z}\) | Process_status_carried_over over \(\mathcal{J} \times \Xi \times \mathcal{T}\) — the state a process carries over into a snapshot: the state it brought into the horizon at the first, the previous snapshot's after that |
| \(\mathit{Process\_tech\_capacity\_expansion}\) | Process_tech_capacity_expansion over \(\mathcal{O}\) |
| \(\mathit{StorageUnit\_additions}\) | StorageUnit_additions over \(\mathcal{I} \times \mathcal{Y}\) |
| \(\mathit{StorageUnit\_capex}\) | StorageUnit_capex (scalar) |
| \(\overleftarrow{\mathit{soc}}\) | StorageUnit_charge_carried_in over \(\Xi \times \mathcal{T} \times \mathcal{S}\) — the charge a unit opens a snapshot with. At the first snapshot it is active in, it is its last such snapshot's less standing loss where it is cyclic. Where it is not, it is the given initial charge, which no standing loss has touched yet. Otherwise it is the previous snapshot's less standing loss. A unit built in a later period opens in that period, and a cyclic one that retires closes on its own last snapshot. Per period, the same holds with each investment period as the horizon |
| \(\mathit{w}^{h}\) | StorageUnit_closing_weight over \(\mathcal{O} \times \Xi \times \mathcal{T} \times \mathcal{S}\) — what the charge a unit holds at a snapshot counts for in a row as its closing level. Where the unit reopens per period, it is the years of the period at the last snapshot of each counted period. Where it does not, it is one at the last counted snapshot it is active in. It is nothing elsewhere |
| \(\mathit{StorageUnit\_injection}\) | StorageUnit_injection over \(\mathcal{N} \times \Xi \times \mathcal{T}\) |
| \(\mathit{last}^{h}\) | StorageUnit_last_counted_active over \(\mathcal{O} \times \Xi \times \mathcal{T} \times \mathcal{S}\) — one at the last snapshot a row counts where a unit is active, and zero elsewhere. A unit that retires before the last counted snapshot closes on its last active level. PyPSA forward-fills the level over the counted snapshots (global_constraints.py:476-479) |
| \(\mathit{StorageUnit\_operational\_limit}\) | StorageUnit_operational_limit over \(\mathcal{O} \times \Xi\) |
| \(\mathit{StorageUnit\_opex}\) | StorageUnit_opex over \(\Xi\) |
| \(\mathit{StorageUnit\_primary\_energy}\) | StorageUnit_primary_energy over \(\mathcal{O} \times \Xi\) |
| \(\mathit{StorageUnit\_tech\_capacity\_expansion}\) | StorageUnit_tech_capacity_expansion over \(\mathcal{O}\) |
| \(\mathit{Store\_additions}\) | Store_additions over \(\mathcal{I} \times \mathcal{Y}\) |
| \(\mathit{Store\_capex}\) | Store_capex (scalar) |
| \(\overline{\mathrm{q}}^{-}\) | Store_charge_cap_pu over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — the most charging through p_store, per unit of e_nom. It is -p_min_pu / max_hours for a store that splits its dispatch, and zero otherwise (stores.py:79-86) |
| \(\mathit{w}^{e}\) | Store_closing_weight over \(\mathcal{O} \times \Xi \times \mathcal{T} \times \mathcal{V}\) — what the energy a store holds at a snapshot counts for in a row as its closing level. Where the store reopens per period, it is the years of the period at the last snapshot of each counted period. Where it does not, it is one at the last counted snapshot it is active in. It is nothing elsewhere |
| \(\underline{\mathrm{q}}^{+}\) | Store_discharge_floor_pu over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — the least discharge, per unit of e_nom. It is zero for a store that splits its dispatch, so p_store carries the charge. Otherwise it is the most charging as a negative discharge, p_min_pu / max_hours (stores.py:79-86) |
| \(\widetilde{\mathrm{e}}^{\mathrm{nom}}\) | Store_e_nom_effective over \(\Xi \times \mathcal{V}\) — the build a store's limits are taken against: the chosen one where it is extendable, the given one otherwise |
| \(\overleftarrow{e}\) | Store_energy_carried_in over \(\Xi \times \mathcal{T} \times \mathcal{V}\) — the energy a store opens a snapshot with. At the first snapshot it is active in, it is its last such snapshot's less standing loss where it is cyclic. Where it is not, it is the given initial energy, which no standing loss has touched yet. Otherwise it is the previous snapshot's less standing loss. A store built in a later period opens in that period, and a cyclic one that retires closes on its own last snapshot. Per period, the same holds with each investment period as the horizon |
| \(\mathit{Store\_injection}\) | Store_injection over \(\mathcal{N} \times \Xi \times \mathcal{T}\) |
| \(\mathit{last}^{e}\) | Store_last_counted_active over \(\mathcal{O} \times \Xi \times \mathcal{T} \times \mathcal{V}\) — one at the last snapshot a row counts where a store is active, and zero elsewhere. A store that retires before the last counted snapshot closes on its last active level. PyPSA forward-fills the level over the counted snapshots (global_constraints.py:530-533) |
| \(\mathit{Store\_operational\_limit}\) | Store_operational_limit over \(\mathcal{O} \times \Xi\) |
| \(\mathit{Store\_opex}\) | Store_opex over \(\Xi\) |
| \(\mathit{Store\_primary\_energy}\) | Store_primary_energy over \(\mathcal{O} \times \Xi\) |
| \(\mathit{Store\_tech\_capacity\_expansion}\) | Store_tech_capacity_expansion over \(\mathcal{O}\) |
| \(\mathit{Transformer\_angle\_sum}\) | Transformer_angle_sum over \(\mathcal{C} \times \Xi \times \mathcal{T}\) |
| \(\mathit{Transformer\_capex}\) | Transformer_capex (scalar) |
| \(\mathit{Transformer\_injection}\) | Transformer_injection over \(\mathcal{N} \times \Xi \times \mathcal{T}\) |
| \(\check{\sigma}\) | Transformer_s_monitored over \(\Xi \times \mathcal{T} \times \mathcal{M}\) — the flow a transformer's post-contingency rows read, as a line's |
| \(\mathit{w}^{E}\) | energy_limit_weight over \(\mathcal{T}\) — what one unit of power at a snapshot counts for in an e_sum row: the generator weighting times the years of the snapshot's period, as in an operational_limit row (constraints.py:2935-2937) |
| \(\mathit{operational\_limit}\) | operational_limit over \(\mathcal{O} \times \Xi\) — what an operational_limit row totals: the weighted energy its generators deliver over the snapshots it counts, plus what its non-cyclic storage draws down. The initial charge it draws from is folded into the row's constant |
| \(\mathit{primary\_energy}\) | primary_energy over \(\mathcal{O} \times \Xi\) — what a primary_energy row totals: weighted generator energy over the snapshots it counts, less the charge left in weighted storage at the close. The initial charge it is compared against is folded into the row's constant |
| \(\mathit{risk\_weighted\_opex}\) | risk_weighted_opex (scalar) |
| \(\mathit{scenario\_opex}\) | scenario_opex over \(\Xi\) — what a future costs to run: every operating term, weighted by the snapshot's hours and its period, before the scenario's own weight. A start and a stop cost what they cost, unweighted, as PyPSA adds them (optimize.py:448-468) |
| \(\mathit{tech\_capacity\_expansion}\) | tech_capacity_expansion over \(\mathcal{O}\) — what a tech_capacity_expansion_limit row totals: the chosen build of the row's carrier-and-bus set |
| \(\mathit{total\_cost}\) | total_cost (scalar) — what the system costs. Capacity counts once per active period at its expected cost over the scenarios. Operation counts in expectation over the scenarios, and a share of it at the tail |
| \(\mathit{transmission\_expansion\_cost}\) | transmission_expansion_cost over \(\mathcal{O} \times \Xi\) — what a transmission_expansion_cost_limit row totals: capital cost times the chosen build of the row's branches |
| \(\mathit{transmission\_volume\_expansion}\) | transmission_volume_expansion over \(\mathcal{O} \times \Xi\) — what a transmission_volume_expansion_limit row totals: length times the chosen build of the row's branches |
Notation#
| Symbol | Meaning |
|---|---|
| \(t \ominus k\) | edge='wrap' — \(t-k\), cyclic |
| \(t \boxminus_{v} k\) | edge=v — \(t-k\), and \(v\) past the edge |
| \(t \ominus^{\mathrm{relation}(t)} k\) | within=relation[...] — \(t-k\) inside the group of \(t\) |
| \(\mathrm{pos}(t)\) | position() — the position of \(t\), counted from \(0\); \(t\) is the label |
| \(\mathrm{pos}_{\mathrm{relation}(t)}(t)\) | position(within=relation[...]) — the position of \(t\) inside its group |
| \(\lvert \mathcal{T} \rvert\) | the size of \(\mathcal{T}\); \(\lvert \mathcal{T} \rvert - 1\) is the last position |
Generator_com_ext#
Generator_com_ext:
description: >-
a committable generator with an extendable, non-modular build that
is active in the snapshot's period. These are PyPSA's `com-ext` rows, whose
status is relaxed against the chosen build
condition: >-
Generator_committable
AND Generator_p_nom_extendable
AND NOT (Generator_p_nom_mod > 0)
AND Generator_active
\[
\mathrm{on}^{\mathrm{com,ext}}_{g,t} \iff \mathrm{com}_{g} \wedge \mathrm{ext}_{g} \wedge \neg \left( \mathrm{p}^{\mathrm{mod}}_{g} > 0 \right) \wedge \mathrm{on}_{g,t} \qquad \forall\, g \in \mathcal{G},\ t \in \mathcal{T}
\]
Generator_committed#
Generator_committed:
description: >-
a committable generator that is active in the snapshot's period: every
unit-commitment row's set
condition: Generator_committable AND Generator_active
\[
\mathrm{on}^{\mathrm{com}}_{g,t} \iff \mathrm{com}_{g} \wedge \mathrm{on}_{g,t} \qquad \forall\, g \in \mathcal{G},\ t \in \mathcal{T}
\]
Generator_maint_ext#
Generator_maint_ext:
description: >-
a maintainable generator with an extendable build that is active in the
snapshot's period, unless it is committable and modular. These are the
maintenance rows against the chosen build
condition: >-
Generator_maintainable
AND Generator_p_nom_extendable
AND NOT (Generator_committable AND Generator_p_nom_mod > 0)
AND Generator_active
\[
\mathrm{on}^{\mathrm{mnt,ext}}_{g,t} \iff \mathrm{mnt}_{g} \wedge \mathrm{ext}_{g} \wedge \neg \left( \mathrm{com}_{g} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0 \right) \wedge \mathrm{on}_{g,t} \qquad \forall\, g \in \mathcal{G},\ t \in \mathcal{T}
\]
Generator_marginal_cost_as_lp#
Generator_marginal_cost_as_lp:
description: >-
linopy's choice of method for all generator cost curves at once
(`method="auto"`): the segment lines where every curve is convex and no
generator with a curve is committable, the incremental rows otherwise.
PyPSA gates the curves by `Generator-status` where one is committable,
and linopy states no segment lines under a gate (`optimize.py:314-318`,
`piecewise.py:200-224`)
condition: >-
count(Generator_marginal_cost_curve AND count(Generator_marginal_cost_curve_bends_down, over=breakpoint) > 0, over=generator) == 0
AND count(Generator_marginal_cost_curve AND Generator_committable, over=generator) == 0
\[
\mathrm{lp} \iff \lvert \{ g \in \mathcal{G} \,:\, \mathrm{curve}_{g} \wedge \lvert \{ b \in \mathcal{B} \,:\, \mathrm{bend}_{b,g} \} \rvert > 0 \} \rvert = 0 \wedge \lvert \{ g \in \mathcal{G} \,:\, \mathrm{curve}_{g} \wedge \mathrm{com}_{g} \} \rvert = 0
\]
Generator_marginal_cost_curve#
Generator_marginal_cost_curve:
description: >-
a generator that pays along a piecewise `marginal_cost` curve: one of
PyPSA's `active_assets` with curve data (`optimize.py:307`,
`piecewise.py:230-257`)
condition: count(Generator_marginal_cost_curve_p, over=breakpoint) > 0 AND Generator_active_asset
\[
\mathrm{curve}_{g} \iff \lvert \{ b \in \mathcal{B} \,:\, \mathrm{p}^{\mathrm{curve}}_{b,g} \text{ is defined} \} \rvert > 0 \wedge \mathrm{act}_{g} \qquad \forall\, g \in \mathcal{G}
\]
Generator_marginal_cost_curve_bends_down#
Generator_marginal_cost_curve_bends_down:
description: >-
a breakpoint where a generator's cost curve gets less steep: the
segment after it rises less per unit of output than the one before
condition: >-
Generator_marginal_cost_curve_p
AND shift(Generator_marginal_cost_curve_p, along=breakpoint, offset=1)
AND shift(Generator_marginal_cost_curve_p, along=breakpoint, offset=-1)
AND (Generator_marginal_cost_curve_cost - shift(Generator_marginal_cost_curve_cost, along=breakpoint, offset=1, edge=0))
* (shift(Generator_marginal_cost_curve_p, along=breakpoint, offset=-1, edge=0) - Generator_marginal_cost_curve_p)
> (shift(Generator_marginal_cost_curve_cost, along=breakpoint, offset=-1, edge=0) - Generator_marginal_cost_curve_cost)
* (Generator_marginal_cost_curve_p - shift(Generator_marginal_cost_curve_p, along=breakpoint, offset=1, edge=0))
\[
\mathrm{bend}_{b,g} \iff \mathrm{p}^{\mathrm{curve}}_{b,g} \text{ is defined} \wedge \mathrm{p}^{\mathrm{curve}}_{b - 1,g} \text{ is defined} \wedge \mathrm{p}^{\mathrm{curve}}_{b + 1,g} \text{ is defined} \wedge \left( \mathrm{C}^{\mathrm{curve}}_{b,g} - \mathrm{C}^{\mathrm{curve}}_{b \boxminus_{0} 1,g} \right) \cdot \left( \mathrm{p}^{\mathrm{curve}}_{b \boxplus_{0} 1,g} - \mathrm{p}^{\mathrm{curve}}_{b,g} \right) > \left( \mathrm{C}^{\mathrm{curve}}_{b \boxplus_{0} 1,g} - \mathrm{C}^{\mathrm{curve}}_{b,g} \right) \cdot \left( \mathrm{p}^{\mathrm{curve}}_{b,g} - \mathrm{p}^{\mathrm{curve}}_{b \boxminus_{0} 1,g} \right) \qquad \forall\, b \in \mathcal{B},\ g \in \mathcal{G}
\]
Generator_ramps_from_previous#
Generator_ramps_from_previous:
description: >-
a snapshot whose ramp uses a previous output. It is any snapshot but its
period's first, and the horizon's first where the generator comes in off
or carries an initial output
condition: >-
position(snapshot, within=snapshot_period[period]) > 0 OR (position(snapshot) == 0 AND
(Generator_status_initial == 0 OR Generator_p_init))
\[
\mathrm{prev}_{g,\xi,t} \iff \mathrm{pos}_{\mathrm{snapshot\_period}(t)}(t) > 0 \vee \mathrm{pos}(t) = 0 \wedge \left( \mathrm{u}^{0}_{g,\xi} = 0 \vee \mathrm{p}^{0}_{g,\xi} \text{ is defined} \right) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Line_lossy#
Line_lossy:
description: >-
a line that is active in the snapshot's period, where the run models
transmission losses
condition: transmission_losses AND Line_active
\[
\mathrm{on}^{s,\mathrm{lossy}}_{k,t} \iff \mathrm{lossy} \wedge \mathrm{on}^{s}_{k,t} \qquad \forall\, k \in \mathcal{K},\ t \in \mathcal{T}
\]
Link_com_ext#
Link_com_ext:
description: >-
a committable link with an extendable, non-modular build that is active in
the snapshot's period. These are PyPSA's `com-ext` rows, whose status is
relaxed against the chosen build
condition: >-
Link_committable
AND Link_p_nom_extendable
AND NOT (Link_p_nom_mod > 0)
AND Link_active
\[
\mathrm{on}^{f,\mathrm{com,ext}}_{l,t} \iff \mathrm{com}^{f}_{l} \wedge \mathrm{ext}^{f}_{l} \wedge \neg \left( \mathrm{f}^{\mathrm{mod}}_{l} > 0 \right) \wedge \mathrm{on}^{f}_{l,t} \qquad \forall\, l \in \mathcal{L},\ t \in \mathcal{T}
\]
Link_committed#
Link_committed:
description: >-
a committable link that is active in the snapshot's period: every
unit-commitment row's set
condition: Link_committable AND Link_active
\[
\mathrm{on}^{f,\mathrm{com}}_{l,t} \iff \mathrm{com}^{f}_{l} \wedge \mathrm{on}^{f}_{l,t} \qquad \forall\, l \in \mathcal{L},\ t \in \mathcal{T}
\]
Link_maint_ext#
Link_maint_ext:
description: >-
a maintainable link with an extendable build that is active in the
snapshot's period, unless it is committable and modular. These are the
maintenance rows against the chosen build
condition: >-
Link_maintainable
AND Link_p_nom_extendable
AND NOT (Link_committable AND Link_p_nom_mod > 0)
AND Link_active
\[
\mathrm{on}^{f,\mathrm{mnt,ext}}_{l,t} \iff \mathrm{mnt}^{f}_{l} \wedge \mathrm{ext}^{f}_{l} \wedge \neg \left( \mathrm{com}^{f}_{l} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0 \right) \wedge \mathrm{on}^{f}_{l,t} \qquad \forall\, l \in \mathcal{L},\ t \in \mathcal{T}
\]
Link_ramps_from_previous#
Link_ramps_from_previous:
description: >-
a snapshot whose ramp uses a previous output. It is any snapshot but its
period's first, and the horizon's first where the link comes in off or
carries an initial output
condition: >-
position(snapshot, within=snapshot_period[period]) > 0 OR (position(snapshot) == 0 AND
(Link_status_initial == 0 OR Link_p_init))
\[
\mathrm{prev}^{f}_{l,\xi,t} \iff \mathrm{pos}_{\mathrm{snapshot\_period}(t)}(t) > 0 \vee \mathrm{pos}(t) = 0 \wedge \left( \mathrm{u}^{f,0}_{l,\xi} = 0 \vee \mathrm{f}^{0}_{l,\xi} \text{ is defined} \right) \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Process_com_ext#
Process_com_ext:
description: >-
a committable process with an extendable, non-modular build that is active
in the snapshot's period. These are PyPSA's `com-ext` rows, whose status
is relaxed against the chosen build
condition: >-
Process_committable
AND Process_p_nom_extendable
AND NOT (Process_p_nom_mod > 0)
AND Process_active
\[
\mathrm{on}^{z,\mathrm{com,ext}}_{j,t} \iff \mathrm{com}^{z}_{j} \wedge \mathrm{ext}^{z}_{j} \wedge \neg \left( \mathrm{z}^{\mathrm{mod}}_{j} > 0 \right) \wedge \mathrm{on}^{z}_{j,t} \qquad \forall\, j \in \mathcal{J},\ t \in \mathcal{T}
\]
Process_committed#
Process_committed:
description: >-
a committable process that is active in the snapshot's period: every
unit-commitment row's set
condition: Process_committable AND Process_active
\[
\mathrm{on}^{z,\mathrm{com}}_{j,t} \iff \mathrm{com}^{z}_{j} \wedge \mathrm{on}^{z}_{j,t} \qquad \forall\, j \in \mathcal{J},\ t \in \mathcal{T}
\]
Process_maint_ext#
Process_maint_ext:
description: >-
a maintainable process with an extendable build that is active in the
snapshot's period, unless it is committable and modular. These are the
maintenance rows against the chosen build
condition: >-
Process_maintainable
AND Process_p_nom_extendable
AND NOT (Process_committable AND Process_p_nom_mod > 0)
AND Process_active
\[
\mathrm{on}^{z,\mathrm{mnt,ext}}_{j,t} \iff \mathrm{mnt}^{z}_{j} \wedge \mathrm{ext}^{z}_{j} \wedge \neg \left( \mathrm{com}^{z}_{j} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0 \right) \wedge \mathrm{on}^{z}_{j,t} \qquad \forall\, j \in \mathcal{J},\ t \in \mathcal{T}
\]
Process_ramps_from_previous#
Process_ramps_from_previous:
description: >-
a snapshot whose ramp uses a previous output. It is any snapshot but its
period's first, and the horizon's first where the process comes in off
or carries an initial output
condition: >-
position(snapshot, within=snapshot_period[period]) > 0 OR (position(snapshot) == 0 AND
(Process_status_initial == 0 OR Process_p_init))
\[
\mathrm{prev}^{z}_{j,\xi,t} \iff \mathrm{pos}_{\mathrm{snapshot\_period}(t)}(t) > 0 \vee \mathrm{pos}(t) = 0 \wedge \left( \mathrm{u}^{z,0}_{j,\xi} = 0 \vee \mathrm{z}^{0}_{j,\xi} \text{ is defined} \right) \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
StorageUnit_ext#
StorageUnit_ext:
description: >-
a storage unit with an extendable build that is active in the snapshot's
period. It is PyPSA's `ext` rows
condition: StorageUnit_p_nom_extendable AND StorageUnit_active
\[
\mathrm{on}^{h,\mathrm{ext}}_{t,s} \iff \mathrm{ext}^{h}_{s} \wedge \mathrm{on}^{h}_{t,s} \qquad \forall\, t \in \mathcal{T},\ s \in \mathcal{S}
\]
StorageUnit_fix#
StorageUnit_fix:
description: >-
a storage unit with a fixed build that is active in the snapshot's period.
It is PyPSA's `fix` rows
condition: NOT StorageUnit_p_nom_extendable AND StorageUnit_active
\[
\mathrm{on}^{h,\mathrm{fix}}_{t,s} \iff \neg \mathrm{ext}^{h}_{s} \wedge \mathrm{on}^{h}_{t,s} \qquad \forall\, t \in \mathcal{T},\ s \in \mathcal{S}
\]
Transformer_lossy#
Transformer_lossy:
description: >-
a transformer that is active in the snapshot's period, where the run models
transmission losses
condition: transmission_losses AND Transformer_active
\[
\mathrm{on}^{\sigma,\mathrm{lossy}}_{t,m} \iff \mathrm{lossy} \wedge \mathrm{on}^{\sigma}_{t,m} \qquad \forall\, t \in \mathcal{T},\ m \in \mathcal{M}
\]
Variable domains#
CVaR
\[
CVaR \in \mathbb{R} \qquad \text{where } \omega \text{ is defined}
\]
CVaR_a
\[
a_{\xi} \ge 0 \qquad \forall\, \xi \in \Xi \,:\, \omega \text{ is defined}
\]
CVaR_theta
\[
\theta \in \mathbb{R} \qquad \text{where } \omega \text{ is defined}
\]
FlowBased_net_position
\[
\mathit{NP}_{n,\xi,t} \in \mathbb{R} \qquad \forall\, n \in \mathcal{N},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{zone}_{n}
\]
Generator_maintenance
\[
0 \le \mu_{g,\xi,t} \le 1 \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}_{g} \wedge \mathrm{on}_{g,t}
\]
Generator_maintenance_capacity
\[
\mu^{\mathrm{nom}}_{g,\xi,t} \ge 0 \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{mnt,ext}}_{g,t}
\]
Generator_maintenance_start
\[
\mu^{\mathrm{up}}_{g,\xi,t} \in \{0, 1\} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}_{g} \wedge \mathrm{on}_{g,t}
\]
Generator_maintenance_status
\[
\mu^{u}_{g,\xi,t} \ge 0 \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}_{g} \wedge \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \neg \left( \mathrm{ext}_{g} \wedge \neg \left( \mathrm{p}^{\mathrm{mod}}_{g} > 0 \right) \right)
\]
Generator_marginal_cost_piecewise
\[
C^{\mathrm{curve}}_{g,\xi,t} \in \mathbb{R} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{curve}_{g} \wedge \mathrm{on}_{g,t}
\]
Generator_n_mod
\[
N_{g} \ge 0, N_{g} \in \mathbb{Z} \qquad \forall\, g \in \mathcal{G} \,:\, \mathrm{ext}_{g} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0 \wedge \mathrm{act}_{g}
\]
Generator_p
\[
p_{g,\xi,t} \in \mathbb{R} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}_{g,t}
\]
Generator_p_nom_ext
\[
P_{g} \in \mathbb{R} \qquad \forall\, g \in \mathcal{G} \,:\, \mathrm{ext}_{g} \wedge \mathrm{act}_{g}
\]
Generator_shut_down
\[
\mathit{dn}_{g,\xi,t} \ge 0, \mathit{dn}_{g,\xi,t} \in \mathbb{Z} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t}
\]
Generator_start_up
\[
\mathit{up}_{g,\xi,t} \ge 0, \mathit{up}_{g,\xi,t} \in \mathbb{Z} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t}
\]
Generator_status
\[
u_{g,\xi,t} \ge 0, u_{g,\xi,t} \in \mathbb{Z} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t}
\]
Line_loss
\[
\ell_{k,\xi,t} \ge 0 \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{s,\mathrm{lossy}}_{k,t}
\]
Line_n_mod
\[
N^{s}_{k} \ge 0, N^{s}_{k} \in \mathbb{Z} \qquad \forall\, k \in \mathcal{K} \,:\, \mathrm{ext}^{s}_{k} \wedge \mathrm{s}^{\mathrm{mod}}_{k} > 0 \wedge \mathrm{act}^{s}_{k}
\]
Line_s
\[
s_{k,\xi,t} \in \mathbb{R} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{s}_{k,t}
\]
Line_s_nom_ext
\[
S_{k} \in \mathbb{R} \qquad \forall\, k \in \mathcal{K} \,:\, \mathrm{ext}^{s}_{k} \wedge \mathrm{act}^{s}_{k}
\]
Link_maintenance
\[
0 \le \mu^{f}_{l,\xi,t} \le 1 \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{f}_{l} \wedge \mathrm{on}^{f}_{l,t}
\]
Link_maintenance_capacity
\[
\mu^{f,\mathrm{nom}}_{l,\xi,t} \ge 0 \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{mnt,ext}}_{l,t}
\]
Link_maintenance_start
\[
\mu^{f,\mathrm{up}}_{l,\xi,t} \in \{0, 1\} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{f}_{l} \wedge \mathrm{on}^{f}_{l,t}
\]
Link_maintenance_status
\[
\mu^{f,u}_{l,\xi,t} \ge 0 \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{f}_{l} \wedge \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \neg \left( \mathrm{ext}^{f}_{l} \wedge \neg \left( \mathrm{f}^{\mathrm{mod}}_{l} > 0 \right) \right)
\]
Link_n_mod
\[
N^{f}_{l} \ge 0, N^{f}_{l} \in \mathbb{Z} \qquad \forall\, l \in \mathcal{L} \,:\, \mathrm{ext}^{f}_{l} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0 \wedge \mathrm{act}^{f}_{l}
\]
Link_p
\[
f_{l,\xi,t} \in \mathbb{R} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f}_{l,t}
\]
Link_p_nom_ext
\[
F_{l} \in \mathbb{R} \qquad \forall\, l \in \mathcal{L} \,:\, \mathrm{ext}^{f}_{l} \wedge \mathrm{act}^{f}_{l}
\]
Link_shut_down
\[
\mathit{dn}^{f}_{l,\xi,t} \ge 0, \mathit{dn}^{f}_{l,\xi,t} \in \mathbb{Z} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t}
\]
Link_start_up
\[
\mathit{up}^{f}_{l,\xi,t} \ge 0, \mathit{up}^{f}_{l,\xi,t} \in \mathbb{Z} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t}
\]
Link_status
\[
u^{f}_{l,\xi,t} \ge 0, u^{f}_{l,\xi,t} \in \mathbb{Z} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t}
\]
Process_maintenance
\[
0 \le \mu^{z}_{j,\xi,t} \le 1 \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{z}_{j} \wedge \mathrm{on}^{z}_{j,t}
\]
Process_maintenance_capacity
\[
\mu^{z,\mathrm{nom}}_{j,\xi,t} \ge 0 \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{mnt,ext}}_{j,t}
\]
Process_maintenance_start
\[
\mu^{z,\mathrm{up}}_{j,\xi,t} \in \{0, 1\} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{z}_{j} \wedge \mathrm{on}^{z}_{j,t}
\]
Process_maintenance_status
\[
\mu^{z,u}_{j,\xi,t} \ge 0 \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{z}_{j} \wedge \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \neg \left( \mathrm{ext}^{z}_{j} \wedge \neg \left( \mathrm{z}^{\mathrm{mod}}_{j} > 0 \right) \right)
\]
Process_n_mod
\[
N^{z}_{j} \ge 0, N^{z}_{j} \in \mathbb{Z} \qquad \forall\, j \in \mathcal{J} \,:\, \mathrm{ext}^{z}_{j} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0 \wedge \mathrm{act}^{z}_{j}
\]
Process_p
\[
z_{j,\xi,t} \in \mathbb{R} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z}_{j,t}
\]
Process_p_nom_ext
\[
Z_{j} \in \mathbb{R} \qquad \forall\, j \in \mathcal{J} \,:\, \mathrm{ext}^{z}_{j} \wedge \mathrm{act}^{z}_{j}
\]
Process_shut_down
\[
\mathit{dn}^{z}_{j,\xi,t} \ge 0, \mathit{dn}^{z}_{j,\xi,t} \in \mathbb{Z} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t}
\]
Process_start_up
\[
\mathit{up}^{z}_{j,\xi,t} \ge 0, \mathit{up}^{z}_{j,\xi,t} \in \mathbb{Z} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t}
\]
Process_status
\[
u^{z}_{j,\xi,t} \ge 0, u^{z}_{j,\xi,t} \in \mathbb{Z} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t}
\]
StorageUnit_n_mod
\[
N^{h}_{s} \ge 0, N^{h}_{s} \in \mathbb{Z} \qquad \forall\, s \in \mathcal{S} \,:\, \mathrm{ext}^{h}_{s} \wedge \mathrm{h}^{\mathrm{mod}}_{s} > 0 \wedge \mathrm{act}^{h}_{s}
\]
StorageUnit_p_dispatch
\[
h^{+}_{\xi,t,s} \in \mathbb{R} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h}_{t,s}
\]
StorageUnit_p_nom_ext
\[
H_{s} \in \mathbb{R} \qquad \forall\, s \in \mathcal{S} \,:\, \mathrm{ext}^{h}_{s} \wedge \mathrm{act}^{h}_{s}
\]
StorageUnit_p_store
\[
h^{-}_{\xi,t,s} \in \mathbb{R} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h}_{t,s}
\]
StorageUnit_spill
\[
0 \le \mathit{spill}_{\xi,t,s} \le \mathrm{inflow}_{\xi,t,s} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{inflow}_{\xi,t,s} > 0 \wedge \mathrm{on}^{h}_{t,s}
\]
StorageUnit_state_of_charge
\[
\mathit{soc}_{\xi,t,s} \in \mathbb{R} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h}_{t,s}
\]
Store_e
\[
e_{\xi,t,v} \in \mathbb{R} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{on}^{e}_{t,v}
\]
Store_e_nom_ext
\[
E_{v} \in \mathbb{R} \qquad \forall\, v \in \mathcal{V} \,:\, \mathrm{ext}^{e}_{v} \wedge \mathrm{act}^{e}_{v}
\]
Store_n_mod
\[
N^{e}_{v} \ge 0, N^{e}_{v} \in \mathbb{Z} \qquad \forall\, v \in \mathcal{V} \,:\, \mathrm{ext}^{e}_{v} \wedge \mathrm{e}^{\mathrm{mod}}_{v} > 0 \wedge \mathrm{act}^{e}_{v}
\]
Store_p
\[
q_{\xi,t,v} \in \mathbb{R} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{on}^{e}_{t,v}
\]
Store_p_store
\[
q^{-}_{\xi,t,v} \in \mathbb{R} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{on}^{e}_{t,v}
\]
Store_spill
\[
0 \le \mathit{spill}^{e}_{\xi,t,v} \le \mathrm{inflow}^{e}_{\xi,t,v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{inflow}^{e}_{\xi,t,v} > 0 \wedge \mathrm{on}^{e}_{t,v}
\]
Transformer_loss
\[
\ell^{\sigma}_{\xi,t,m} \ge 0 \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \mathrm{on}^{\sigma,\mathrm{lossy}}_{t,m}
\]
Transformer_n_mod
\[
N^{\sigma}_{m} \ge 0, N^{\sigma}_{m} \in \mathbb{Z} \qquad \forall\, m \in \mathcal{M} \,:\, \mathrm{ext}^{\sigma}_{m} \wedge \sigma^{\mathrm{mod}}_{m} > 0 \wedge \mathrm{act}^{\sigma}_{m}
\]
Transformer_phase_shift
\[
\mathrm{Transformer\_phase\_shift\_min}_{\xi,m} \le \mathit{Transformer\_phase\_shift}_{\xi,t,m} \le \mathrm{Transformer\_phase\_shift\_max}_{\xi,m} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \mathrm{Transformer\_phase\_shift\_varying}_{\xi,m} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Transformer_s
\[
\sigma_{\xi,t,m} \in \mathbb{R} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \mathrm{on}^{\sigma}_{t,m}
\]
Transformer_s_nom_ext
\[
\Sigma_{m} \in \mathbb{R} \qquad \forall\, m \in \mathcal{M} \,:\, \mathrm{ext}^{\sigma}_{m} \wedge \mathrm{act}^{\sigma}_{m}
\]
Bus_injection#
Bus_injection:
dims: [scenario, snapshot, bus]
expression: >-
FlowBased_injection
+ Generator_injection
+ Line_injection
+ Link_injection
+ Load_injection
+ Process_injection
+ StorageUnit_injection
+ Store_injection
+ Transformer_injection
description: >-
what every component puts into a bus, less what it takes out of it.
PyPSA writes each term into the balance, and a load on its right-hand
side
\[
\mathit{Bus\_injection}_{n,\xi,t} = \mathit{NP}^{\mathrm{inj}}_{n,\xi,t} + \mathit{Generator\_injection}_{n,\xi,t} + \mathit{Line\_injection}_{n,\xi,t} + \mathit{Link\_injection}_{n,\xi,t} + \mathrm{Load\_injection}_{n,\xi,t} + \mathit{Process\_injection}_{n,\xi,t} + \mathit{StorageUnit\_injection}_{n,\xi,t} + \mathit{Store\_injection}_{n,\xi,t} + \mathit{Transformer\_injection}_{n,\xi,t} \qquad \forall\, n \in \mathcal{N},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Carrier_additions#
Carrier_additions:
dims: [period, carrier]
expression: >-
Generator_additions
+ Line_additions
+ Link_additions
+ Process_additions
+ StorageUnit_additions
+ Store_additions
description: >-
what a carrier adds in a period: every extendable component of that
carrier, counting each build in the first period it is active in. Like
PyPSA, it sums only the components that carry a carrier attribute, so a
transformer, which has none, counts in no carrier
\[
\mathit{Carrier\_additions}_{i,y} = \mathit{Generator\_additions}_{i,y} + \mathit{Line\_additions}_{i,y} + \mathit{Link\_additions}_{i,y} + \mathit{Process\_additions}_{i,y} + \mathit{StorageUnit\_additions}_{i,y} + \mathit{Store\_additions}_{i,y} \qquad \forall\, i \in \mathcal{I},\ y \in \mathcal{Y}
\]
Carrier_relative_growth#
Carrier_relative_growth:
description: >-
the share of the previous period's additions in a carrier's growth
limit, PyPSA's `max_relative_growth` clipped at zero. A negative share
thus adds nothing and never tightens the limit
dims: [carrier]
cases:
positive: { when: Carrier_max_relative_growth > 0, expression: Carrier_max_relative_growth }
otherwise: 0
\[
\mathrm{r}^{+}_{i} = \begin{cases} \mathrm{r}_{i} & \text{if } \mathrm{r}_{i} > 0 \\ 0 & \text{otherwise} \end{cases} \qquad \forall\, i \in \mathcal{I}
\]
Cycle_angle_sum#
Cycle_angle_sum:
dims: [scenario, snapshot, cycle]
expression: Line_angle_sum + Transformer_angle_sum
description: >-
the voltage angle differences around a cycle: every branch flow times
its cycle weight, and every transformer phase shift
\[
\mathit{Cycle\_angle\_sum}_{c,\xi,t} = \mathit{Line\_angle\_sum}_{c,\xi,t} + \mathit{Transformer\_angle\_sum}_{c,\xi,t} \qquad \forall\, c \in \mathcal{C},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
FlowBased_injection#
FlowBased_injection:
description: what a zone bus exports to its flow-based domain leaves the bus (`constraints.py:1578`)
dims: [scenario, snapshot, bus]
expression: -FlowBased_net_position
\[
\mathit{NP}^{\mathrm{inj}}_{n,\xi,t} = -\mathit{NP}_{n,\xi,t} \qquad \forall\, n \in \mathcal{N},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_additions#
Generator_additions:
expression: >-
sum(Generator_p_nom_ext * Generator_first_active, over=generator, by=Generator_carrier[carrier])
\[
\mathit{Generator\_additions}_{i,y} = \sum_{g \in \mathcal{G} \,:\, \mathrm{Generator\_carrier}(g) = i} P_{g} \cdot \mathrm{new}_{g,y} \qquad \forall\, i \in \mathcal{I},\ y \in \mathcal{Y}
\]
Generator_capex#
Generator_capex:
expression: sum(scenario_weight * Generator_p_nom_ext * Generator_capital_cost * Generator_capital_weight)
\[
\mathit{Generator\_capex} = \sum_{g \in \mathcal{G},\ \xi \in \Xi} \pi_{\xi} \cdot P_{g} \cdot \mathrm{c}^{\mathrm{cap}}_{g,\xi} \cdot \mathrm{W}_{g}
\]
Generator_commitment_opex#
Generator_commitment_opex:
expression: >-
sum(sum(((Generator_status * Generator_stand_by_cost) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=generator), over=snapshot)
+ sum(sum((Generator_start_up * Generator_start_up_cost) * at(period_weight_objective, by=snapshot_period[period]), over=generator), over=snapshot)
+ sum(sum((Generator_shut_down * Generator_shut_down_cost) * at(period_weight_objective, by=snapshot_period[period]), over=generator), over=snapshot)
\[
\mathit{Generator\_commitment\_opex}_{\xi} = \sum_{t \in \mathcal{T}} \sum_{g \in \mathcal{G}} u_{g,\xi,t} \cdot \mathrm{c}^{\mathrm{on}}_{g,\xi,t} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{g \in \mathcal{G}} \mathit{up}_{g,\xi,t} \cdot \mathrm{c}^{\mathrm{up}}_{g,\xi,t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{g \in \mathcal{G}} \mathit{dn}_{g,\xi,t} \cdot \mathrm{c}^{\mathrm{dn}}_{g,\xi,t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} \qquad \forall\, \xi \in \Xi
\]
Generator_injection#
Generator_injection:
expression: sum(Generator_sign * Generator_p, over=generator, by=Generator_bus[bus])
\[
\mathit{Generator\_injection}_{n,\xi,t} = \sum_{g \in \mathcal{G} \,:\, \mathrm{Generator\_bus}(g) = n} \mathrm{sgn}_{g} \cdot p_{g,\xi,t} \qquad \forall\, n \in \mathcal{N},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_maintenance_modules#
Generator_maintenance_modules:
description: >-
how many modules of a modular committable generator can be on in
maintenance: its build cap in modules where the build is extendable,
and the modules installed where it is fixed (`constraints.py:588-594`)
dims: [scenario, generator]
cases:
extendable: { when: Generator_p_nom_extendable, expression: Generator_p_nom_max / Generator_p_nom_mod }
otherwise: Generator_modules_installed
\[
\mathrm{Generator\_maintenance\_modules}_{g,\xi} = \begin{cases} \frac{\overline{\mathrm{p}}^{\mathrm{nom}}_{g,\xi}}{\mathrm{p}^{\mathrm{mod}}_{g}} & \text{if } \mathrm{ext}_{g} \\ \mathrm{N}^{\mathrm{fix}}_{g,\xi} & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi
\]
Generator_modules#
Generator_modules:
description: >-
how many modules a committed unit's status counts: the chosen number
where the build is extendable and modular, and the modules installed
otherwise (`constraints.py:656-661`)
dims: [scenario, generator]
cases:
chosen: { when: Generator_p_nom_extendable AND Generator_p_nom_mod > 0, expression: Generator_n_mod }
otherwise: Generator_modules_installed
\[
\mathrm{M}_{g,\xi} = \begin{cases} N_{g} & \text{if } \mathrm{ext}_{g} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0 \\ \mathrm{N}^{\mathrm{fix}}_{g,\xi} & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi
\]
Generator_operational_limit#
Generator_operational_limit:
expression: >-
sum(sum((Generator_p * GlobalConstraint_energy_weight) * Generator_operational_limit_weight, over=snapshot), over=generator)
\[
\mathit{Generator\_operational\_limit}_{o,\xi} = \sum_{g \in \mathcal{G}} \sum_{t \in \mathcal{T}} p_{g,\xi,t} \cdot \mathit{w}^{\mathrm{gc}}_{o,\xi,t} \cdot \mathrm{b}_{g,o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
Generator_opex#
Generator_opex:
expression: >-
sum(sum((Generator_running_cost * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=generator), over=snapshot)
+ sum(sum((((Generator_p * Generator_p) * Generator_marginal_cost_quadratic) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=generator), over=snapshot)
\[
\mathit{Generator\_opex}_{\xi} = \sum_{t \in \mathcal{T}} \sum_{g \in \mathcal{G}} C^{\mathrm{run}}_{g,\xi,t} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{g \in \mathcal{G}} p_{g,\xi,t} \cdot p_{g,\xi,t} \cdot \mathrm{c}^{(2)}_{g,\xi,t} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} \qquad \forall\, \xi \in \Xi
\]
Generator_p_carried_over#
Generator_p_carried_over:
description: >-
the output a generator carries over into a snapshot. At the first, it is
the `p_init` it brought in, which is zero where it came in off. After
that, it is the previous snapshot's
dims: [scenario, snapshot, generator]
cases:
opening: { when: "position(snapshot) == 0", expression: Generator_p_init }
otherwise: shift(Generator_p, along=snapshot, offset=1)
\[
\overleftarrow{p}^{\circ}_{g,\xi,t} = \begin{cases} \mathrm{p}^{0}_{g,\xi} & \text{if } \mathrm{pos}(t) = 0 \\ p_{g,\xi,t - 1} & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_p_nom_committed#
Generator_p_nom_committed:
description: >-
the build a committed unit's ramp rows are taken against: one module
where the build is modular, since its status counts modules, and the
given build otherwise (`constraints.py:1131-1134`)
dims: [scenario, generator]
cases:
modular_build: { when: "Generator_p_nom_mod > 0", expression: Generator_p_nom_mod }
otherwise: Generator_p_nom
\[
\widehat{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} = \begin{cases} \mathrm{p}^{\mathrm{mod}}_{g} & \text{if } \mathrm{p}^{\mathrm{mod}}_{g} > 0 \\ \mathrm{p}^{\mathrm{nom}}_{g,\xi} & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi
\]
Generator_p_nom_effective#
Generator_p_nom_effective:
description: "the build a generator's limits are taken against: the chosen one where it is extendable, the given one otherwise"
dims: [scenario, generator]
cases:
extendable: { when: Generator_p_nom_extendable, expression: Generator_p_nom_ext }
otherwise: Generator_p_nom
\[
\widetilde{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} = \begin{cases} P_{g} & \text{if } \mathrm{ext}_{g} \\ \mathrm{p}^{\mathrm{nom}}_{g,\xi} & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi
\]
Generator_previous_p#
Generator_previous_p:
description: >-
the output a generator carries into a snapshot. It is nothing at the
first snapshot it is active in past the first of the horizon: PyPSA
fills an output it did not build with zero (`constraints.py:716`).
It is the output carried over otherwise
dims: [scenario, snapshot, generator]
cases:
opening_late: { when: "position(snapshot) > 0 AND NOT shift(Generator_active, along=snapshot, offset=1)", expression: 0 }
otherwise: Generator_p_carried_over
\[
\overleftarrow{p}_{g,\xi,t} = \begin{cases} 0 & \text{if } \mathrm{pos}(t) > 0 \wedge \neg \mathrm{on}_{g,t - 1} \\ \overleftarrow{p}^{\circ}_{g,\xi,t} & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_previous_status#
Generator_previous_status:
description: >-
the commitment state a generator carries into a snapshot. It is off at
the first snapshot it is active in past the first of the horizon. PyPSA
reads a status it did not build as off there (`constraints.py:391`). It
is the state carried over otherwise
dims: [scenario, snapshot, generator]
cases:
opening_late: { when: "position(snapshot) > 0 AND NOT shift(Generator_active, along=snapshot, offset=1)", expression: 0 }
otherwise: Generator_status_carried_over
\[
\overleftarrow{u}_{g,\xi,t} = \begin{cases} 0 & \text{if } \mathrm{pos}(t) > 0 \wedge \neg \mathrm{on}_{g,t - 1} \\ \overleftarrow{u}^{\circ}_{g,\xi,t} & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_primary_energy#
Generator_primary_energy:
expression: >-
sum(sum((Generator_p * GlobalConstraint_energy_weight) * Generator_primary_energy_weight, over=snapshot), over=generator)
\[
\mathit{Generator\_primary\_energy}_{o,\xi} = \sum_{g \in \mathcal{G}} \sum_{t \in \mathcal{T}} p_{g,\xi,t} \cdot \mathit{w}^{\mathrm{gc}}_{o,\xi,t} \cdot \mathrm{a}_{g,o,\xi,t} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
Generator_ramp_down_allowance#
Generator_ramp_down_allowance:
description: >-
how far a generator may lower output between two snapshots. It is its
ramp limit of the build while it stays on, plus its shut-down ramp in
the snapshot it turns off
dims: [scenario, snapshot, generator]
cases:
committed:
when: Generator_committable
expression: >-
Generator_ramp_down_rate * Generator_p_nom_committed * Generator_status
+ Generator_shut_down_rate * Generator_p_nom_committed
* (Generator_ramp_previous_status - Generator_status)
otherwise: Generator_ramp_down_rate * Generator_p_nom_effective
\[
\Delta^{-}_{g,\xi,t} = \begin{cases} \widetilde{\mathrm{rd}}_{g,\xi,t} \cdot \widehat{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} \cdot u_{g,\xi,t} + \widetilde{\mathrm{rd}}^{\mathrm{dn}}_{g,\xi} \cdot \widehat{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} \cdot \left( \overleftarrow{\bar u}_{g,\xi,t} - u_{g,\xi,t} \right) & \text{if } \mathrm{com}_{g} \\ \widetilde{\mathrm{rd}}_{g,\xi,t} \cdot \widetilde{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_ramp_down_rate#
Generator_ramp_down_rate:
description: >-
the ramp limit in a unit's down row, PyPSA's `ramp_limit_down`. It is
the full build where PyPSA has none, since a shut-down ramp alone builds
the row
dims: [scenario, snapshot, generator]
cases:
given: { when: Generator_ramp_limit_down, expression: Generator_ramp_limit_down }
otherwise: 1
\[
\widetilde{\mathrm{rd}}_{g,\xi,t} = \begin{cases} \mathrm{rd}_{g,\xi,t} & \text{if } \mathrm{rd}_{g,\xi,t} \text{ is defined} \\ 1 & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_ramp_previous_status#
Generator_ramp_previous_status:
description: >-
the status a committed unit's ramp rows carry into a snapshot. At the
first snapshot, a fixed modular unit that came in running has every
module on, since the state it brought in is on or off
(`constraints.py:1152-1157`). The transition and down-time rows count it as one module
(PyPSA/PyPSA#2114). It is the state carried in otherwise
dims: [scenario, snapshot, generator]
cases:
opening_modules:
when: "position(snapshot) == 0 AND NOT Generator_p_nom_extendable AND Generator_p_nom_mod > 0"
expression: Generator_status_initial * Generator_modules_installed
otherwise: Generator_previous_status
\[
\overleftarrow{\bar u}_{g,\xi,t} = \begin{cases} \mathrm{u}^{0}_{g,\xi} \cdot \mathrm{N}^{\mathrm{fix}}_{g,\xi} & \text{if } \mathrm{pos}(t) = 0 \wedge \neg \mathrm{ext}_{g} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0 \\ \overleftarrow{u}_{g,\xi,t} & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_ramp_up_allowance#
Generator_ramp_up_allowance:
description: >-
how far a generator may raise output between two snapshots. It is its
ramp limit of the build while it stays on, plus its start-up ramp in the
snapshot it turns on
dims: [scenario, snapshot, generator]
cases:
committed:
when: Generator_committable
expression: >-
Generator_ramp_up_rate * Generator_p_nom_committed * Generator_ramp_previous_status
+ Generator_start_up_rate * Generator_p_nom_committed
* (Generator_status - Generator_ramp_previous_status)
otherwise: Generator_ramp_up_rate * Generator_p_nom_effective
\[
\Delta^{+}_{g,\xi,t} = \begin{cases} \widetilde{\mathrm{ru}}_{g,\xi,t} \cdot \widehat{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} \cdot \overleftarrow{\bar u}_{g,\xi,t} + \widetilde{\mathrm{ru}}^{\mathrm{up}}_{g,\xi} \cdot \widehat{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} \cdot \left( u_{g,\xi,t} - \overleftarrow{\bar u}_{g,\xi,t} \right) & \text{if } \mathrm{com}_{g} \\ \widetilde{\mathrm{ru}}_{g,\xi,t} \cdot \widetilde{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_ramp_up_rate#
Generator_ramp_up_rate:
description: >-
the ramp limit in a unit's up row, PyPSA's `ramp_limit_up`. It is the
full build where PyPSA has none, since a start-up ramp alone builds the
row
dims: [scenario, snapshot, generator]
cases:
given: { when: Generator_ramp_limit_up, expression: Generator_ramp_limit_up }
otherwise: 1
\[
\widetilde{\mathrm{ru}}_{g,\xi,t} = \begin{cases} \mathrm{ru}_{g,\xi,t} & \text{if } \mathrm{ru}_{g,\xi,t} \text{ is defined} \\ 1 & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_running_cost#
Generator_running_cost:
description: >-
what a generator's output costs in a snapshot: the value read off its
piecewise curve where it has one, and its linear `marginal_cost`
otherwise. A curve replaces the linear cost, as PyPSA drops the unit
from that term (`optimize.py:333-348`)
dims: [scenario, snapshot, generator]
cases:
curve: { when: Generator_marginal_cost_curve, expression: Generator_marginal_cost_piecewise }
otherwise: Generator_p * Generator_marginal_cost
\[
C^{\mathrm{run}}_{g,\xi,t} = \begin{cases} C^{\mathrm{curve}}_{g,\xi,t} & \text{if } \mathrm{curve}_{g} \\ p_{g,\xi,t} \cdot \mathrm{c}_{g,\xi,t} & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_shut_down_rate#
Generator_shut_down_rate:
description: >-
the shut-down ramp in a unit's down row (PyPSA's
`ramp_limit_shut_down`), or the full build where it has none
dims: [scenario, generator]
cases:
given: { when: Generator_ramp_limit_shut_down, expression: Generator_ramp_limit_shut_down }
otherwise: 1
\[
\widetilde{\mathrm{rd}}^{\mathrm{dn}}_{g,\xi} = \begin{cases} \mathrm{rd}^{\mathrm{dn}}_{g,\xi} & \text{if } \mathrm{rd}^{\mathrm{dn}}_{g,\xi} \text{ is defined} \\ 1 & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi
\]
Generator_start_up_rate#
Generator_start_up_rate:
description: >-
the start-up ramp in a unit's up row (PyPSA's `ramp_limit_start_up`), or
the full build where it has none
dims: [scenario, generator]
cases:
given: { when: Generator_ramp_limit_start_up, expression: Generator_ramp_limit_start_up }
otherwise: 1
\[
\widetilde{\mathrm{ru}}^{\mathrm{up}}_{g,\xi} = \begin{cases} \mathrm{ru}^{\mathrm{up}}_{g,\xi} & \text{if } \mathrm{ru}^{\mathrm{up}}_{g,\xi} \text{ is defined} \\ 1 & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi
\]
Generator_status_carried_over#
Generator_status_carried_over:
description: >-
the state a generator carries over into a snapshot: the state it brought
into the horizon at the first, the previous snapshot's after that
dims: [scenario, snapshot, generator]
cases:
opening: { when: "position(snapshot) == 0", expression: Generator_status_initial }
otherwise: shift(Generator_status, along=snapshot, offset=1)
\[
\overleftarrow{u}^{\circ}_{g,\xi,t} = \begin{cases} \mathrm{u}^{0}_{g,\xi} & \text{if } \mathrm{pos}(t) = 0 \\ u_{g,\xi,t - 1} & \text{otherwise} \end{cases} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_tech_capacity_expansion#
Generator_tech_capacity_expansion:
expression: sum(Generator_p_nom_ext * Generator_tech_capacity_weight, over=generator)
\[
\mathit{Generator\_tech\_capacity\_expansion}_{o} = \sum_{g \in \mathcal{G}} P_{g} \cdot \mathrm{m}_{g,o} \qquad \forall\, o \in \mathcal{O}
\]
GlobalConstraint_energy_weight#
GlobalConstraint_energy_weight:
description: >-
what one unit of power at a snapshot counts for in a row. Where the row
counts the snapshot, it is the generator weighting times the years of
the snapshot's period. Where it does not, it is nothing
dims: [scenario, global_constraint, snapshot]
cases:
counted:
when: GlobalConstraint_counts_snapshot
expression: snapshot_weightings_generators * at(period_weight_years, by=snapshot_period[period])
otherwise: 0
\[
\mathit{w}^{\mathrm{gc}}_{o,\xi,t} = \begin{cases} \mathrm{w}^{\mathrm{gen}}_{t} \cdot \mathrm{w}^{\mathrm{yr}}_{\mathrm{snapshot\_period}(t)} & \text{if } \mathrm{in}_{o,\xi,t} \\ 0 & \text{otherwise} \end{cases} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Line_additions#
Line_additions:
expression: >-
sum(Line_s_nom_ext * Line_first_active, over=line, by=Line_carrier[carrier])
\[
\mathit{Line\_additions}_{i,y} = \sum_{k \in \mathcal{K} \,:\, \mathrm{Line\_carrier}(k) = i} S_{k} \cdot \mathrm{new}^{s}_{k,y} \qquad \forall\, i \in \mathcal{I},\ y \in \mathcal{Y}
\]
Line_angle_sum#
\[
\mathit{Line\_angle\_sum}_{c,\xi,t} = \sum_{k \in \mathcal{K}} s_{k,\xi,t} \cdot \mathrm{x}_{c,k,\mathrm{snapshot\_period}(t)} \qquad \forall\, c \in \mathcal{C},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Line_capex#
Line_capex:
expression: sum(scenario_weight * Line_s_nom_ext * Line_capital_cost * Line_capital_weight)
\[
\mathit{Line\_capex} = \sum_{k \in \mathcal{K},\ \xi \in \Xi} \pi_{\xi} \cdot S_{k} \cdot \mathrm{c}^{\mathrm{cap},s}_{k,\xi} \cdot \mathrm{W}^{s}_{k}
\]
Line_injection#
Line_injection:
expression: >-
-sum(Line_s, over=line, by=Line_bus0[bus])
+ sum(Line_s, over=line, by=Line_bus1[bus])
- (0.5 * sum(Line_loss, over=line, by=Line_bus0[bus]))
- (0.5 * sum(Line_loss, over=line, by=Line_bus1[bus]))
\[
\mathit{Line\_injection}_{n,\xi,t} = -\left( \sum_{k \in \mathcal{K} \,:\, \mathrm{Line\_bus0}(k) = n} s_{k,\xi,t} \right) + \sum_{k \in \mathcal{K} \,:\, \mathrm{Line\_bus1}(k) = n} s_{k,\xi,t} - 0.5 \cdot \left( \sum_{k \in \mathcal{K} \,:\, \mathrm{Line\_bus0}(k) = n} \ell_{k,\xi,t} \right) - 0.5 \cdot \left( \sum_{k \in \mathcal{K} \,:\, \mathrm{Line\_bus1}(k) = n} \ell_{k,\xi,t} \right) \qquad \forall\, n \in \mathcal{N},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Line_s_monitored#
Line_s_monitored:
description: >-
the flow in a line's post-contingency rows. It is its flow where it
is active, and nothing where it does not. PyPSA builds those rows for every
branch of the sub-network in every snapshot
dims: [scenario, snapshot, line]
cases:
standing: { when: Line_active, expression: Line_s }
otherwise: 0
\[
\check{s}_{k,\xi,t} = \begin{cases} s_{k,\xi,t} & \text{if } \mathrm{on}^{s}_{k,t} \\ 0 & \text{otherwise} \end{cases} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Line_tech_capacity_expansion#
Line_tech_capacity_expansion:
expression: sum(Line_s_nom_ext * Line_tech_capacity_weight, over=line)
\[
\mathit{Line\_tech\_capacity\_expansion}_{o} = \sum_{k \in \mathcal{K}} S_{k} \cdot \mathrm{m}^{l}_{o,k} \qquad \forall\, o \in \mathcal{O}
\]
Line_transmission_expansion_cost#
Line_transmission_expansion_cost:
expression: sum(Line_s_nom_ext * Line_expansion_cost_weight, over=line)
\[
\mathit{Line\_transmission\_expansion\_cost}_{o,\xi} = \sum_{k \in \mathcal{K}} S_{k} \cdot \mathrm{cc}_{o,k,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
Line_transmission_volume_expansion#
\[
\mathit{Line\_transmission\_volume\_expansion}_{o,\xi} = \sum_{k \in \mathcal{K}} S_{k} \cdot \mathrm{len}_{o,k,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
Link_additions#
Link_additions:
expression: >-
sum(Link_p_nom_ext * Link_first_active, over=link, by=Link_carrier[carrier])
\[
\mathit{Link\_additions}_{i,y} = \sum_{l \in \mathcal{L} \,:\, \mathrm{Link\_carrier}(l) = i} F_{l} \cdot \mathrm{new}^{f}_{l,y} \qquad \forall\, i \in \mathcal{I},\ y \in \mathcal{Y}
\]
Link_capex#
Link_capex:
expression: sum(scenario_weight * Link_p_nom_ext * Link_capital_cost * Link_capital_weight)
\[
\mathit{Link\_capex} = \sum_{l \in \mathcal{L},\ \xi \in \Xi} \pi_{\xi} \cdot F_{l} \cdot \mathrm{c}^{\mathrm{cap},f}_{l,\xi} \cdot \mathrm{W}^{f}_{l}
\]
Link_commitment_opex#
Link_commitment_opex:
expression: >-
sum(sum(((Link_status * Link_stand_by_cost) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=link), over=snapshot)
+ sum(sum((Link_start_up * Link_start_up_cost) * at(period_weight_objective, by=snapshot_period[period]), over=link), over=snapshot)
+ sum(sum((Link_shut_down * Link_shut_down_cost) * at(period_weight_objective, by=snapshot_period[period]), over=link), over=snapshot)
\[
\mathit{Link\_commitment\_opex}_{\xi} = \sum_{t \in \mathcal{T}} \sum_{l \in \mathcal{L}} u^{f}_{l,\xi,t} \cdot \mathrm{c}^{f,\mathrm{on}}_{l,\xi,t} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{l \in \mathcal{L}} \mathit{up}^{f}_{l,\xi,t} \cdot \mathrm{c}^{f,\mathrm{up}}_{l,\xi,t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{l \in \mathcal{L}} \mathit{dn}^{f}_{l,\xi,t} \cdot \mathrm{c}^{f,\mathrm{dn}}_{l,\xi,t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} \qquad \forall\, \xi \in \Xi
\]
Link_injection#
Link_injection:
expression: >-
-sum(Link_p, over=link, by=Link_bus0[bus])
+ sum(Link_output_arrival, over=link_output, by=Link_output_bus[bus])
\[
\mathit{Link\_injection}_{n,\xi,t} = -\left( \sum_{l \in \mathcal{L} \,:\, \mathrm{Link\_bus0}(l) = n} f_{l,\xi,t} \right) + \sum_{o \in \mathcal{O} \,:\, \mathrm{Link\_output\_bus}(o) = n} \overrightarrow{f}_{o,\xi,t} \qquad \forall\, n \in \mathcal{N},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Link_maintenance_modules#
Link_maintenance_modules:
description: >-
how many modules of a modular committable link can be on in
maintenance: its build cap in modules where the build is extendable,
and the modules installed where it is fixed (`constraints.py:588-594`)
dims: [scenario, link]
cases:
extendable: { when: Link_p_nom_extendable, expression: Link_p_nom_max / Link_p_nom_mod }
otherwise: Link_modules_installed
\[
\mathrm{Link\_maintenance\_modules}_{l,\xi} = \begin{cases} \frac{\overline{\mathrm{f}}^{\mathrm{nom}}_{l,\xi}}{\mathrm{f}^{\mathrm{mod}}_{l}} & \text{if } \mathrm{ext}^{f}_{l} \\ \mathrm{N}^{f,\mathrm{fix}}_{l,\xi} & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi
\]
Link_modules#
Link_modules:
description: >-
how many modules a committed link's status counts: the chosen number
where the build is extendable and modular, and the modules installed
otherwise (`constraints.py:656-661`)
dims: [scenario, link]
cases:
chosen: { when: Link_p_nom_extendable AND Link_p_nom_mod > 0, expression: Link_n_mod }
otherwise: Link_modules_installed
\[
\mathrm{M}^{f}_{l,\xi} = \begin{cases} N^{f}_{l} & \text{if } \mathrm{ext}^{f}_{l} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0 \\ \mathrm{N}^{f,\mathrm{fix}}_{l,\xi} & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi
\]
Link_opex#
Link_opex:
expression: >-
sum(sum(((Link_p * Link_marginal_cost) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=link), over=snapshot)
+ sum(sum((((Link_p * Link_p) * Link_marginal_cost_quadratic) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=link), over=snapshot)
\[
\mathit{Link\_opex}_{\xi} = \sum_{t \in \mathcal{T}} \sum_{l \in \mathcal{L}} f_{l,\xi,t} \cdot \mathrm{c}^{f}_{l,\xi,t} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{l \in \mathcal{L}} f_{l,\xi,t} \cdot f_{l,\xi,t} \cdot \mathrm{c}^{f,(2)}_{l,\xi,t} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} \qquad \forall\, \xi \in \Xi
\]
Link_output_arrival#
Link_output_arrival:
description: >-
what a link delivers to an output port at a snapshot. It is its flow
delayed by the port's `delay` within its investment period. It is
multiplied by the port's efficiency at the snapshot the flow arrives.
Where the port is `cyclic_delay`, the delayed flow wraps from the
period's end. Where it is not, the flow still in transit at the period's
first snapshots is lost. A port that does not delay (`delay` zero)
delivers its flow unshifted, cyclic or not
dims: [scenario, snapshot, link_output]
cases:
wrapping:
when: Link_output_cyclic_delay
expression: shift(at(Link_p, by=Link_output_link[link]), along=snapshot, offset=Link_output_delay, edge='wrap', within=snapshot_period[period]) * Link_efficiency
otherwise: shift(at(Link_p, by=Link_output_link[link]), along=snapshot, offset=Link_output_delay, edge=0, within=snapshot_period[period]) * Link_efficiency
\[
\overrightarrow{f}_{o,\xi,t} = \begin{cases} f_{\mathrm{Link\_output\_link}(o),\xi,t \ominus^{\mathrm{snapshot\_period}(t)} \mathrm{d}^{f}} \cdot \eta_{o,\xi,t} & \text{if } \mathrm{cyc}^{f}_{o} \\ f_{\mathrm{Link\_output\_link}(o),\xi,t \boxminus_{0}^{\mathrm{snapshot\_period}(t)} \mathrm{d}^{f}} \cdot \eta_{o,\xi,t} & \text{otherwise} \end{cases} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Link_p_carried_over#
Link_p_carried_over:
description: >-
the output a link carries over into a snapshot. At the first, it is the
`p_init` it brought in, which is zero where it came in off. After that,
it is the previous snapshot's
dims: [scenario, snapshot, link]
cases:
opening: { when: "position(snapshot) == 0", expression: Link_p_init }
otherwise: shift(Link_p, along=snapshot, offset=1)
\[
\overleftarrow{f}^{\circ}_{l,\xi,t} = \begin{cases} \mathrm{f}^{0}_{l,\xi} & \text{if } \mathrm{pos}(t) = 0 \\ f_{l,\xi,t - 1} & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Link_p_nom_committed#
Link_p_nom_committed:
description: >-
the build a committed link's ramp rows are taken against: one module
where the build is modular, since its status counts modules, and the
given build otherwise (`constraints.py:1131-1134`)
dims: [scenario, link]
cases:
modular_build: { when: "Link_p_nom_mod > 0", expression: Link_p_nom_mod }
otherwise: Link_p_nom
\[
\widehat{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} = \begin{cases} \mathrm{f}^{\mathrm{mod}}_{l} & \text{if } \mathrm{f}^{\mathrm{mod}}_{l} > 0 \\ \mathrm{f}^{\mathrm{nom}}_{l,\xi} & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi
\]
Link_p_nom_effective#
Link_p_nom_effective:
description: "the build a link's limits are taken against: the chosen one where it is extendable, the given one otherwise"
dims: [scenario, link]
cases:
extendable: { when: Link_p_nom_extendable, expression: Link_p_nom_ext }
otherwise: Link_p_nom
\[
\widetilde{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} = \begin{cases} F_{l} & \text{if } \mathrm{ext}^{f}_{l} \\ \mathrm{f}^{\mathrm{nom}}_{l,\xi} & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi
\]
Link_previous_p#
Link_previous_p:
description: >-
the output a link carries into a snapshot. It is nothing at the first
snapshot it is active in past the first of the horizon: PyPSA fills an
output it did not build with zero (`constraints.py:716`). It is the
output carried over otherwise
dims: [scenario, snapshot, link]
cases:
opening_late: { when: "position(snapshot) > 0 AND NOT shift(Link_active, along=snapshot, offset=1)", expression: 0 }
otherwise: Link_p_carried_over
\[
\overleftarrow{f}_{l,\xi,t} = \begin{cases} 0 & \text{if } \mathrm{pos}(t) > 0 \wedge \neg \mathrm{on}^{f}_{l,t - 1} \\ \overleftarrow{f}^{\circ}_{l,\xi,t} & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Link_previous_status#
Link_previous_status:
description: >-
the commitment state a link carries into a snapshot. It is off at the
first snapshot it is active in past the first of the horizon. PyPSA reads a
status it did not build as off there (`constraints.py:391`). It is the
state carried over otherwise
dims: [scenario, snapshot, link]
cases:
opening_late: { when: "position(snapshot) > 0 AND NOT shift(Link_active, along=snapshot, offset=1)", expression: 0 }
otherwise: Link_status_carried_over
\[
\overleftarrow{u}^{f}_{l,\xi,t} = \begin{cases} 0 & \text{if } \mathrm{pos}(t) > 0 \wedge \neg \mathrm{on}^{f}_{l,t - 1} \\ \overleftarrow{u}^{\circ f}_{l,\xi,t} & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Link_ramp_down_allowance#
Link_ramp_down_allowance:
description: >-
how far a link may lower flow between two snapshots. It is its ramp
limit of the build while it stays on, plus its shut-down ramp in the
snapshot it turns off
dims: [scenario, snapshot, link]
cases:
committed:
when: Link_committable
expression: >-
Link_ramp_down_rate * Link_p_nom_committed * Link_status
+ Link_shut_down_rate * Link_p_nom_committed
* (Link_ramp_previous_status - Link_status)
otherwise: Link_ramp_down_rate * Link_p_nom_effective
\[
\Delta^{f,-}_{l,\xi,t} = \begin{cases} \widetilde{\mathrm{rd}}^{f}_{l,\xi,t} \cdot \widehat{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} \cdot u^{f}_{l,\xi,t} + \widetilde{\mathrm{rd}}^{f,\mathrm{dn}}_{l,\xi} \cdot \widehat{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} \cdot \left( \overleftarrow{\bar u}^{f}_{l,\xi,t} - u^{f}_{l,\xi,t} \right) & \text{if } \mathrm{com}^{f}_{l} \\ \widetilde{\mathrm{rd}}^{f}_{l,\xi,t} \cdot \widetilde{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Link_ramp_down_rate#
Link_ramp_down_rate:
description: >-
the ramp limit in a link's down row, PyPSA's `ramp_limit_down`. It is
the full build where PyPSA has none, since a shut-down ramp alone builds
the row
dims: [scenario, snapshot, link]
cases:
given: { when: Link_ramp_limit_down, expression: Link_ramp_limit_down }
otherwise: 1
\[
\widetilde{\mathrm{rd}}^{f}_{l,\xi,t} = \begin{cases} \mathrm{rd}^{f}_{l,\xi,t} & \text{if } \mathrm{rd}^{f}_{l,\xi,t} \text{ is defined} \\ 1 & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Link_ramp_previous_status#
Link_ramp_previous_status:
description: >-
the status a committed link's ramp rows carry into a snapshot. At the
first snapshot, a fixed modular link that came in running has every
module on, since the state it brought in is on or off
(`constraints.py:1152-1157`). The transition and down-time rows count it as one module
(PyPSA/PyPSA#2114). It is the state carried in otherwise
dims: [scenario, snapshot, link]
cases:
opening_modules:
when: "position(snapshot) == 0 AND NOT Link_p_nom_extendable AND Link_p_nom_mod > 0"
expression: Link_status_initial * Link_modules_installed
otherwise: Link_previous_status
\[
\overleftarrow{\bar u}^{f}_{l,\xi,t} = \begin{cases} \mathrm{u}^{f,0}_{l,\xi} \cdot \mathrm{N}^{f,\mathrm{fix}}_{l,\xi} & \text{if } \mathrm{pos}(t) = 0 \wedge \neg \mathrm{ext}^{f}_{l} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0 \\ \overleftarrow{u}^{f}_{l,\xi,t} & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Link_ramp_up_allowance#
Link_ramp_up_allowance:
description: >-
how far a link may raise flow between two snapshots. It is its ramp
limit of the build while it stays on, plus its start-up ramp in the
snapshot it turns on
dims: [scenario, snapshot, link]
cases:
committed:
when: Link_committable
expression: >-
Link_ramp_up_rate * Link_p_nom_committed * Link_ramp_previous_status
+ Link_start_up_rate * Link_p_nom_committed
* (Link_status - Link_ramp_previous_status)
otherwise: Link_ramp_up_rate * Link_p_nom_effective
\[
\Delta^{f,+}_{l,\xi,t} = \begin{cases} \widetilde{\mathrm{ru}}^{f}_{l,\xi,t} \cdot \widehat{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} \cdot \overleftarrow{\bar u}^{f}_{l,\xi,t} + \widetilde{\mathrm{ru}}^{f,\mathrm{up}}_{l,\xi} \cdot \widehat{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} \cdot \left( u^{f}_{l,\xi,t} - \overleftarrow{\bar u}^{f}_{l,\xi,t} \right) & \text{if } \mathrm{com}^{f}_{l} \\ \widetilde{\mathrm{ru}}^{f}_{l,\xi,t} \cdot \widetilde{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Link_ramp_up_rate#
Link_ramp_up_rate:
description: >-
the ramp limit in a link's up row, PyPSA's `ramp_limit_up`. It is the
full build where PyPSA has none, since a start-up ramp alone builds the
row
dims: [scenario, snapshot, link]
cases:
given: { when: Link_ramp_limit_up, expression: Link_ramp_limit_up }
otherwise: 1
\[
\widetilde{\mathrm{ru}}^{f}_{l,\xi,t} = \begin{cases} \mathrm{ru}^{f}_{l,\xi,t} & \text{if } \mathrm{ru}^{f}_{l,\xi,t} \text{ is defined} \\ 1 & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Link_shut_down_rate#
Link_shut_down_rate:
description: >-
the shut-down ramp in a link's down row (PyPSA's
`ramp_limit_shut_down`), or the full build where it has none
dims: [scenario, link]
cases:
given: { when: Link_ramp_limit_shut_down, expression: Link_ramp_limit_shut_down }
otherwise: 1
\[
\widetilde{\mathrm{rd}}^{f,\mathrm{dn}}_{l,\xi} = \begin{cases} \mathrm{rd}^{f,\mathrm{dn}}_{l,\xi} & \text{if } \mathrm{rd}^{f,\mathrm{dn}}_{l,\xi} \text{ is defined} \\ 1 & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi
\]
Link_start_up_rate#
Link_start_up_rate:
description: >-
the start-up ramp in a link's up row (PyPSA's `ramp_limit_start_up`), or
the full build where it has none
dims: [scenario, link]
cases:
given: { when: Link_ramp_limit_start_up, expression: Link_ramp_limit_start_up }
otherwise: 1
\[
\widetilde{\mathrm{ru}}^{f,\mathrm{up}}_{l,\xi} = \begin{cases} \mathrm{ru}^{f,\mathrm{up}}_{l,\xi} & \text{if } \mathrm{ru}^{f,\mathrm{up}}_{l,\xi} \text{ is defined} \\ 1 & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi
\]
Link_status_carried_over#
Link_status_carried_over:
description: >-
the state a link carries over into a snapshot: the state it brought into
the horizon at the first, the previous snapshot's after that
dims: [scenario, snapshot, link]
cases:
opening: { when: "position(snapshot) == 0", expression: Link_status_initial }
otherwise: shift(Link_status, along=snapshot, offset=1)
\[
\overleftarrow{u}^{\circ f}_{l,\xi,t} = \begin{cases} \mathrm{u}^{f,0}_{l,\xi} & \text{if } \mathrm{pos}(t) = 0 \\ u^{f}_{l,\xi,t - 1} & \text{otherwise} \end{cases} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Link_tech_capacity_expansion#
Link_tech_capacity_expansion:
expression: sum(Link_p_nom_ext * Link_tech_capacity_weight, over=link)
\[
\mathit{Link\_tech\_capacity\_expansion}_{o} = \sum_{l \in \mathcal{L}} F_{l} \cdot \mathrm{m}^{f}_{o,l} \qquad \forall\, o \in \mathcal{O}
\]
Link_transmission_expansion_cost#
Link_transmission_expansion_cost:
expression: sum(Link_p_nom_ext * Link_expansion_cost_weight, over=link)
\[
\mathit{Link\_transmission\_expansion\_cost}_{o,\xi} = \sum_{l \in \mathcal{L}} F_{l} \cdot \mathrm{cc}^{f}_{o,l,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
Link_transmission_volume_expansion#
\[
\mathit{Link\_transmission\_volume\_expansion}_{o,\xi} = \sum_{l \in \mathcal{L}} F_{l} \cdot \mathrm{len}^{f}_{o,l,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
Load_demand#
Load_demand:
description: >-
what a load draws from its bus's balance. It is its demand times its
sign where it is active, and nothing where it is not. PyPSA drops an
inactive load from the balance (`constraints.py:1591-1592`)
dims: [scenario, snapshot, load]
cases:
active: { when: Load_active, expression: Load_sign * Load_p_set }
otherwise: 0
\[
\check{\mathrm{load}}_{d,\xi,t} = \begin{cases} \mathrm{sgn}^{\mathrm{load}}_{d} \cdot \mathrm{load}_{d,\xi,t} & \text{if } \mathrm{on}^{\mathrm{load}}_{d} \\ 0 & \text{otherwise} \end{cases} \qquad \forall\, d \in \mathcal{D},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Load_injection#
\[
\mathrm{Load\_injection}_{n,\xi,t} = \sum_{d \in \mathcal{D} \,:\, \mathrm{Load\_bus}(d) = n} \check{\mathrm{load}}_{d,\xi,t} \qquad \forall\, n \in \mathcal{N},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Outage_s#
Outage_s:
description: >-
the flow an outage takes off its branch: the outaged line's or
transformer's flow before it goes out
dims: [scenario, snapshot, outage]
cases:
line: { when: Outage_line, expression: "at(Line_s_monitored, by=Outage_line[line])" }
otherwise: at(Transformer_s_monitored, by=Outage_transformer[transformer])
\[
\hat{s}_{\kappa,\xi,t} = \begin{cases} \check{s}_{\mathrm{Outage\_line}(\kappa),\xi,t} & \text{if } \mathrm{Outage\_line}(\kappa) \text{ is defined} \\ \check{\sigma}_{\xi,t,\mathrm{Outage\_transformer}(\kappa)} & \text{otherwise} \end{cases} \qquad \forall\, \kappa \in \mathcal{K}^{\mathrm{out}},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Process_additions#
Process_additions:
expression: >-
sum(Process_p_nom_ext * Process_first_active, over=process, by=Process_carrier[carrier])
\[
\mathit{Process\_additions}_{i,y} = \sum_{j \in \mathcal{J} \,:\, \mathrm{Process\_carrier}(j) = i} Z_{j} \cdot \mathrm{new}^{z}_{y,j} \qquad \forall\, i \in \mathcal{I},\ y \in \mathcal{Y}
\]
Process_capex#
Process_capex:
expression: sum(scenario_weight * Process_p_nom_ext * Process_capital_cost * Process_capital_weight)
\[
\mathit{Process\_capex} = \sum_{j \in \mathcal{J},\ \xi \in \Xi} \pi_{\xi} \cdot Z_{j} \cdot \mathrm{c}^{\mathrm{cap},z}_{j,\xi} \cdot \mathrm{W}^{z}_{j}
\]
Process_commitment_opex#
Process_commitment_opex:
expression: >-
sum(sum(((Process_status * Process_stand_by_cost) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=process), over=snapshot)
+ sum(sum((Process_start_up * Process_start_up_cost) * at(period_weight_objective, by=snapshot_period[period]), over=process), over=snapshot)
+ sum(sum((Process_shut_down * Process_shut_down_cost) * at(period_weight_objective, by=snapshot_period[period]), over=process), over=snapshot)
\[
\mathit{Process\_commitment\_opex}_{\xi} = \sum_{t \in \mathcal{T}} \sum_{j \in \mathcal{J}} u^{z}_{j,\xi,t} \cdot \mathrm{c}^{z,\mathrm{on}}_{j,\xi,t} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{j \in \mathcal{J}} \mathit{up}^{z}_{j,\xi,t} \cdot \mathrm{c}^{z,\mathrm{up}}_{j,\xi,t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{j \in \mathcal{J}} \mathit{dn}^{z}_{j,\xi,t} \cdot \mathrm{c}^{z,\mathrm{dn}}_{j,\xi,t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} \qquad \forall\, \xi \in \Xi
\]
Process_injection#
Process_injection:
expression: >-
sum(Process_output_arrival, over=process_output, by=Process_output_bus[bus])
\[
\mathit{Process\_injection}_{n,\xi,t} = \sum_{r \in \mathcal{R} \,:\, \mathrm{Process\_output\_bus}(r) = n} \overrightarrow{z}_{r,\xi,t} \qquad \forall\, n \in \mathcal{N},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Process_maintenance_modules#
Process_maintenance_modules:
description: >-
how many modules of a modular committable process can be on in
maintenance: its build cap in modules where the build is extendable,
and the modules installed where it is fixed (`constraints.py:588-594`)
dims: [scenario, process]
cases:
extendable: { when: Process_p_nom_extendable, expression: Process_p_nom_max / Process_p_nom_mod }
otherwise: Process_modules_installed
\[
\mathrm{Process\_maintenance\_modules}_{j,\xi} = \begin{cases} \frac{\overline{\mathrm{z}}^{\mathrm{nom}}_{j,\xi}}{\mathrm{z}^{\mathrm{mod}}_{j}} & \text{if } \mathrm{ext}^{z}_{j} \\ \mathrm{N}^{z,\mathrm{fix}}_{j,\xi} & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi
\]
Process_modules#
Process_modules:
description: >-
how many modules a committed process's status counts: the chosen number
where the build is extendable and modular, and the modules installed
otherwise (`constraints.py:656-661`)
dims: [scenario, process]
cases:
chosen: { when: Process_p_nom_extendable AND Process_p_nom_mod > 0, expression: Process_n_mod }
otherwise: Process_modules_installed
\[
\mathrm{M}^{z}_{j,\xi} = \begin{cases} N^{z}_{j} & \text{if } \mathrm{ext}^{z}_{j} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0 \\ \mathrm{N}^{z,\mathrm{fix}}_{j,\xi} & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi
\]
Process_opex#
Process_opex:
expression: >-
sum(sum(((Process_p * Process_marginal_cost) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=process), over=snapshot)
+ sum(sum((((Process_p * Process_p) * Process_marginal_cost_quadratic) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=process), over=snapshot)
\[
\mathit{Process\_opex}_{\xi} = \sum_{t \in \mathcal{T}} \sum_{j \in \mathcal{J}} z_{j,\xi,t} \cdot \mathrm{c}^{z}_{j,\xi,t} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{j \in \mathcal{J}} z_{j,\xi,t} \cdot z_{j,\xi,t} \cdot \mathrm{c}^{z,(2)}_{j,\xi,t} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} \qquad \forall\, \xi \in \Xi
\]
Process_output_arrival#
Process_output_arrival:
description: >-
what a process transfers at a port at a snapshot. It is its internal
power delayed by the port's `delay` within its investment period. It is
multiplied by the port's rate at the snapshot the transfer arrives.
Where the port is `cyclic_delay`, the delayed transfer wraps from the
period's end. Where it is not, the energy still in transit at the
period's first snapshots is lost. A port that does not delay (`delay`
zero) transfers at once, cyclic or not
dims: [scenario, snapshot, process_output]
cases:
wrapping:
when: Process_output_cyclic_delay
expression: shift(at(Process_p, by=Process_output_process[process]), along=snapshot, offset=Process_output_delay, edge='wrap', within=snapshot_period[period]) * Process_rate
otherwise: shift(at(Process_p, by=Process_output_process[process]), along=snapshot, offset=Process_output_delay, edge=0, within=snapshot_period[period]) * Process_rate
\[
\overrightarrow{z}_{r,\xi,t} = \begin{cases} z_{\mathrm{Process\_output\_process}(r),\xi,t \ominus^{\mathrm{snapshot\_period}(t)} \mathrm{d}^{z}} \cdot \alpha_{r,\xi,t} & \text{if } \mathrm{cyc}^{z}_{r} \\ z_{\mathrm{Process\_output\_process}(r),\xi,t \boxminus_{0}^{\mathrm{snapshot\_period}(t)} \mathrm{d}^{z}} \cdot \alpha_{r,\xi,t} & \text{otherwise} \end{cases} \qquad \forall\, r \in \mathcal{R},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Process_p_carried_over#
Process_p_carried_over:
description: >-
the output a process carries over into a snapshot. At the first, it is
the `p_init` it brought in, which is zero where it came in off. After
that, it is the previous snapshot's
dims: [scenario, snapshot, process]
cases:
opening: { when: "position(snapshot) == 0", expression: Process_p_init }
otherwise: shift(Process_p, along=snapshot, offset=1)
\[
\overleftarrow{z}^{\circ}_{j,\xi,t} = \begin{cases} \mathrm{z}^{0}_{j,\xi} & \text{if } \mathrm{pos}(t) = 0 \\ z_{j,\xi,t - 1} & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Process_p_nom_committed#
Process_p_nom_committed:
description: >-
the build a committed process's ramp rows are taken against: one module
where the build is modular, since its status counts modules, and the
given build otherwise (`constraints.py:1131-1134`)
dims: [scenario, process]
cases:
modular_build: { when: "Process_p_nom_mod > 0", expression: Process_p_nom_mod }
otherwise: Process_p_nom
\[
\widehat{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} = \begin{cases} \mathrm{z}^{\mathrm{mod}}_{j} & \text{if } \mathrm{z}^{\mathrm{mod}}_{j} > 0 \\ \mathrm{z}^{\mathrm{nom}}_{j,\xi} & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi
\]
Process_p_nom_effective#
Process_p_nom_effective:
description: "the build a process's limits are taken against: the chosen one where it is extendable, the given one otherwise"
dims: [scenario, process]
cases:
extendable: { when: Process_p_nom_extendable, expression: Process_p_nom_ext }
otherwise: Process_p_nom
\[
\widetilde{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} = \begin{cases} Z_{j} & \text{if } \mathrm{ext}^{z}_{j} \\ \mathrm{z}^{\mathrm{nom}}_{j,\xi} & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi
\]
Process_previous_p#
Process_previous_p:
description: >-
the output a process carries into a snapshot. It is nothing at the first
snapshot it is active in past the first of the horizon: PyPSA fills an
output it did not build with zero (`constraints.py:716`). It is the
output carried over otherwise
dims: [scenario, snapshot, process]
cases:
opening_late: { when: "position(snapshot) > 0 AND NOT shift(Process_active, along=snapshot, offset=1)", expression: 0 }
otherwise: Process_p_carried_over
\[
\overleftarrow{z}_{j,\xi,t} = \begin{cases} 0 & \text{if } \mathrm{pos}(t) > 0 \wedge \neg \mathrm{on}^{z}_{j,t - 1} \\ \overleftarrow{z}^{\circ}_{j,\xi,t} & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Process_previous_status#
Process_previous_status:
description: >-
the commitment state a process carries into a snapshot. It is off at the
first snapshot it is active in past the first of the horizon. PyPSA reads a
status it did not build as off there (`constraints.py:391`). It is the
state carried over otherwise
dims: [scenario, snapshot, process]
cases:
opening_late: { when: "position(snapshot) > 0 AND NOT shift(Process_active, along=snapshot, offset=1)", expression: 0 }
otherwise: Process_status_carried_over
\[
\overleftarrow{u}^{z}_{j,\xi,t} = \begin{cases} 0 & \text{if } \mathrm{pos}(t) > 0 \wedge \neg \mathrm{on}^{z}_{j,t - 1} \\ \overleftarrow{u}^{\circ z}_{j,\xi,t} & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Process_ramp_down_allowance#
Process_ramp_down_allowance:
description: >-
how far a process may lower internal power between two snapshots. It is
its ramp limit of the build while it stays on, plus its shut-down ramp
in the snapshot it turns off
dims: [scenario, snapshot, process]
cases:
committed:
when: Process_committable
expression: >-
Process_ramp_down_rate * Process_p_nom_committed * Process_status
+ Process_shut_down_rate * Process_p_nom_committed
* (Process_ramp_previous_status - Process_status)
otherwise: Process_ramp_down_rate * Process_p_nom_effective
\[
\Delta^{z,-}_{j,\xi,t} = \begin{cases} \widetilde{\mathrm{rd}}^{z}_{j,\xi,t} \cdot \widehat{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} \cdot u^{z}_{j,\xi,t} + \widetilde{\mathrm{rd}}^{z,\mathrm{dn}}_{j,\xi} \cdot \widehat{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} \cdot \left( \overleftarrow{\bar u}^{z}_{j,\xi,t} - u^{z}_{j,\xi,t} \right) & \text{if } \mathrm{com}^{z}_{j} \\ \widetilde{\mathrm{rd}}^{z}_{j,\xi,t} \cdot \widetilde{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Process_ramp_down_rate#
Process_ramp_down_rate:
description: >-
the ramp limit in a process's down row, PyPSA's `ramp_limit_down`. It is
the full build where PyPSA has none, since a shut-down ramp alone builds
the row
dims: [scenario, snapshot, process]
cases:
given: { when: Process_ramp_limit_down, expression: Process_ramp_limit_down }
otherwise: 1
\[
\widetilde{\mathrm{rd}}^{z}_{j,\xi,t} = \begin{cases} \mathrm{rd}^{z}_{j,\xi,t} & \text{if } \mathrm{rd}^{z}_{j,\xi,t} \text{ is defined} \\ 1 & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Process_ramp_previous_status#
Process_ramp_previous_status:
description: >-
the status a committed process's ramp rows carry into a snapshot. At the
first snapshot, a fixed modular process that came in running has every
module on, since the state it brought in is on or off
(`constraints.py:1152-1157`). The transition and down-time rows count it as one module
(PyPSA/PyPSA#2114). It is the state carried in otherwise
dims: [scenario, snapshot, process]
cases:
opening_modules:
when: "position(snapshot) == 0 AND NOT Process_p_nom_extendable AND Process_p_nom_mod > 0"
expression: Process_status_initial * Process_modules_installed
otherwise: Process_previous_status
\[
\overleftarrow{\bar u}^{z}_{j,\xi,t} = \begin{cases} \mathrm{u}^{z,0}_{j,\xi} \cdot \mathrm{N}^{z,\mathrm{fix}}_{j,\xi} & \text{if } \mathrm{pos}(t) = 0 \wedge \neg \mathrm{ext}^{z}_{j} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0 \\ \overleftarrow{u}^{z}_{j,\xi,t} & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Process_ramp_up_allowance#
Process_ramp_up_allowance:
description: >-
how far a process may raise internal power between two snapshots. It is
its ramp limit of the build while it stays on, plus its start-up ramp in
the snapshot it turns on
dims: [scenario, snapshot, process]
cases:
committed:
when: Process_committable
expression: >-
Process_ramp_up_rate * Process_p_nom_committed * Process_ramp_previous_status
+ Process_start_up_rate * Process_p_nom_committed
* (Process_status - Process_ramp_previous_status)
otherwise: Process_ramp_up_rate * Process_p_nom_effective
\[
\Delta^{z,+}_{j,\xi,t} = \begin{cases} \widetilde{\mathrm{ru}}^{z}_{j,\xi,t} \cdot \widehat{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} \cdot \overleftarrow{\bar u}^{z}_{j,\xi,t} + \widetilde{\mathrm{ru}}^{z,\mathrm{up}}_{j,\xi} \cdot \widehat{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} \cdot \left( u^{z}_{j,\xi,t} - \overleftarrow{\bar u}^{z}_{j,\xi,t} \right) & \text{if } \mathrm{com}^{z}_{j} \\ \widetilde{\mathrm{ru}}^{z}_{j,\xi,t} \cdot \widetilde{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Process_ramp_up_rate#
Process_ramp_up_rate:
description: >-
the ramp limit in a process's up row, PyPSA's `ramp_limit_up`. It is the
full build where PyPSA has none, since a start-up ramp alone builds the
row
dims: [scenario, snapshot, process]
cases:
given: { when: Process_ramp_limit_up, expression: Process_ramp_limit_up }
otherwise: 1
\[
\widetilde{\mathrm{ru}}^{z}_{j,\xi,t} = \begin{cases} \mathrm{ru}^{z}_{j,\xi,t} & \text{if } \mathrm{ru}^{z}_{j,\xi,t} \text{ is defined} \\ 1 & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Process_shut_down_rate#
Process_shut_down_rate:
description: >-
the shut-down ramp in a process's down row (PyPSA's
`ramp_limit_shut_down`), or the full build where it has none
dims: [scenario, process]
cases:
given: { when: Process_ramp_limit_shut_down, expression: Process_ramp_limit_shut_down }
otherwise: 1
\[
\widetilde{\mathrm{rd}}^{z,\mathrm{dn}}_{j,\xi} = \begin{cases} \mathrm{rd}^{z,\mathrm{dn}}_{j,\xi} & \text{if } \mathrm{rd}^{z,\mathrm{dn}}_{j,\xi} \text{ is defined} \\ 1 & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi
\]
Process_start_up_rate#
Process_start_up_rate:
description: >-
the start-up ramp in a process's up row (PyPSA's `ramp_limit_start_up`),
or the full build where it has none
dims: [scenario, process]
cases:
given: { when: Process_ramp_limit_start_up, expression: Process_ramp_limit_start_up }
otherwise: 1
\[
\widetilde{\mathrm{ru}}^{z,\mathrm{up}}_{j,\xi} = \begin{cases} \mathrm{ru}^{z,\mathrm{up}}_{j,\xi} & \text{if } \mathrm{ru}^{z,\mathrm{up}}_{j,\xi} \text{ is defined} \\ 1 & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi
\]
Process_status_carried_over#
Process_status_carried_over:
description: >-
the state a process carries over into a snapshot: the state it brought
into the horizon at the first, the previous snapshot's after that
dims: [scenario, snapshot, process]
cases:
opening: { when: "position(snapshot) == 0", expression: Process_status_initial }
otherwise: shift(Process_status, along=snapshot, offset=1)
\[
\overleftarrow{u}^{\circ z}_{j,\xi,t} = \begin{cases} \mathrm{u}^{z,0}_{j,\xi} & \text{if } \mathrm{pos}(t) = 0 \\ u^{z}_{j,\xi,t - 1} & \text{otherwise} \end{cases} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Process_tech_capacity_expansion#
Process_tech_capacity_expansion:
expression: sum(Process_p_nom_ext * Process_tech_capacity_weight, over=process)
\[
\mathit{Process\_tech\_capacity\_expansion}_{o} = \sum_{j \in \mathcal{J}} Z_{j} \cdot \mathrm{m}^{z}_{o,j} \qquad \forall\, o \in \mathcal{O}
\]
StorageUnit_additions#
StorageUnit_additions:
expression: >-
sum(StorageUnit_p_nom_ext * StorageUnit_first_active, over=storage_unit, by=StorageUnit_carrier[carrier])
\[
\mathit{StorageUnit\_additions}_{i,y} = \sum_{s \in \mathcal{S} \,:\, \mathrm{StorageUnit\_carrier}(s) = i} H_{s} \cdot \mathrm{new}^{h}_{y,s} \qquad \forall\, i \in \mathcal{I},\ y \in \mathcal{Y}
\]
StorageUnit_capex#
StorageUnit_capex:
expression: sum(scenario_weight * StorageUnit_p_nom_ext * StorageUnit_capital_cost * StorageUnit_capital_weight)
\[
\mathit{StorageUnit\_capex} = \sum_{\xi \in \Xi,\ s \in \mathcal{S}} \pi_{\xi} \cdot H_{s} \cdot \mathrm{c}^{\mathrm{cap},h}_{\xi,s} \cdot \mathrm{W}^{h}_{s}
\]
StorageUnit_charge_carried_in#
StorageUnit_charge_carried_in:
description: >-
the charge a unit opens a snapshot with. At the first snapshot it is active
in, it is its last such snapshot's less standing loss where it is
cyclic. Where it is not, it is the given initial charge, which no
standing loss has touched yet. Otherwise it is the previous snapshot's
less standing loss. A unit built in a later period opens in that period,
and a cyclic one that retires closes on its own last snapshot. Per
period, the same holds with each investment period as the horizon
dims: [scenario, snapshot, storage_unit]
cases:
cyclic:
when: >-
StorageUnit_cyclic_state_of_charge AND NOT StorageUnit_cyclic_state_of_charge_per_period
AND NOT StorageUnit_state_of_charge_initial_per_period
AND (position(snapshot) == 0 OR StorageUnit_opens_late)
expression: >-
StorageUnit_retention
* shift(shift(StorageUnit_state_of_charge, along=snapshot, offset=1, edge='wrap'), along=snapshot, offset=StorageUnit_inactive_snapshots, edge='wrap')
opening:
when: >-
NOT StorageUnit_cyclic_state_of_charge AND NOT StorageUnit_cyclic_state_of_charge_per_period
AND NOT StorageUnit_state_of_charge_initial_per_period
AND (position(snapshot) == 0 OR StorageUnit_opens_late)
expression: StorageUnit_state_of_charge_initial
period_cyclic:
when: StorageUnit_cyclic_state_of_charge_per_period
expression: >-
StorageUnit_retention
* shift(StorageUnit_state_of_charge, along=snapshot, offset=1, edge='wrap', within=snapshot_period[period])
period_opening:
when: >-
StorageUnit_state_of_charge_initial_per_period AND NOT StorageUnit_cyclic_state_of_charge_per_period
AND position(snapshot, within=snapshot_period[period]) == 0
expression: StorageUnit_state_of_charge_initial
otherwise: StorageUnit_retention * shift(StorageUnit_state_of_charge, along=snapshot, offset=1)
\[
\overleftarrow{\mathit{soc}}_{\xi,t,s} = \begin{cases} \rho_{\xi,t,s} \cdot \mathit{soc}_{\xi,\left( t \ominus \mathrm{idle} \right) \ominus 1,s} & \text{if } \mathrm{cyc}_{\xi,s} \wedge \neg \mathrm{cyc}^{y}_{\xi,s} \wedge \neg \mathrm{reset}_{\xi,s} \wedge \left( \mathrm{pos}(t) = 0 \vee \mathrm{open}_{t,s} \right) \\ \mathrm{soc}^{0}_{\xi,s} & \text{if } \neg \mathrm{cyc}_{\xi,s} \wedge \neg \mathrm{cyc}^{y}_{\xi,s} \wedge \neg \mathrm{reset}_{\xi,s} \wedge \left( \mathrm{pos}(t) = 0 \vee \mathrm{open}_{t,s} \right) \\ \rho_{\xi,t,s} \cdot \mathit{soc}_{\xi,t \ominus^{\mathrm{snapshot\_period}(t)} 1,s} & \text{if } \mathrm{cyc}^{y}_{\xi,s} \\ \mathrm{soc}^{0}_{\xi,s} & \text{if } \mathrm{reset}_{\xi,s} \wedge \neg \mathrm{cyc}^{y}_{\xi,s} \wedge \mathrm{pos}_{\mathrm{snapshot\_period}(t)}(t) = 0 \\ \rho_{\xi,t,s} \cdot \mathit{soc}_{\xi,t - 1,s} & \text{otherwise} \end{cases} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S}
\]
StorageUnit_closing_weight#
StorageUnit_closing_weight:
description: >-
what the charge a unit holds at a snapshot counts for in a row as its
closing level. Where the unit reopens per period, it is the years of the
period at the last snapshot of each counted period. Where it does not,
it is one at the last counted snapshot it is active in. It is nothing
elsewhere
dims: [scenario, global_constraint, snapshot, storage_unit]
cases:
per_period:
when: StorageUnit_state_of_charge_initial_per_period AND GlobalConstraint_counts_snapshot AND position(snapshot, within=snapshot_period[period]) == -1
expression: at(period_weight_years, by=snapshot_period[period])
carried_over:
when: NOT StorageUnit_state_of_charge_initial_per_period
expression: StorageUnit_last_counted_active
otherwise: 0
\[
\mathit{w}^{h}_{o,\xi,t,s} = \begin{cases} \mathit{last}^{h}_{o,\xi,t,s} & \text{if } \neg \mathrm{reset}_{\xi,s} \\ \mathrm{w}^{\mathrm{yr}}_{\mathrm{snapshot\_period}(t)} & \text{if } \mathrm{reset}_{\xi,s} \wedge \mathrm{in}_{o,\xi,t} \wedge \mathrm{pos}_{\mathrm{snapshot\_period}(t)}(t) = \lvert \mathcal{T}_{\mathrm{snapshot\_period}(t)} \rvert - 1 \\ 0 & \text{otherwise} \end{cases} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S}
\]
StorageUnit_injection#
StorageUnit_injection:
expression: >-
sum(StorageUnit_sign * (StorageUnit_p_dispatch - StorageUnit_p_store), over=storage_unit, by=StorageUnit_bus[bus])
\[
\mathit{StorageUnit\_injection}_{n,\xi,t} = \sum_{s \in \mathcal{S} \,:\, \mathrm{StorageUnit\_bus}(s) = n} \mathrm{sgn}^{h}_{s} \cdot \left( h^{+}_{\xi,t,s} - h^{-}_{\xi,t,s} \right) \qquad \forall\, n \in \mathcal{N},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
StorageUnit_last_counted_active#
StorageUnit_last_counted_active:
description: >-
one at the last snapshot a row counts where a unit is active, and zero
elsewhere. A unit that retires before the last counted snapshot closes
on its last active level. PyPSA forward-fills the level over the counted
snapshots (`global_constraints.py:476-479`)
dims: [scenario, global_constraint, snapshot, storage_unit]
cases:
last:
when: >-
GlobalConstraint_counts_snapshot AND StorageUnit_active
AND NOT shift(GlobalConstraint_counts_snapshot AND StorageUnit_active, along=snapshot, offset=-1)
expression: 1
otherwise: 0
\[
\mathit{last}^{h}_{o,\xi,t,s} = \begin{cases} 1 & \text{if } \mathrm{in}_{o,\xi,t} \wedge \mathrm{on}^{h}_{t,s} \wedge \neg \left( \mathrm{in}_{o,\xi,t + 1} \wedge \mathrm{on}^{h}_{t + 1,s} \right) \\ 0 & \text{otherwise} \end{cases} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S}
\]
StorageUnit_operational_limit#
StorageUnit_operational_limit:
expression: >-
-sum(sum((StorageUnit_state_of_charge * StorageUnit_closing_weight) * StorageUnit_operational_limit_weight, over=snapshot), over=storage_unit)
\[
\mathit{StorageUnit\_operational\_limit}_{o,\xi} = -\left( \sum_{s \in \mathcal{S}} \sum_{t \in \mathcal{T}} \mathit{soc}_{\xi,t,s} \cdot \mathit{w}^{h}_{o,\xi,t,s} \cdot \mathrm{b}^{h}_{o,\xi,s} \right) \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
StorageUnit_opex#
StorageUnit_opex:
expression: >-
sum(sum(((StorageUnit_p_dispatch * StorageUnit_marginal_cost) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=storage_unit), over=snapshot)
+ sum(sum((((StorageUnit_p_dispatch * StorageUnit_p_dispatch) * StorageUnit_marginal_cost_quadratic) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=storage_unit), over=snapshot)
+ sum(sum(((StorageUnit_state_of_charge * StorageUnit_marginal_cost_storage) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=storage_unit), over=snapshot)
+ sum(sum(((StorageUnit_spill * StorageUnit_spill_cost) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=storage_unit), over=snapshot)
\[
\mathit{StorageUnit\_opex}_{\xi} = \sum_{t \in \mathcal{T}} \sum_{s \in \mathcal{S}} h^{+}_{\xi,t,s} \cdot \mathrm{c}^{h}_{\xi,t,s} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{s \in \mathcal{S}} h^{+}_{\xi,t,s} \cdot h^{+}_{\xi,t,s} \cdot \mathrm{c}^{h,(2)}_{\xi,t,s} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{s \in \mathcal{S}} \mathit{soc}_{\xi,t,s} \cdot \mathrm{c}^{\mathrm{soc}}_{\xi,t,s} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{s \in \mathcal{S}} \mathit{spill}_{\xi,t,s} \cdot \mathrm{c}^{\mathrm{spill}}_{\xi,t,s} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} \qquad \forall\, \xi \in \Xi
\]
StorageUnit_primary_energy#
StorageUnit_primary_energy:
expression: >-
-sum(sum((StorageUnit_state_of_charge * StorageUnit_closing_weight) * StorageUnit_primary_energy_weight, over=snapshot), over=storage_unit)
\[
\mathit{StorageUnit\_primary\_energy}_{o,\xi} = -\left( \sum_{s \in \mathcal{S}} \sum_{t \in \mathcal{T}} \mathit{soc}_{\xi,t,s} \cdot \mathit{w}^{h}_{o,\xi,t,s} \cdot \mathrm{a}^{h}_{o,\xi,s} \right) \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
StorageUnit_tech_capacity_expansion#
StorageUnit_tech_capacity_expansion:
expression: sum(StorageUnit_p_nom_ext * StorageUnit_tech_capacity_weight, over=storage_unit)
\[
\mathit{StorageUnit\_tech\_capacity\_expansion}_{o} = \sum_{s \in \mathcal{S}} H_{s} \cdot \mathrm{m}^{h}_{o,s} \qquad \forall\, o \in \mathcal{O}
\]
Store_additions#
Store_additions:
expression: >-
sum(Store_e_nom_ext * Store_first_active, over=store, by=Store_carrier[carrier])
\[
\mathit{Store\_additions}_{i,y} = \sum_{v \in \mathcal{V} \,:\, \mathrm{Store\_carrier}(v) = i} E_{v} \cdot \mathrm{new}^{e}_{y,v} \qquad \forall\, i \in \mathcal{I},\ y \in \mathcal{Y}
\]
Store_capex#
Store_capex:
expression: sum(scenario_weight * Store_e_nom_ext * Store_capital_cost * Store_capital_weight)
\[
\mathit{Store\_capex} = \sum_{\xi \in \Xi,\ v \in \mathcal{V}} \pi_{\xi} \cdot E_{v} \cdot \mathrm{c}^{\mathrm{cap},e}_{\xi,v} \cdot \mathrm{W}^{e}_{v}
\]
Store_charge_cap_pu#
Store_charge_cap_pu:
description: >-
the most charging through `p_store`, per unit of `e_nom`. It is
`-p_min_pu / max_hours` for a store that splits its dispatch, and zero
otherwise (`stores.py:79-86`)
dims: [scenario, snapshot, store]
cases:
split: { when: Store_split_dispatch, expression: -Store_p_min_pu / Store_max_hours }
otherwise: 0
\[
\overline{\mathrm{q}}^{-}_{\xi,t,v} = \begin{cases} \frac{-\underline{\mathrm{q}}_{\xi,t,v}}{\mathrm{T}^{e}_{\xi,v}} & \text{if } \mathrm{split}^{e}_{\xi,v} \\ 0 & \text{otherwise} \end{cases} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V}
\]
Store_closing_weight#
Store_closing_weight:
description: >-
what the energy a store holds at a snapshot counts for in a row as its
closing level. Where the store reopens per period, it is the years of
the period at the last snapshot of each counted period. Where it does
not, it is one at the last counted snapshot it is active in. It is nothing
elsewhere
dims: [scenario, global_constraint, snapshot, store]
cases:
per_period:
when: Store_e_initial_per_period AND GlobalConstraint_counts_snapshot AND position(snapshot, within=snapshot_period[period]) == -1
expression: at(period_weight_years, by=snapshot_period[period])
carried_over:
when: NOT Store_e_initial_per_period
expression: Store_last_counted_active
otherwise: 0
\[
\mathit{w}^{e}_{o,\xi,t,v} = \begin{cases} \mathit{last}^{e}_{o,\xi,t,v} & \text{if } \neg \mathrm{reset}^{e}_{\xi,v} \\ \mathrm{w}^{\mathrm{yr}}_{\mathrm{snapshot\_period}(t)} & \text{if } \mathrm{reset}^{e}_{\xi,v} \wedge \mathrm{in}_{o,\xi,t} \wedge \mathrm{pos}_{\mathrm{snapshot\_period}(t)}(t) = \lvert \mathcal{T}_{\mathrm{snapshot\_period}(t)} \rvert - 1 \\ 0 & \text{otherwise} \end{cases} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V}
\]
Store_discharge_floor_pu#
Store_discharge_floor_pu:
description: >-
the least discharge, per unit of `e_nom`. It is zero for a store that
splits its dispatch, so `p_store` carries the charge. Otherwise it is the
most charging as a negative discharge, `p_min_pu / max_hours`
(`stores.py:79-86`)
dims: [scenario, snapshot, store]
cases:
split: { when: Store_split_dispatch, expression: 0 }
otherwise: Store_p_min_pu / Store_max_hours
\[
\underline{\mathrm{q}}^{+}_{\xi,t,v} = \begin{cases} 0 & \text{if } \mathrm{split}^{e}_{\xi,v} \\ \frac{\underline{\mathrm{q}}_{\xi,t,v}}{\mathrm{T}^{e}_{\xi,v}} & \text{otherwise} \end{cases} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V}
\]
Store_e_nom_effective#
Store_e_nom_effective:
description: "the build a store's limits are taken against: the chosen one where it is extendable, the given one otherwise"
dims: [scenario, store]
cases:
extendable: { when: Store_e_nom_extendable, expression: Store_e_nom_ext }
otherwise: Store_e_nom
\[
\widetilde{\mathrm{e}}^{\mathrm{nom}}_{\xi,v} = \begin{cases} E_{v} & \text{if } \mathrm{ext}^{e}_{v} \\ \mathrm{e}^{\mathrm{nom}}_{\xi,v} & \text{otherwise} \end{cases} \qquad \forall\, \xi \in \Xi,\ v \in \mathcal{V}
\]
Store_energy_carried_in#
Store_energy_carried_in:
description: >-
the energy a store opens a snapshot with. At the first snapshot it
is active in, it is its last such snapshot's less standing loss where it is
cyclic. Where it is not, it is the given initial energy, which no
standing loss has touched yet. Otherwise it is the previous snapshot's
less standing loss. A store built in a later period opens in that
period, and a cyclic one that retires closes on its own last snapshot.
Per period, the same holds with each investment period as the horizon
dims: [scenario, snapshot, store]
cases:
cyclic:
when: >-
Store_e_cyclic AND NOT Store_e_cyclic_per_period AND NOT Store_e_initial_per_period
AND (position(snapshot) == 0 OR Store_opens_late)
expression: >-
Store_retention
* shift(shift(Store_e, along=snapshot, offset=1, edge='wrap'), along=snapshot, offset=Store_inactive_snapshots, edge='wrap')
opening:
when: >-
NOT Store_e_cyclic AND NOT Store_e_cyclic_per_period AND NOT Store_e_initial_per_period
AND (position(snapshot) == 0 OR Store_opens_late)
expression: Store_e_initial
period_cyclic:
when: Store_e_cyclic_per_period
expression: Store_retention * shift(Store_e, along=snapshot, offset=1, edge='wrap', within=snapshot_period[period])
period_opening:
when: Store_e_initial_per_period AND NOT Store_e_cyclic_per_period AND position(snapshot, within=snapshot_period[period]) == 0
expression: Store_e_initial
otherwise: Store_retention * shift(Store_e, along=snapshot, offset=1)
\[
\overleftarrow{e}_{\xi,t,v} = \begin{cases} \rho^{e}_{\xi,t,v} \cdot e_{\xi,\left( t \ominus \mathrm{idle}^{e} \right) \ominus 1,v} & \text{if } \mathrm{cyc}^{e}_{\xi,v} \wedge \neg \mathrm{cyc}^{e,y}_{\xi,v} \wedge \neg \mathrm{reset}^{e}_{\xi,v} \wedge \left( \mathrm{pos}(t) = 0 \vee \mathrm{open}^{e}_{t,v} \right) \\ \mathrm{e}^{0}_{\xi,v} & \text{if } \neg \mathrm{cyc}^{e}_{\xi,v} \wedge \neg \mathrm{cyc}^{e,y}_{\xi,v} \wedge \neg \mathrm{reset}^{e}_{\xi,v} \wedge \left( \mathrm{pos}(t) = 0 \vee \mathrm{open}^{e}_{t,v} \right) \\ \rho^{e}_{\xi,t,v} \cdot e_{\xi,t \ominus^{\mathrm{snapshot\_period}(t)} 1,v} & \text{if } \mathrm{cyc}^{e,y}_{\xi,v} \\ \mathrm{e}^{0}_{\xi,v} & \text{if } \mathrm{reset}^{e}_{\xi,v} \wedge \neg \mathrm{cyc}^{e,y}_{\xi,v} \wedge \mathrm{pos}_{\mathrm{snapshot\_period}(t)}(t) = 0 \\ \rho^{e}_{\xi,t,v} \cdot e_{\xi,t - 1,v} & \text{otherwise} \end{cases} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V}
\]
Store_injection#
\[
\mathit{Store\_injection}_{n,\xi,t} = \sum_{v \in \mathcal{V} \,:\, \mathrm{Store\_bus}(v) = n} \mathrm{sgn}^{q}_{v} \cdot q_{\xi,t,v} \qquad \forall\, n \in \mathcal{N},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Store_last_counted_active#
Store_last_counted_active:
description: >-
one at the last snapshot a row counts where a store is active, and zero
elsewhere. A store that retires before the last counted snapshot closes
on its last active level. PyPSA forward-fills the level over the counted
snapshots (`global_constraints.py:530-533`)
dims: [scenario, global_constraint, snapshot, store]
cases:
last:
when: >-
GlobalConstraint_counts_snapshot AND Store_active
AND NOT shift(GlobalConstraint_counts_snapshot AND Store_active, along=snapshot, offset=-1)
expression: 1
otherwise: 0
\[
\mathit{last}^{e}_{o,\xi,t,v} = \begin{cases} 1 & \text{if } \mathrm{in}_{o,\xi,t} \wedge \mathrm{on}^{e}_{t,v} \wedge \neg \left( \mathrm{in}_{o,\xi,t + 1} \wedge \mathrm{on}^{e}_{t + 1,v} \right) \\ 0 & \text{otherwise} \end{cases} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V}
\]
Store_operational_limit#
Store_operational_limit:
expression: >-
-sum(sum((Store_e * Store_closing_weight) * Store_operational_limit_weight, over=snapshot), over=store)
\[
\mathit{Store\_operational\_limit}_{o,\xi} = -\left( \sum_{v \in \mathcal{V}} \sum_{t \in \mathcal{T}} e_{\xi,t,v} \cdot \mathit{w}^{e}_{o,\xi,t,v} \cdot \mathrm{b}^{e}_{o,\xi,v} \right) \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
Store_opex#
Store_opex:
expression: >-
sum(sum(((Store_p * Store_marginal_cost) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=store), over=snapshot)
+ sum(sum((((Store_p * Store_p) * Store_marginal_cost_quadratic) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=store), over=snapshot)
+ sum(sum(((Store_e * Store_marginal_cost_storage) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=store), over=snapshot)
+ sum(sum((((Store_p + Store_p_store) * Store_marginal_cost_dispatch) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=store), over=snapshot)
+ sum(sum(((Store_p_store * Store_marginal_cost_store) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=store), over=snapshot)
+ sum(sum(((Store_spill * Store_spill_cost) * snapshot_weightings_objective) * at(period_weight_objective, by=snapshot_period[period]), over=store), over=snapshot)
\[
\mathit{Store\_opex}_{\xi} = \sum_{t \in \mathcal{T}} \sum_{v \in \mathcal{V}} q_{\xi,t,v} \cdot \mathrm{c}^{q}_{\xi,t,v} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{v \in \mathcal{V}} q_{\xi,t,v} \cdot q_{\xi,t,v} \cdot \mathrm{c}^{q,(2)}_{\xi,t,v} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{v \in \mathcal{V}} e_{\xi,t,v} \cdot \mathrm{c}^{e}_{\xi,t,v} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{v \in \mathcal{V}} \left( q_{\xi,t,v} + q^{-}_{\xi,t,v} \right) \cdot \mathrm{c}^{q,+}_{\xi,t,v} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{v \in \mathcal{V}} q^{-}_{\xi,t,v} \cdot \mathrm{c}^{q,-}_{\xi,t,v} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} + \sum_{t \in \mathcal{T}} \sum_{v \in \mathcal{V}} \mathit{spill}^{e}_{\xi,t,v} \cdot \mathrm{c}^{\mathrm{spill},e}_{\xi,t,v} \cdot \mathrm{w}_{t} \cdot \mathrm{w}^{y}_{\mathrm{snapshot\_period}(t)} \qquad \forall\, \xi \in \Xi
\]
Store_primary_energy#
Store_primary_energy:
expression: >-
-sum(sum((Store_e * Store_closing_weight) * Store_primary_energy_weight, over=snapshot), over=store)
\[
\mathit{Store\_primary\_energy}_{o,\xi} = -\left( \sum_{v \in \mathcal{V}} \sum_{t \in \mathcal{T}} e_{\xi,t,v} \cdot \mathit{w}^{e}_{o,\xi,t,v} \cdot \mathrm{a}^{e}_{o,\xi,v} \right) \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
Store_tech_capacity_expansion#
Store_tech_capacity_expansion:
expression: sum(Store_e_nom_ext * Store_tech_capacity_weight, over=store)
\[
\mathit{Store\_tech\_capacity\_expansion}_{o} = \sum_{v \in \mathcal{V}} E_{v} \cdot \mathrm{m}^{e}_{o,v} \qquad \forall\, o \in \mathcal{O}
\]
Transformer_angle_sum#
Transformer_angle_sum:
expression: >-
sum(Transformer_s * at(Transformer_cycle_weight, by=snapshot_period[period]), over=transformer)
+ sum(Transformer_phase_shift_weight, over=transformer)
+ sum(Transformer_phase_shift * at(Transformer_phase_shift_cycle_weight, by=snapshot_period[period]), over=transformer)
\[
\mathit{Transformer\_angle\_sum}_{c,\xi,t} = \sum_{m \in \mathcal{M}} \sigma_{\xi,t,m} \cdot \mathrm{x}^{\sigma}_{c,\mathrm{snapshot\_period}(t),m} + \sum_{m \in \mathcal{M}} \vartheta_{c,\xi,t,m} + \sum_{m \in \mathcal{M}} \mathit{Transformer\_phase\_shift}_{\xi,t,m} \cdot \mathrm{Transformer\_phase\_shift\_cycle\_weight}_{c,\mathrm{snapshot\_period}(t),\xi,m} \qquad \forall\, c \in \mathcal{C},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Transformer_capex#
Transformer_capex:
expression: sum(scenario_weight * Transformer_s_nom_ext * Transformer_capital_cost * Transformer_capital_weight)
\[
\mathit{Transformer\_capex} = \sum_{\xi \in \Xi,\ m \in \mathcal{M}} \pi_{\xi} \cdot \Sigma_{m} \cdot \mathrm{c}^{\mathrm{cap},\sigma}_{\xi,m} \cdot \mathrm{W}^{\sigma}_{m}
\]
Transformer_injection#
Transformer_injection:
expression: >-
-sum(Transformer_s, over=transformer, by=Transformer_bus0[bus])
+ sum(Transformer_s, over=transformer, by=Transformer_bus1[bus])
- (0.5 * sum(Transformer_loss, over=transformer, by=Transformer_bus0[bus]))
- (0.5 * sum(Transformer_loss, over=transformer, by=Transformer_bus1[bus]))
\[
\mathit{Transformer\_injection}_{n,\xi,t} = -\left( \sum_{m \in \mathcal{M} \,:\, \mathrm{Transformer\_bus0}(m) = n} \sigma_{\xi,t,m} \right) + \sum_{m \in \mathcal{M} \,:\, \mathrm{Transformer\_bus1}(m) = n} \sigma_{\xi,t,m} - 0.5 \cdot \left( \sum_{m \in \mathcal{M} \,:\, \mathrm{Transformer\_bus0}(m) = n} \ell^{\sigma}_{\xi,t,m} \right) - 0.5 \cdot \left( \sum_{m \in \mathcal{M} \,:\, \mathrm{Transformer\_bus1}(m) = n} \ell^{\sigma}_{\xi,t,m} \right) \qquad \forall\, n \in \mathcal{N},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Transformer_s_monitored#
Transformer_s_monitored:
description: the flow a transformer's post-contingency rows read, as a line's
dims: [scenario, snapshot, transformer]
cases:
standing: { when: Transformer_active, expression: Transformer_s }
otherwise: 0
\[
\check{\sigma}_{\xi,t,m} = \begin{cases} \sigma_{\xi,t,m} & \text{if } \mathrm{on}^{\sigma}_{t,m} \\ 0 & \text{otherwise} \end{cases} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M}
\]
energy_limit_weight#
energy_limit_weight:
description: >-
what one unit of power at a snapshot counts for in an `e_sum` row: the
generator weighting times the years of the snapshot's period, as in an
`operational_limit` row (`constraints.py:2935-2937`)
dims: [snapshot]
expression: snapshot_weightings_generators * at(period_weight_years, by=snapshot_period[period])
\[
\mathit{w}^{E}_{t} = \mathrm{w}^{\mathrm{gen}}_{t} \cdot \mathrm{w}^{\mathrm{yr}}_{\mathrm{snapshot\_period}(t)} \qquad \forall\, t \in \mathcal{T}
\]
operational_limit#
operational_limit:
dims: [scenario, global_constraint]
expression: >-
Generator_operational_limit
+ StorageUnit_operational_limit
+ Store_operational_limit
description: >-
what an `operational_limit` row totals: the weighted energy its
generators deliver over the snapshots it counts, plus what its
non-cyclic storage draws down. The initial charge it draws from is
folded into the row's constant
\[
\mathit{operational\_limit}_{o,\xi} = \mathit{Generator\_operational\_limit}_{o,\xi} + \mathit{StorageUnit\_operational\_limit}_{o,\xi} + \mathit{Store\_operational\_limit}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
primary_energy#
primary_energy:
dims: [scenario, global_constraint]
expression: Generator_primary_energy + StorageUnit_primary_energy + Store_primary_energy
description: >-
what a `primary_energy` row totals: weighted generator energy over the
snapshots it counts, less the charge left in weighted storage at the
close. The initial charge it is compared against is folded into the
row's constant
\[
\mathit{primary\_energy}_{o,\xi} = \mathit{Generator\_primary\_energy}_{o,\xi} + \mathit{StorageUnit\_primary\_energy}_{o,\xi} + \mathit{Store\_primary\_energy}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
risk_weighted_opex#
risk_weighted_opex:
expression: (1 - CVaR_omega) * sum(scenario_weight * scenario_opex, over=scenario) + CVaR_omega * CVaR
\[
\mathit{risk\_weighted\_opex} = \left( 1 - \omega \right) \cdot \left( \sum_{\xi \in \Xi} \pi_{\xi} \cdot \mathit{scenario\_opex}_{\xi} \right) + \omega \cdot CVaR
\]
scenario_opex#
scenario_opex:
dims: [scenario]
expression: >-
Generator_opex
+ Generator_commitment_opex
+ Link_opex
+ Link_commitment_opex
+ Process_opex
+ Process_commitment_opex
+ StorageUnit_opex
+ Store_opex
description: >-
what a future costs to run: every operating term, weighted by the
snapshot's hours and its period, before the scenario's own weight. A
start and a stop cost what they cost, unweighted, as PyPSA adds them
(`optimize.py:448-468`)
\[
\mathit{scenario\_opex}_{\xi} = \mathit{Generator\_opex}_{\xi} + \mathit{Generator\_commitment\_opex}_{\xi} + \mathit{Link\_opex}_{\xi} + \mathit{Link\_commitment\_opex}_{\xi} + \mathit{Process\_opex}_{\xi} + \mathit{Process\_commitment\_opex}_{\xi} + \mathit{StorageUnit\_opex}_{\xi} + \mathit{Store\_opex}_{\xi} \qquad \forall\, \xi \in \Xi
\]
tech_capacity_expansion#
tech_capacity_expansion:
dims: [global_constraint]
expression: >-
Generator_tech_capacity_expansion
+ Line_tech_capacity_expansion
+ Link_tech_capacity_expansion
+ Process_tech_capacity_expansion
+ StorageUnit_tech_capacity_expansion
+ Store_tech_capacity_expansion
description: >-
what a `tech_capacity_expansion_limit` row totals: the chosen build of
the row's carrier-and-bus set
\[
\mathit{tech\_capacity\_expansion}_{o} = \mathit{Generator\_tech\_capacity\_expansion}_{o} + \mathit{Line\_tech\_capacity\_expansion}_{o} + \mathit{Link\_tech\_capacity\_expansion}_{o} + \mathit{Process\_tech\_capacity\_expansion}_{o} + \mathit{StorageUnit\_tech\_capacity\_expansion}_{o} + \mathit{Store\_tech\_capacity\_expansion}_{o} \qquad \forall\, o \in \mathcal{O}
\]
total_cost#
total_cost:
dims: []
expression: >-
Generator_capex
+ Link_capex
+ StorageUnit_capex
+ Store_capex
+ Line_capex
+ Process_capex
+ Transformer_capex
+ risk_weighted_opex
description: >-
what the system costs. Capacity counts once per active period at its
expected cost over the scenarios. Operation counts in expectation over
the scenarios, and a share of it at the tail
\[
\mathit{total\_cost} = \mathit{Generator\_capex} + \mathit{Link\_capex} + \mathit{StorageUnit\_capex} + \mathit{Store\_capex} + \mathit{Line\_capex} + \mathit{Process\_capex} + \mathit{Transformer\_capex} + \mathit{risk\_weighted\_opex}
\]
transmission_expansion_cost#
transmission_expansion_cost:
dims: [scenario, global_constraint]
expression: Line_transmission_expansion_cost + Link_transmission_expansion_cost
description: >-
what a `transmission_expansion_cost_limit` row totals: capital cost
times the chosen build of the row's branches
\[
\mathit{transmission\_expansion\_cost}_{o,\xi} = \mathit{Line\_transmission\_expansion\_cost}_{o,\xi} + \mathit{Link\_transmission\_expansion\_cost}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
transmission_volume_expansion#
transmission_volume_expansion:
dims: [scenario, global_constraint]
expression: Line_transmission_volume_expansion + Link_transmission_volume_expansion
description: >-
what a `transmission_volume_expansion_limit` row totals: length times
the chosen build of the row's branches
\[
\mathit{transmission\_volume\_expansion}_{o,\xi} = \mathit{Line\_transmission\_volume\_expansion}_{o,\xi} + \mathit{Link\_transmission\_volume\_expansion}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi
\]
Objective#
\[
\min \mathit{total\_cost}
\]
Bus-nodal_balance#
Bus_nodal_balance
Bus_nodal_balance:
description: >-
`Bus-nodal_balance` — what is generated at a bus meets the load there,
less half of every incident line's and transformer's loss. What is
generated includes storage dispatch and stores. It is net of what the
links take away, plus what arrives over them after losses and any delay
at every port they deliver to. Each process port draws or delivers at
its own rate, and each passive branch carries its flow. PyPSA dissipates
a branch's loss half at either end. Each generator, storage unit, store
and load term enters with its component's `sign`
(`constraints.py:1480-1481`, `:1592`), and an inactive load not at all.
A bus nothing is attached to has no row. PyPSA refuses one that carries
load. The file states no assumption for it.
dims: [scenario, snapshot, bus]
where: Bus_attached
expression: Bus_injection == 0
\[
-\mathit{NP}_{n,\xi,t} + \sum_{g \in \mathcal{G} \,:\, \mathrm{Generator\_bus}(g) = n} \mathrm{sgn}_{g} \cdot p_{g,\xi,t} - \left( \sum_{k \in \mathcal{K} \,:\, \mathrm{Line\_bus0}(k) = n} s_{k,\xi,t} \right) + \sum_{k \in \mathcal{K} \,:\, \mathrm{Line\_bus1}(k) = n} s_{k,\xi,t} - 0.5 \cdot \left( \sum_{k \in \mathcal{K} \,:\, \mathrm{Line\_bus0}(k) = n} \ell_{k,\xi,t} \right) - 0.5 \cdot \left( \sum_{k \in \mathcal{K} \,:\, \mathrm{Line\_bus1}(k) = n} \ell_{k,\xi,t} \right) - \left( \sum_{l \in \mathcal{L} \,:\, \mathrm{Link\_bus0}(l) = n} f_{l,\xi,t} \right) + \sum_{o \in \mathcal{O} \,:\, \mathrm{Link\_output\_bus}(o) = n} \overrightarrow{f}_{o,\xi,t} + \sum_{d \in \mathcal{D} \,:\, \mathrm{Load\_bus}(d) = n} \check{\mathrm{load}}_{d,\xi,t} + \sum_{r \in \mathcal{R} \,:\, \mathrm{Process\_output\_bus}(r) = n} \overrightarrow{z}_{r,\xi,t} + \sum_{s \in \mathcal{S} \,:\, \mathrm{StorageUnit\_bus}(s) = n} \mathrm{sgn}^{h}_{s} \cdot \left( h^{+}_{\xi,t,s} - h^{-}_{\xi,t,s} \right) + \sum_{v \in \mathcal{V} \,:\, \mathrm{Store\_bus}(v) = n} \mathrm{sgn}^{q}_{v} \cdot q_{\xi,t,v} - \left( \sum_{m \in \mathcal{M} \,:\, \mathrm{Transformer\_bus0}(m) = n} \sigma_{\xi,t,m} \right) + \sum_{m \in \mathcal{M} \,:\, \mathrm{Transformer\_bus1}(m) = n} \sigma_{\xi,t,m} - 0.5 \cdot \left( \sum_{m \in \mathcal{M} \,:\, \mathrm{Transformer\_bus0}(m) = n} \ell^{\sigma}_{\xi,t,m} \right) - 0.5 \cdot \left( \sum_{m \in \mathcal{M} \,:\, \mathrm{Transformer\_bus1}(m) = n} \ell^{\sigma}_{\xi,t,m} \right) = 0 \qquad \forall\, n \in \mathcal{N},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{Bus\_attached}_{n,t}
\]
CVaR-def#
CVaR_def
CVaR_def:
description: "`CVaR-def` — the tail's average is at least where it starts plus the expected excess over the tail's probability"
dims: []
where: CVaR_omega
expression: CVaR_theta + 1 / (1 - CVaR_alpha) * sum(scenario_weight * CVaR_a, over=scenario) <= CVaR
\[
\theta + \frac{1}{1 - \alpha} \cdot \left( \sum_{\xi \in \Xi} \pi_{\xi} \cdot a_{\xi} \right) \le CVaR \qquad \text{where } \omega \text{ is defined}
\]
CVaR-excess-{s}#
CVaR_excess
CVaR_excess:
description: "`CVaR-excess-{s}` — a scenario's operating cost beyond the tail's start is its excess. PyPSA names one row per scenario"
dims: [scenario]
where: CVaR_omega
expression: CVaR_a - scenario_opex + CVaR_theta >= 0
\[
a_{\xi} - \mathit{scenario\_opex}_{\xi} + \theta \ge 0 \qquad \forall\, \xi \in \Xi \,:\, \omega \text{ is defined}
\]
Carrier-growth_limit#
Carrier_growth_limit
Carrier_growth_limit:
description: >-
`Carrier-growth_limit` — what a carrier adds in a period is at most its
allowance plus a share of what it added the period before. It counts
what its extendable components add, each build in the first period it
is active in. The first period has no predecessor, so `edge=0` leaves it
the bare allowance
dims: [carrier, period]
where: Carrier_max_growth
expression: >-
Carrier_additions
- shift(Carrier_additions, along=period, offset=1, edge=0) * Carrier_relative_growth
<= Carrier_max_growth
\[
\mathit{Carrier\_additions}_{i,y} - \mathit{Carrier\_additions}_{i,y \boxminus_{0} 1} \cdot \mathrm{r}^{+}_{i} \le \overline{\Delta}_{i} \qquad \forall\, i \in \mathcal{I},\ y \in \mathcal{Y} \,:\, \overline{\Delta}_{i} \text{ is defined}
\]
FlowBased-balance#
FlowBased_balance
FlowBased_balance:
description: "`FlowBased-balance` — the net positions of each domain's zones sum to zero"
dims: [scenario, snapshot, flow_based_domain]
expression: sum(FlowBased_net_position, over=bus, by=Bus_flow_based_domain[flow_based_domain]) == 0
\[
\sum_{n \in \mathcal{N} \,:\, \mathrm{Bus\_flow\_based\_domain}(n) = l} \mathit{NP}_{n,\xi,t} = 0 \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator-com-down-time#
Generator_com_down_time
Generator_com_down_time:
description: >-
`Generator-com-down-time` — a unit stopped within its own minimum down time
is still off, and so are the modules of a modular unit that stopped.
The down time it brought into the horizon counts as one stop before the
first snapshot, also where the unit is modular and all its modules
stopped (`constraints.py:656-683`, PyPSA/PyPSA#2114)
dims: [scenario, snapshot, generator]
where: Generator_committed AND Generator_min_down_time > 0
expression: >-
sum_back(Generator_shut_down, along=snapshot, window=Generator_min_down_time)
+ Generator_down_time_brought_in <= Generator_modules - Generator_status
\[
\sum_{t' \in \mathcal{T} \,:\, 0 \le t - t' < \mathrm{DT}} \mathit{dn}_{g,\xi,t'} + \mathrm{rest}_{g,\xi,t} \le \mathrm{M}_{g,\xi} - u_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \mathrm{DT}_{g,\xi} > 0
\]
Generator-com-ext-p-lower#
Generator_com_ext_p_lower
Generator_com_ext_p_lower:
description: >-
`Generator-com-ext-p-lower` — a committed extendable unit outputs at
least its minimum of the chosen build. Off, the big M releases the row
dims: [scenario, snapshot, generator]
where: Generator_com_ext
expression: >-
Generator_p >=
Generator_p_min_pu * (Generator_p_nom_ext - Generator_maintenance_pu * Generator_maintenance_capacity)
+ Generator_big_m * Generator_status - Generator_big_m
\[
p_{g,\xi,t} \ge \underline{\mathrm{p}}_{g,\xi,t} \cdot \left( P_{g} - \gamma_{g,\xi} \cdot \mu^{\mathrm{nom}}_{g,\xi,t} \right) + \mathrm{M}_{g,\xi} \cdot u_{g,\xi,t} - \mathrm{M}_{g,\xi} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com,ext}}_{g,t}
\]
Generator-com-ext-p-lower-nonneg#
Generator_com_ext_p_lower_nonneg
Generator_com_ext_p_lower_nonneg:
description: >-
`Generator-com-ext-p-lower-nonneg` — where no minimum-per-unit is
negative, output is also plainly non-negative, a row the big-M lower
cannot assert while the unit is off
dims: [scenario, snapshot, generator]
where: Generator_com_ext AND Generator_p_min_pu_nonneg
expression: Generator_p >= 0
\[
p_{g,\xi,t} \ge 0 \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com,ext}}_{g,t} \wedge \mathrm{nonneg}_{g}
\]
Generator-com-ext-p-upper-bigM#
Generator_com_ext_p_upper_big_m
Generator_com_ext_p_upper_big_m:
description: "`Generator-com-ext-p-upper-bigM` — off, a unit outputs nothing. On, the big M is no bound"
dims: [scenario, snapshot, generator]
where: Generator_com_ext
expression: Generator_p <= Generator_big_m * Generator_status
\[
p_{g,\xi,t} \le \mathrm{M}_{g,\xi} \cdot u_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com,ext}}_{g,t}
\]
Generator-com-ext-p-upper-cap#
Generator_com_ext_p_upper_cap
Generator_com_ext_p_upper_cap:
description: >-
`Generator-com-ext-p-upper-cap` — a committed extendable unit outputs
at most what is available of the chosen build, whatever its status
dims: [scenario, snapshot, generator]
where: Generator_com_ext
expression: Generator_p <= Generator_p_max_pu * (Generator_p_nom_ext - Generator_maintenance_pu * Generator_maintenance_capacity)
\[
p_{g,\xi,t} \le \overline{\mathrm{p}}_{g,\xi,t} \cdot \left( P_{g} - \gamma_{g,\xi} \cdot \mu^{\mathrm{nom}}_{g,\xi,t} \right) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com,ext}}_{g,t}
\]
Generator-com-mod-p-lower#
Generator_com_mod_p_lower
Generator_com_mod_p_lower:
description: >-
`Generator-com-mod-p-lower` — a committed modular unit outputs at least
its minimum of one module, whether the build is fixed or a decision
dims: [scenario, snapshot, generator]
where: Generator_committed AND Generator_p_nom_mod > 0
expression: Generator_p >= Generator_p_min_pu * Generator_p_nom_mod * (Generator_status - Generator_maintenance_pu * Generator_maintenance_status)
\[
p_{g,\xi,t} \ge \underline{\mathrm{p}}_{g,\xi,t} \cdot \mathrm{p}^{\mathrm{mod}}_{g} \cdot \left( u_{g,\xi,t} - \gamma_{g,\xi} \cdot \mu^{u}_{g,\xi,t} \right) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0
\]
Generator-com-mod-p-upper#
Generator_com_mod_p_upper
Generator_com_mod_p_upper:
description: >-
`Generator-com-mod-p-upper` — a committed modular unit outputs at most
one module's share, whether the build is fixed or a decision
dims: [scenario, snapshot, generator]
where: Generator_committed AND Generator_p_nom_mod > 0
expression: Generator_p <= Generator_p_max_pu * Generator_p_nom_mod * (Generator_status - Generator_maintenance_pu * Generator_maintenance_status)
\[
p_{g,\xi,t} \le \overline{\mathrm{p}}_{g,\xi,t} \cdot \mathrm{p}^{\mathrm{mod}}_{g} \cdot \left( u_{g,\xi,t} - \gamma_{g,\xi} \cdot \mu^{u}_{g,\xi,t} \right) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0
\]
Generator-com-p-lower#
Generator_com_p_lower
Generator_com_p_lower:
description: "`Generator-com-p-lower` — a committed unit outputs at least its minimum. Off, it outputs at least nothing"
dims: [scenario, snapshot, generator]
where: Generator_committed AND NOT Generator_p_nom_extendable AND NOT (Generator_p_nom_mod > 0)
expression: Generator_p >= Generator_p_min_pu * Generator_p_nom * (Generator_status - Generator_maintenance_pu * Generator_maintenance_status)
\[
p_{g,\xi,t} \ge \underline{\mathrm{p}}_{g,\xi,t} \cdot \mathrm{p}^{\mathrm{nom}}_{g,\xi} \cdot \left( u_{g,\xi,t} - \gamma_{g,\xi} \cdot \mu^{u}_{g,\xi,t} \right) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \neg \mathrm{ext}_{g} \wedge \neg \left( \mathrm{p}^{\mathrm{mod}}_{g} > 0 \right)
\]
Generator-com-p-upper#
Generator_com_p_upper
Generator_com_p_upper:
description: "`Generator-com-p-upper` — a committed unit outputs at most what is available. Off, it outputs at most nothing"
dims: [scenario, snapshot, generator]
where: Generator_committed AND NOT Generator_p_nom_extendable AND NOT (Generator_p_nom_mod > 0)
expression: Generator_p <= Generator_p_max_pu * Generator_p_nom * (Generator_status - Generator_maintenance_pu * Generator_maintenance_status)
\[
p_{g,\xi,t} \le \overline{\mathrm{p}}_{g,\xi,t} \cdot \mathrm{p}^{\mathrm{nom}}_{g,\xi} \cdot \left( u_{g,\xi,t} - \gamma_{g,\xi} \cdot \mu^{u}_{g,\xi,t} \right) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \neg \mathrm{ext}_{g} \wedge \neg \left( \mathrm{p}^{\mathrm{mod}}_{g} > 0 \right)
\]
Generator-com-transition-shut-down#
Generator_com_transition_shut_down
Generator_com_transition_shut_down:
description: "`Generator-com-transition-shut-down` — turning off is a stop, counted against the state the unit carried into the snapshot"
dims: [scenario, snapshot, generator]
where: Generator_committed
expression: Generator_shut_down >= Generator_previous_status - Generator_status
\[
\mathit{dn}_{g,\xi,t} \ge \overleftarrow{u}_{g,\xi,t} - u_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t}
\]
Generator-com-transition-start-up#
Generator_com_transition_start_up
Generator_com_transition_start_up:
description: "`Generator-com-transition-start-up` — turning on is a start, counted against the state the unit carried into the snapshot"
dims: [scenario, snapshot, generator]
where: Generator_committed
expression: Generator_start_up >= Generator_status - Generator_previous_status
\[
\mathit{up}_{g,\xi,t} \ge u_{g,\xi,t} - \overleftarrow{u}_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t}
\]
Generator-com-up-time#
Generator_com_up_time
Generator_com_up_time:
description: >-
`Generator-com-up-time` — a unit started within its own minimum up time is
still on. The up time it brought into the horizon counts as a start
before the first snapshot
dims: [scenario, snapshot, generator]
where: Generator_committed AND Generator_min_up_time > 0
expression: >-
sum_back(Generator_start_up, along=snapshot, window=Generator_min_up_time)
+ Generator_up_time_brought_in <= Generator_status
\[
\sum_{t' \in \mathcal{T} \,:\, 0 \le t - t' < \mathrm{UT}} \mathit{up}_{g,\xi,t'} + \mathrm{hold}_{g,\xi,t} \le u_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \mathrm{UT}_{g,\xi} > 0
\]
Generator-e_sum_max#
Generator_e_sum_max
Generator_e_sum_max:
description: "`Generator-e_sum_max` — energy over the horizon, each period weighted by its years, is at most its budget. A budget of infinity is no row, and so is a generator active in no snapshot"
dims: [scenario, generator]
where: Generator_e_sum_max AND count(Generator_active, over=snapshot) > 0
expression: sum(Generator_p * energy_limit_weight, over=snapshot) <= Generator_e_sum_max
\[
\sum_{t \in \mathcal{T}} p_{g,\xi,t} \cdot \mathit{w}^{E}_{t} \le \overline{\mathrm{E}}_{g,\xi} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi \,:\, \overline{\mathrm{E}}_{g,\xi} \text{ is defined} \wedge \lvert \{ t \in \mathcal{T} \,:\, \mathrm{on}_{g,t} \} \rvert > 0
\]
Generator-e_sum_min#
Generator_e_sum_min
Generator_e_sum_min:
description: "`Generator-e_sum_min` — energy over the horizon, each period weighted by its years, is at least its floor. A floor of minus infinity is no row, and so is a generator active in no snapshot"
dims: [scenario, generator]
where: Generator_e_sum_min AND count(Generator_active, over=snapshot) > 0
expression: sum(Generator_p * energy_limit_weight, over=snapshot) >= Generator_e_sum_min
\[
\sum_{t \in \mathcal{T}} p_{g,\xi,t} \cdot \mathit{w}^{E}_{t} \ge \underline{\mathrm{E}}_{g,\xi} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi \,:\, \underline{\mathrm{E}}_{g,\xi} \text{ is defined} \wedge \lvert \{ t \in \mathcal{T} \,:\, \mathrm{on}_{g,t} \} \rvert > 0
\]
Generator-ext-p-lower#
Generator_ext_p_lower
Generator_ext_p_lower:
description: "`Generator-ext-p-lower` — an extendable generator outputs at least its minimum of the chosen build"
dims: [scenario, snapshot, generator]
where: Generator_p_nom_extendable AND not Generator_committable AND Generator_active
expression: Generator_p >= Generator_p_min_pu * (Generator_p_nom_ext - Generator_maintenance_pu * Generator_maintenance_capacity)
\[
p_{g,\xi,t} \ge \underline{\mathrm{p}}_{g,\xi,t} \cdot \left( P_{g} - \gamma_{g,\xi} \cdot \mu^{\mathrm{nom}}_{g,\xi,t} \right) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{ext}_{g} \wedge \neg \mathrm{com}_{g} \wedge \mathrm{on}_{g,t}
\]
Generator-ext-p_nom-lower#
Generator_ext_p_nom_lower
Generator_ext_p_nom_lower:
description: "`Generator-ext-p_nom-lower` — the chosen build is at least its floor in every scenario"
dims: [scenario, generator]
where: Generator_p_nom_extendable AND Generator_active_asset
expression: Generator_p_nom_ext >= Generator_p_nom_min
\[
P_{g} \ge \underline{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi \,:\, \mathrm{ext}_{g} \wedge \mathrm{act}_{g}
\]
Generator-ext-p_nom-upper#
Generator_ext_p_nom_upper
Generator_ext_p_nom_upper:
description: "`Generator-ext-p_nom-upper` — the chosen build is at most its cap in every scenario. A cap of infinity is no row"
dims: [scenario, generator]
where: Generator_p_nom_extendable AND Generator_active_asset AND Generator_p_nom_max
expression: Generator_p_nom_ext <= Generator_p_nom_max
\[
P_{g} \le \overline{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi \,:\, \mathrm{ext}_{g} \wedge \mathrm{act}_{g} \wedge \overline{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} \text{ is defined}
\]
Generator-ext-p-upper#
Generator_ext_p_upper
Generator_ext_p_upper:
description: "`Generator-ext-p-upper` — an extendable generator outputs at most what is available of the chosen build"
dims: [scenario, snapshot, generator]
where: Generator_p_nom_extendable AND not Generator_committable AND Generator_active
expression: Generator_p <= Generator_p_max_pu * (Generator_p_nom_ext - Generator_maintenance_pu * Generator_maintenance_capacity)
\[
p_{g,\xi,t} \le \overline{\mathrm{p}}_{g,\xi,t} \cdot \left( P_{g} - \gamma_{g,\xi} \cdot \mu^{\mathrm{nom}}_{g,\xi,t} \right) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{ext}_{g} \wedge \neg \mathrm{com}_{g} \wedge \mathrm{on}_{g,t}
\]
Generator-fix-p-lower#
Generator_fix_p_lower
Generator_fix_p_lower:
description: "`Generator-fix-p-lower` — a fixed generator outputs at least its minimum"
dims: [scenario, snapshot, generator]
where: not Generator_p_nom_extendable AND not Generator_committable AND Generator_active
expression: Generator_p >= Generator_p_min_pu * Generator_p_nom * (1 - Generator_maintenance_pu * Generator_maintenance)
\[
p_{g,\xi,t} \ge \underline{\mathrm{p}}_{g,\xi,t} \cdot \mathrm{p}^{\mathrm{nom}}_{g,\xi} \cdot \left( 1 - \gamma_{g,\xi} \cdot \mu_{g,\xi,t} \right) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \neg \mathrm{ext}_{g} \wedge \neg \mathrm{com}_{g} \wedge \mathrm{on}_{g,t}
\]
Generator-fix-p-upper#
Generator_fix_p_upper
Generator_fix_p_upper:
description: "`Generator-fix-p-upper` — a fixed generator outputs at most what is available"
dims: [scenario, snapshot, generator]
where: not Generator_p_nom_extendable AND not Generator_committable AND Generator_active
expression: Generator_p <= Generator_p_max_pu * Generator_p_nom * (1 - Generator_maintenance_pu * Generator_maintenance)
\[
p_{g,\xi,t} \le \overline{\mathrm{p}}_{g,\xi,t} \cdot \mathrm{p}^{\mathrm{nom}}_{g,\xi} \cdot \left( 1 - \gamma_{g,\xi} \cdot \mu_{g,\xi,t} \right) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \neg \mathrm{ext}_{g} \wedge \neg \mathrm{com}_{g} \wedge \mathrm{on}_{g,t}
\]
Generator-maint-event-count#
Generator_maint_event_count
Generator_maint_event_count:
description: "`Generator-maint-event-count` — a maintainable generator holds its number of maintenance events over the horizon"
dims: [scenario, generator]
where: Generator_maintainable
expression: sum(Generator_maintenance_start, over=snapshot) == Generator_maintenance_events
\[
\sum_{t \in \mathcal{T}} \mu^{\mathrm{up}}_{g,\xi,t} = \mathrm{n}^{\mathrm{mnt}}_{g,\xi} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi \,:\, \mathrm{mnt}_{g}
\]
Generator-maint-modstatus-lb#
Generator_maint_modstatus_lb
Generator_maint_modstatus_lb:
description: "`Generator-maint-modstatus-lb` — in maintenance, every module on is on in maintenance"
dims: [scenario, snapshot, generator]
where: Generator_maintainable AND Generator_committed AND Generator_p_nom_mod > 0
expression: Generator_maintenance_status >= Generator_status - Generator_maintenance_modules * (1 - Generator_maintenance)
\[
\mu^{u}_{g,\xi,t} \ge u_{g,\xi,t} - \mathrm{Generator\_maintenance\_modules}_{g,\xi} \cdot \left( 1 - \mu_{g,\xi,t} \right) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}_{g} \wedge \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0
\]
Generator-maint-modstatus-le-maint#
Generator_maint_modstatus_le_maint
Generator_maint_modstatus_le_maint:
description: >-
`Generator-maint-modstatus-le-maint` — out of maintenance, no module is
on in maintenance. In it, at most its modules in maintenance
dims: [scenario, snapshot, generator]
where: Generator_maintainable AND Generator_committed AND Generator_p_nom_mod > 0
expression: Generator_maintenance_status <= Generator_maintenance_modules * Generator_maintenance
\[
\mu^{u}_{g,\xi,t} \le \mathrm{Generator\_maintenance\_modules}_{g,\xi} \cdot \mu_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}_{g} \wedge \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0
\]
Generator-maint-modstatus-le-status#
Generator_maint_modstatus_le_status
Generator_maint_modstatus_le_status:
description: "`Generator-maint-modstatus-le-status` — the modules on in maintenance are at most the modules on"
dims: [scenario, snapshot, generator]
where: Generator_maintainable AND Generator_committed AND Generator_p_nom_mod > 0
expression: Generator_maintenance_status <= Generator_status
\[
\mu^{u}_{g,\xi,t} \le u_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}_{g} \wedge \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0
\]
Generator-maint-start-horizon#
Generator_maint_start_horizon
Generator_maint_start_horizon:
description: "`Generator-maint-start-horizon` — no event starts where it could not run its whole duration"
dims: [scenario, snapshot, generator]
where: Generator_maintainable AND Generator_active AND Generator_maintenance_start_blocked
expression: Generator_maintenance_start == 0
\[
\mu^{\mathrm{up}}_{g,\xi,t} = 0 \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}_{g} \wedge \mathrm{on}_{g,t} \wedge \mathrm{blk}_{g,\xi,t}
\]
Generator-maint-status-lb#
Generator_maint_status_lb
Generator_maint_status_lb:
description: "`Generator-maint-status-lb` — on and in maintenance, the status in maintenance is one"
dims: [scenario, snapshot, generator]
where: Generator_maintainable AND Generator_committed AND NOT Generator_p_nom_extendable AND NOT (Generator_p_nom_mod > 0)
expression: Generator_maintenance_status >= Generator_status + Generator_maintenance - 1
\[
\mu^{u}_{g,\xi,t} \ge u_{g,\xi,t} + \mu_{g,\xi,t} - 1 \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}_{g} \wedge \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \neg \mathrm{ext}_{g} \wedge \neg \left( \mathrm{p}^{\mathrm{mod}}_{g} > 0 \right)
\]
Generator-maint-status-le-maint#
Generator_maint_status_le_maint
Generator_maint_status_le_maint:
description: "`Generator-maint-status-le-maint` — out of maintenance, the status in maintenance is zero"
dims: [scenario, snapshot, generator]
where: Generator_maintainable AND Generator_committed AND NOT Generator_p_nom_extendable AND NOT (Generator_p_nom_mod > 0)
expression: Generator_maintenance_status <= Generator_maintenance
\[
\mu^{u}_{g,\xi,t} \le \mu_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}_{g} \wedge \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \neg \mathrm{ext}_{g} \wedge \neg \left( \mathrm{p}^{\mathrm{mod}}_{g} > 0 \right)
\]
Generator-maint-status-le-status#
Generator_maint_status_le_status
Generator_maint_status_le_status:
description: "`Generator-maint-status-le-status` — the status in maintenance is at most the status"
dims: [scenario, snapshot, generator]
where: Generator_maintainable AND Generator_committed AND NOT Generator_p_nom_extendable AND NOT (Generator_p_nom_mod > 0)
expression: Generator_maintenance_status <= Generator_status
\[
\mu^{u}_{g,\xi,t} \le u_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}_{g} \wedge \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \neg \mathrm{ext}_{g} \wedge \neg \left( \mathrm{p}^{\mathrm{mod}}_{g} > 0 \right)
\]
Generator-maint-window#
Generator_maint_window
Generator_maint_window:
description: >-
`Generator-maint-window` — a generator is in maintenance exactly where
an event it started covers the snapshot. Two events do not overlap,
since the maintenance status is at most one
dims: [scenario, snapshot, generator]
where: Generator_maintainable AND Generator_active
expression: Generator_maintenance == sum(Generator_maintenance_start, over=start, by=Generator_maintenance_cover[covered])
\[
\mu_{g,\xi,t} = \sum_{t' \in \mathcal{T} \,:\, \left( t,\ g,\ \xi,\ t' \right) \in \mathrm{Generator\_maintenance\_cover}} \mu^{\mathrm{up}}_{g,\xi,t'} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}_{g} \wedge \mathrm{on}_{g,t}
\]
Generator-maintcap_lower_nommax#
Generator_maintcap_lower_nommax
Generator_maintcap_lower_nommax:
description: "`Generator-maintcap_lower_nommax` — in maintenance, the whole chosen build is taken off"
dims: [scenario, snapshot, generator]
where: Generator_maint_ext
expression: Generator_maintenance_capacity >= Generator_p_nom_ext - Generator_p_nom_max * (1 - Generator_maintenance)
\[
\mu^{\mathrm{nom}}_{g,\xi,t} \ge P_{g} - \overline{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} \cdot \left( 1 - \mu_{g,\xi,t} \right) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{mnt,ext}}_{g,t}
\]
Generator-maintcap_lower_nommin#
Generator_maintcap_lower_nommin
Generator_maintcap_lower_nommin:
description: "`Generator-maintcap_lower_nommin` — in maintenance, at least the floor of the build is taken off"
dims: [scenario, snapshot, generator]
where: Generator_maint_ext AND Generator_p_nom_min > 0
expression: Generator_maintenance_capacity >= Generator_p_nom_min * Generator_maintenance
\[
\mu^{\mathrm{nom}}_{g,\xi,t} \ge \underline{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} \cdot \mu_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{mnt,ext}}_{g,t} \wedge \underline{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} > 0
\]
Generator-maintcap_upper#
Generator_maintcap_upper
Generator_maintcap_upper:
description: >-
`Generator-maintcap_upper` — the build taken off is at most the chosen build in
maintenance, and at most the build less its floor out of it
dims: [scenario, snapshot, generator]
where: Generator_maint_ext
expression: Generator_maintenance_capacity <= Generator_p_nom_ext - Generator_p_nom_min * (1 - Generator_maintenance)
\[
\mu^{\mathrm{nom}}_{g,\xi,t} \le P_{g} - \underline{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} \cdot \left( 1 - \mu_{g,\xi,t} \right) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{mnt,ext}}_{g,t}
\]
Generator-maintcap_upper_nommax#
Generator_maintcap_upper_nommax
Generator_maintcap_upper_nommax:
description: "`Generator-maintcap_upper_nommax` — out of maintenance, no build is taken off"
dims: [scenario, snapshot, generator]
where: Generator_maint_ext
expression: Generator_maintenance_capacity <= Generator_p_nom_max * Generator_maintenance
\[
\mu^{\mathrm{nom}}_{g,\xi,t} \le \overline{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} \cdot \mu_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{mnt,ext}}_{g,t}
\]
Generator-p_nom_modularity#
Generator_p_nom_modularity
Generator_p_nom_modularity:
description: "`Generator-p_nom_modularity` — the chosen build is a whole number of modules"
dims: [generator]
where: Generator_p_nom_extendable AND Generator_p_nom_mod > 0 AND Generator_active_asset
expression: Generator_p_nom_ext == Generator_p_nom_mod * Generator_n_mod
\[
P_{g} = \mathrm{p}^{\mathrm{mod}}_{g} \cdot N_{g} \qquad \forall\, g \in \mathcal{G} \,:\, \mathrm{ext}_{g} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0 \wedge \mathrm{act}_{g}
\]
Generator-p_nom_set#
Generator_p_nom_set
Generator_p_nom_set:
description: "`Generator-p_nom_set` — the chosen build pinned, wherever a value is given"
dims: [generator]
where: Generator_p_nom_extendable AND Generator_active_asset AND Generator_p_nom_set
expression: Generator_p_nom_ext == Generator_p_nom_set
\[
P_{g} = \mathrm{p}^{\mathrm{nom,set}}_{g} \qquad \forall\, g \in \mathcal{G} \,:\, \mathrm{ext}_{g} \wedge \mathrm{act}_{g} \wedge \mathrm{p}^{\mathrm{nom,set}}_{g} \text{ is defined}
\]
Generator-p-ramp_limit_down#
Generator_p_ramp_limit_down
Generator_p_ramp_limit_down:
description: >-
`Generator-p-ramp_limit_down` — a generator lowers output no faster than
its ramp limit of the build. A committed one lowers it no further than
its shut-down ramp in the snapshot it turns off. A unit that came into
the horizon running has a row at the first snapshot only where its
`p_init` gives the output it brought in. No unit has one at the start of
a later investment period, nor does any unit a big M releases instead
dims: [scenario, snapshot, generator]
where: >-
(Generator_ramp_limit_down OR Generator_ramp_limit_shut_down)
AND NOT (Generator_committable AND Generator_p_nom_extendable AND NOT (Generator_p_nom_mod > 0))
AND Generator_ramps_from_previous
AND Generator_active
expression: Generator_previous_p - Generator_p <= Generator_ramp_down_allowance
\[
\overleftarrow{p}_{g,\xi,t} - p_{g,\xi,t} \le \Delta^{-}_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \left( \mathrm{rd}_{g,\xi,t} \text{ is defined} \vee \mathrm{rd}^{\mathrm{dn}}_{g,\xi} \text{ is defined} \right) \wedge \neg \left( \mathrm{com}_{g} \wedge \mathrm{ext}_{g} \wedge \neg \left( \mathrm{p}^{\mathrm{mod}}_{g} > 0 \right) \right) \wedge \mathrm{prev}_{g,\xi,t} \wedge \mathrm{on}_{g,t}
\]
Generator-p-ramp_limit_down-run-bigM#
Generator_p_ramp_limit_down_run_big_m
Generator_p_ramp_limit_down_run_big_m:
description: >-
`Generator-p-ramp_limit_down-run-bigM` — a committed extendable unit
lowers output no faster than its limit of the chosen build. The big M
releases the row in the snapshot it turns off
dims: [scenario, snapshot, generator]
where: >-
Generator_com_ext
AND (Generator_ramp_limit_down OR Generator_ramp_limit_shut_down)
AND Generator_ramps_from_previous
expression: >-
Generator_p - Generator_previous_p >=
-Generator_ramp_down_rate * Generator_p_nom_ext
- Generator_big_m + Generator_big_m * Generator_status
\[
p_{g,\xi,t} - \overleftarrow{p}_{g,\xi,t} \ge -\widetilde{\mathrm{rd}}_{g,\xi,t} \cdot P_{g} - \mathrm{M}_{g,\xi} + \mathrm{M}_{g,\xi} \cdot u_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com,ext}}_{g,t} \wedge \left( \mathrm{rd}_{g,\xi,t} \text{ is defined} \vee \mathrm{rd}^{\mathrm{dn}}_{g,\xi} \text{ is defined} \right) \wedge \mathrm{prev}_{g,\xi,t}
\]
Generator-p-ramp_limit_down-shut-bigM#
Generator_p_ramp_limit_down_shut_big_m
Generator_p_ramp_limit_down_shut_big_m:
description: >-
`Generator-p-ramp_limit_down-shut-bigM` — in the snapshot it turns off,
a committed extendable unit ramps no further than its shut-down ramp of
the chosen build. The big M releases the row everywhere else
dims: [scenario, snapshot, generator]
where: >-
Generator_com_ext
AND (Generator_ramp_limit_down OR Generator_ramp_limit_shut_down)
AND Generator_ramps_from_previous
expression: >-
Generator_previous_p - Generator_p <=
Generator_shut_down_rate * Generator_p_nom_ext
+ Generator_big_m - Generator_big_m * Generator_shut_down
\[
\overleftarrow{p}_{g,\xi,t} - p_{g,\xi,t} \le \widetilde{\mathrm{rd}}^{\mathrm{dn}}_{g,\xi} \cdot P_{g} + \mathrm{M}_{g,\xi} - \mathrm{M}_{g,\xi} \cdot \mathit{dn}_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com,ext}}_{g,t} \wedge \left( \mathrm{rd}_{g,\xi,t} \text{ is defined} \vee \mathrm{rd}^{\mathrm{dn}}_{g,\xi} \text{ is defined} \right) \wedge \mathrm{prev}_{g,\xi,t}
\]
Generator-p-ramp_limit_up#
Generator_p_ramp_limit_up
Generator_p_ramp_limit_up:
description: >-
`Generator-p-ramp_limit_up` — a generator raises output no faster than
its ramp limit of the build. A committed one raises it no further than
its start-up ramp in the snapshot it turns on. A unit that came into the
horizon running has a row at the first snapshot only where its `p_init`
gives the output it brought in. No unit has one at the start of a later
investment period, nor does any unit a big M releases instead
dims: [scenario, snapshot, generator]
where: >-
(Generator_ramp_limit_up OR Generator_ramp_limit_start_up)
AND NOT (Generator_committable AND Generator_p_nom_extendable AND NOT (Generator_p_nom_mod > 0))
AND Generator_ramps_from_previous
AND Generator_active
expression: Generator_p - Generator_previous_p <= Generator_ramp_up_allowance
\[
p_{g,\xi,t} - \overleftarrow{p}_{g,\xi,t} \le \Delta^{+}_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \left( \mathrm{ru}_{g,\xi,t} \text{ is defined} \vee \mathrm{ru}^{\mathrm{up}}_{g,\xi} \text{ is defined} \right) \wedge \neg \left( \mathrm{com}_{g} \wedge \mathrm{ext}_{g} \wedge \neg \left( \mathrm{p}^{\mathrm{mod}}_{g} > 0 \right) \right) \wedge \mathrm{prev}_{g,\xi,t} \wedge \mathrm{on}_{g,t}
\]
Generator-p-ramp_limit_up-run-bigM#
Generator_p_ramp_limit_up_run_big_m
Generator_p_ramp_limit_up_run_big_m:
description: >-
`Generator-p-ramp_limit_up-run-bigM` — a committed extendable unit
raises output no faster than its limit of the chosen build. The big M
releases the row in the snapshot it turns on
dims: [scenario, snapshot, generator]
where: >-
Generator_com_ext
AND (Generator_ramp_limit_up OR Generator_ramp_limit_start_up)
AND Generator_ramps_from_previous
expression: >-
Generator_p - Generator_previous_p <=
Generator_ramp_up_rate * Generator_p_nom_ext
+ Generator_big_m - Generator_big_m * Generator_previous_status
\[
p_{g,\xi,t} - \overleftarrow{p}_{g,\xi,t} \le \widetilde{\mathrm{ru}}_{g,\xi,t} \cdot P_{g} + \mathrm{M}_{g,\xi} - \mathrm{M}_{g,\xi} \cdot \overleftarrow{u}_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com,ext}}_{g,t} \wedge \left( \mathrm{ru}_{g,\xi,t} \text{ is defined} \vee \mathrm{ru}^{\mathrm{up}}_{g,\xi} \text{ is defined} \right) \wedge \mathrm{prev}_{g,\xi,t}
\]
Generator-p-ramp_limit_up-start-bigM#
Generator_p_ramp_limit_up_start_big_m
Generator_p_ramp_limit_up_start_big_m:
description: >-
`Generator-p-ramp_limit_up-start-bigM` — in the snapshot it turns on, a
committed extendable unit ramps no further than its start-up ramp of the
chosen build. The big M releases the row everywhere else
dims: [scenario, snapshot, generator]
where: >-
Generator_com_ext
AND (Generator_ramp_limit_up OR Generator_ramp_limit_start_up)
AND Generator_ramps_from_previous
expression: >-
Generator_p - Generator_previous_p <=
Generator_start_up_rate * Generator_p_nom_ext
+ Generator_big_m - Generator_big_m * Generator_start_up
\[
p_{g,\xi,t} - \overleftarrow{p}_{g,\xi,t} \le \widetilde{\mathrm{ru}}^{\mathrm{up}}_{g,\xi} \cdot P_{g} + \mathrm{M}_{g,\xi} - \mathrm{M}_{g,\xi} \cdot \mathit{up}_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com,ext}}_{g,t} \wedge \left( \mathrm{ru}_{g,\xi,t} \text{ is defined} \vee \mathrm{ru}^{\mathrm{up}}_{g,\xi} \text{ is defined} \right) \wedge \mathrm{prev}_{g,\xi,t}
\]
Generator-p_set#
Generator_p_set
Generator_p_set:
description: "`Generator-p_set` — output pinned to the given schedule, wherever one is given"
dims: [scenario, snapshot, generator]
where: Generator_p_set AND Generator_active
expression: Generator_p == Generator_p_set
\[
p_{g,\xi,t} = \mathrm{p}^{\mathrm{set}}_{g,\xi,t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{p}^{\mathrm{set}}_{g,\xi,t} \text{ is defined} \wedge \mathrm{on}_{g,t}
\]
Generator-shut_down-p-fixed-upper#
Generator_shut_down_p_fixed_upper
Generator_shut_down_p_fixed_upper:
description: >-
`Generator-shut_down-p-fixed-upper` — a stop is at most the modules in
place, an explicit row as PyPSA writes it. The bound is one where the
build is not modular, and the fixed build's whole count of modules where
it is
dims: [scenario, snapshot, generator]
where: Generator_committed AND NOT (Generator_p_nom_extendable AND Generator_p_nom_mod > 0)
expression: Generator_shut_down <= Generator_modules_installed
\[
\mathit{dn}_{g,\xi,t} \le \mathrm{N}^{\mathrm{fix}}_{g,\xi} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \neg \left( \mathrm{ext}_{g} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0 \right)
\]
Generator-shut_down-p_nom-variable-upper#
Generator_shut_down_p_nom_variable_upper
Generator_shut_down_p_nom_variable_upper:
description: "`Generator-shut_down-p_nom-variable-upper` — a modular unit stops only where a module is built"
dims: [scenario, snapshot, generator]
where: Generator_committed AND Generator_p_nom_extendable AND Generator_p_nom_mod > 0
expression: Generator_shut_down <= Generator_n_mod
\[
\mathit{dn}_{g,\xi,t} \le N_{g} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \mathrm{ext}_{g} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0
\]
Generator-start_up-p-fixed-upper#
Generator_start_up_p_fixed_upper
Generator_start_up_p_fixed_upper:
description: >-
`Generator-start_up-p-fixed-upper` — a start is at most the modules in
place, an explicit row as PyPSA writes it. The bound is one where the
build is not modular, and the fixed build's whole count of modules where
it is
dims: [scenario, snapshot, generator]
where: Generator_committed AND NOT (Generator_p_nom_extendable AND Generator_p_nom_mod > 0)
expression: Generator_start_up <= Generator_modules_installed
\[
\mathit{up}_{g,\xi,t} \le \mathrm{N}^{\mathrm{fix}}_{g,\xi} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \neg \left( \mathrm{ext}_{g} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0 \right)
\]
Generator-start_up-p_nom-variable-upper#
Generator_start_up_p_nom_variable_upper
Generator_start_up_p_nom_variable_upper:
description: "`Generator-start_up-p_nom-variable-upper` — a modular unit starts only where a module is built"
dims: [scenario, snapshot, generator]
where: Generator_committed AND Generator_p_nom_extendable AND Generator_p_nom_mod > 0
expression: Generator_start_up <= Generator_n_mod
\[
\mathit{up}_{g,\xi,t} \le N_{g} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \mathrm{ext}_{g} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0
\]
Generator-status-p-fixed-upper#
Generator_status_p_fixed_upper
Generator_status_p_fixed_upper:
description: >-
`Generator-status-p-fixed-upper` — a status is at most the modules in
place, an explicit row as PyPSA writes it. The bound is one where the
build is not modular, and the fixed build's whole count of modules where
it is
dims: [scenario, snapshot, generator]
where: Generator_committed AND NOT (Generator_p_nom_extendable AND Generator_p_nom_mod > 0)
expression: Generator_status <= Generator_modules_installed
\[
u_{g,\xi,t} \le \mathrm{N}^{\mathrm{fix}}_{g,\xi} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \neg \left( \mathrm{ext}_{g} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0 \right)
\]
Generator-status-p_nom-variable-upper#
Generator_status_p_nom_variable_upper
Generator_status_p_nom_variable_upper:
description: "`Generator-status-p_nom-variable-upper` — a modular unit is on only where a module is built"
dims: [scenario, snapshot, generator]
where: Generator_committed AND Generator_p_nom_extendable AND Generator_p_nom_mod > 0
expression: Generator_status <= Generator_n_mod
\[
u_{g,\xi,t} \le N_{g} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com}}_{g,t} \wedge \mathrm{ext}_{g} \wedge \mathrm{p}^{\mathrm{mod}}_{g} > 0
\]
flow_based#
GlobalConstraint_flow_based_ub
GlobalConstraint_flow_based_ub:
description: >-
`flow_based` — a CNEC's flow, the zones' net positions and the corridor
links' flows each times its factor, is at most its remaining available
margin. PyPSA names the rows `GlobalConstraint-ram`, one per CNEC and
snapshot, and reads no `constant` (`flow_based.py:262-293`)
dims: [scenario, snapshot, global_constraint]
where: >-
GlobalConstraint_type == 'flow_based' AND GlobalConstraint_sense == '<='
AND GlobalConstraint_ram AND GlobalConstraint_counts_snapshot
expression: >-
sum(GlobalConstraint_zonal_ptdf * FlowBased_net_position, over=bus)
+ sum(GlobalConstraint_link_ptdf * Link_p, over=link)
<= GlobalConstraint_ram
\[
\sum_{n \in \mathcal{N}} \mathrm{PTDF}^{z}_{n,o,\xi,t} \cdot \mathit{NP}_{n,\xi,t} + \sum_{l \in \mathcal{L}} \mathrm{PTDF}^{f}_{o,l,\xi,t} \cdot f_{l,\xi,t} \le \mathrm{RAM}_{o,\xi,t} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{type}_{o} = \text{'}\mathrm{flow\_based}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{<=}\text{'} \wedge \mathrm{RAM}_{o,\xi,t} \text{ is defined} \wedge \mathrm{in}_{o,\xi,t}
\]
operational_limit#
GlobalConstraint_operational_limit_eq
GlobalConstraint_operational_limit_eq:
description: "`operational_limit` — its total, at its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'operational_limit' AND GlobalConstraint_sense == '=='
expression: operational_limit == GlobalConstraint_constant
\[
\mathit{operational\_limit}_{o,\xi} = \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{operational\_limit}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{==}\text{'}
\]
operational_limit#
GlobalConstraint_operational_limit_lb
GlobalConstraint_operational_limit_lb:
description: "`operational_limit` — its total, at least its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'operational_limit' AND GlobalConstraint_sense == '>='
expression: operational_limit >= GlobalConstraint_constant
\[
\mathit{operational\_limit}_{o,\xi} \ge \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{operational\_limit}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{>=}\text{'}
\]
operational_limit#
GlobalConstraint_operational_limit_ub
GlobalConstraint_operational_limit_ub:
description: "`operational_limit` — its total, at most its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'operational_limit' AND GlobalConstraint_sense == '<='
expression: operational_limit <= GlobalConstraint_constant
\[
\mathit{operational\_limit}_{o,\xi} \le \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{operational\_limit}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{<=}\text{'}
\]
primary_energy#
GlobalConstraint_primary_energy_eq
GlobalConstraint_primary_energy_eq:
description: "`primary_energy` — its total, at its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'primary_energy' AND GlobalConstraint_sense == '=='
expression: primary_energy == GlobalConstraint_constant
\[
\mathit{primary\_energy}_{o,\xi} = \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{primary\_energy}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{==}\text{'}
\]
primary_energy#
GlobalConstraint_primary_energy_lb
GlobalConstraint_primary_energy_lb:
description: "`primary_energy` — its total, at least its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'primary_energy' AND GlobalConstraint_sense == '>='
expression: primary_energy >= GlobalConstraint_constant
\[
\mathit{primary\_energy}_{o,\xi} \ge \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{primary\_energy}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{>=}\text{'}
\]
primary_energy#
GlobalConstraint_primary_energy_ub
GlobalConstraint_primary_energy_ub:
description: "`primary_energy` — its total, at most its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'primary_energy' AND GlobalConstraint_sense == '<='
expression: primary_energy <= GlobalConstraint_constant
\[
\mathit{primary\_energy}_{o,\xi} \le \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{primary\_energy}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{<=}\text{'}
\]
tech_capacity_expansion_limit#
GlobalConstraint_tech_capacity_expansion_limit_eq
GlobalConstraint_tech_capacity_expansion_limit_eq:
description: "`tech_capacity_expansion_limit` — its total, at its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'tech_capacity_expansion_limit' AND GlobalConstraint_sense == '=='
expression: tech_capacity_expansion == GlobalConstraint_constant
\[
\mathit{tech\_capacity\_expansion}_{o} = \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{tech\_capacity\_expansion\_limit}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{==}\text{'}
\]
tech_capacity_expansion_limit#
GlobalConstraint_tech_capacity_expansion_limit_lb
GlobalConstraint_tech_capacity_expansion_limit_lb:
description: "`tech_capacity_expansion_limit` — its total, at least its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'tech_capacity_expansion_limit' AND GlobalConstraint_sense == '>='
expression: tech_capacity_expansion >= GlobalConstraint_constant
\[
\mathit{tech\_capacity\_expansion}_{o} \ge \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{tech\_capacity\_expansion\_limit}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{>=}\text{'}
\]
tech_capacity_expansion_limit#
GlobalConstraint_tech_capacity_expansion_limit_ub
GlobalConstraint_tech_capacity_expansion_limit_ub:
description: "`tech_capacity_expansion_limit` — its total, at most its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'tech_capacity_expansion_limit' AND GlobalConstraint_sense == '<='
expression: tech_capacity_expansion <= GlobalConstraint_constant
\[
\mathit{tech\_capacity\_expansion}_{o} \le \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{tech\_capacity\_expansion\_limit}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{<=}\text{'}
\]
transmission_expansion_cost_limit#
GlobalConstraint_transmission_expansion_cost_limit_eq
GlobalConstraint_transmission_expansion_cost_limit_eq:
description: "`transmission_expansion_cost_limit` — its total, at its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'transmission_expansion_cost_limit' AND GlobalConstraint_sense == '=='
expression: transmission_expansion_cost == GlobalConstraint_constant
\[
\mathit{transmission\_expansion\_cost}_{o,\xi} = \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{transmission\_expansion\_cost\_limit}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{==}\text{'}
\]
transmission_expansion_cost_limit#
GlobalConstraint_transmission_expansion_cost_limit_lb
GlobalConstraint_transmission_expansion_cost_limit_lb:
description: "`transmission_expansion_cost_limit` — its total, at least its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'transmission_expansion_cost_limit' AND GlobalConstraint_sense == '>='
expression: transmission_expansion_cost >= GlobalConstraint_constant
\[
\mathit{transmission\_expansion\_cost}_{o,\xi} \ge \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{transmission\_expansion\_cost\_limit}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{>=}\text{'}
\]
transmission_expansion_cost_limit#
GlobalConstraint_transmission_expansion_cost_limit_ub
GlobalConstraint_transmission_expansion_cost_limit_ub:
description: "`transmission_expansion_cost_limit` — its total, at most its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'transmission_expansion_cost_limit' AND GlobalConstraint_sense == '<='
expression: transmission_expansion_cost <= GlobalConstraint_constant
\[
\mathit{transmission\_expansion\_cost}_{o,\xi} \le \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{transmission\_expansion\_cost\_limit}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{<=}\text{'}
\]
transmission_volume_expansion_limit#
GlobalConstraint_transmission_volume_expansion_limit_eq
GlobalConstraint_transmission_volume_expansion_limit_eq:
description: "`transmission_volume_expansion_limit` — its total, at its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'transmission_volume_expansion_limit' AND GlobalConstraint_sense == '=='
expression: transmission_volume_expansion == GlobalConstraint_constant
\[
\mathit{transmission\_volume\_expansion}_{o,\xi} = \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{transmission\_volume\_expansion\_limit}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{==}\text{'}
\]
transmission_volume_expansion_limit#
GlobalConstraint_transmission_volume_expansion_limit_lb
GlobalConstraint_transmission_volume_expansion_limit_lb:
description: "`transmission_volume_expansion_limit` — its total, at least its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'transmission_volume_expansion_limit' AND GlobalConstraint_sense == '>='
expression: transmission_volume_expansion >= GlobalConstraint_constant
\[
\mathit{transmission\_volume\_expansion}_{o,\xi} \ge \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{transmission\_volume\_expansion\_limit}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{>=}\text{'}
\]
transmission_volume_expansion_limit#
GlobalConstraint_transmission_volume_expansion_limit_ub
GlobalConstraint_transmission_volume_expansion_limit_ub:
description: "`transmission_volume_expansion_limit` — its total, at most its constant"
dims: [scenario, global_constraint]
where: GlobalConstraint_type == 'transmission_volume_expansion_limit' AND GlobalConstraint_sense == '<='
expression: transmission_volume_expansion <= GlobalConstraint_constant
\[
\mathit{transmission\_volume\_expansion}_{o,\xi} \le \mathrm{K}_{o,\xi} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{transmission\_volume\_expansion\_limit}\text{'} \wedge \mathrm{sense}_{o,\xi} = \text{'}\mathrm{<=}\text{'}
\]
Kirchhoff-Voltage-Law#
Kirchhoff_Voltage_Law
Kirchhoff_Voltage_Law:
description: >-
`Kirchhoff-Voltage-Law` — around every independent cycle the
impedance-weighted flows sum to nothing. This makes the linear power
flow physical rather than transport. A transformer's flow weighs its
effective reactance, and its phase shift enters the cycle sum too. The
shift is a constant where it is fixed. Where a phase-shifting
transformer chooses its shift, it is the shift decision times its cycle
weight. A snapshot uses the cycles of its own period, of the branches
that are active in it (`constraints.py:1710-1729`). A cycle label that
period's basis does not reach has no row
dims: [scenario, snapshot, cycle]
where: at(Cycle_in_basis, by=snapshot_period[period])
expression: Cycle_angle_sum == 0
\[
\sum_{k \in \mathcal{K}} s_{k,\xi,t} \cdot \mathrm{x}_{c,k,\mathrm{snapshot\_period}(t)} + \sum_{m \in \mathcal{M}} \sigma_{\xi,t,m} \cdot \mathrm{x}^{\sigma}_{c,\mathrm{snapshot\_period}(t),m} + \sum_{m \in \mathcal{M}} \vartheta_{c,\xi,t,m} + \sum_{m \in \mathcal{M}} \mathit{Transformer\_phase\_shift}_{\xi,t,m} \cdot \mathrm{Transformer\_phase\_shift\_cycle\_weight}_{c,\mathrm{snapshot\_period}(t),\xi,m} = 0 \qquad \forall\, c \in \mathcal{C},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{Cycle\_in\_basis}_{c,\mathrm{snapshot\_period}(t)}
\]
Line-ext-s-lower#
Line_ext_s_lower
Line_ext_s_lower:
description: "`Line-ext-s-lower` — an extendable line carries at least the negative of its rating of the chosen build, the loss counted against it"
dims: [scenario, snapshot, line]
where: Line_s_nom_extendable AND Line_active
expression: Line_s - Line_loss >= -Line_s_max_pu * Line_s_nom_ext
\[
s_{k,\xi,t} - \ell_{k,\xi,t} \ge -\overline{\mathrm{s}}_{k,\xi,t} \cdot S_{k} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{ext}^{s}_{k} \wedge \mathrm{on}^{s}_{k,t}
\]
Line-ext-s-lower-security-for-{c}-outage-in-sub-network-{n}#
Line_ext_s_lower_security
Line_ext_s_lower_security:
description: >-
`Line-ext-s-lower-security-for-{c}-outage-in-sub-network-{n}` — after
any one outage, an extendable line carries at least the negative of its
rating of the chosen build. Its flow takes on its share of the outaged
branch's flow. PyPSA names one row per outaged component `c` and
sub-network `n`. Under `multi_investment_periods` it appends
`-period-{p}` for each period `p`. This entry states them all over the
outage dimension. A line inactive in the snapshot's period has no row
dims: [scenario, snapshot, line, outage]
where: Line_s_nom_extendable AND at(Line_BODF, by=snapshot_period[period]) AND Line_active
expression: Line_s_monitored + at(Line_BODF, by=snapshot_period[period]) * Outage_s >= -Line_s_max_pu * Line_s_nom_ext
\[
\check{s}_{k,\xi,t} + \beta_{k,\kappa,\mathrm{snapshot\_period}(t)} \cdot \hat{s}_{\kappa,\xi,t} \ge -\overline{\mathrm{s}}_{k,\xi,t} \cdot S_{k} \qquad \forall\, k \in \mathcal{K},\ \kappa \in \mathcal{K}^{\mathrm{out}},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{ext}^{s}_{k} \wedge \beta_{k,\kappa,\mathrm{snapshot\_period}(t)} \text{ is defined} \wedge \mathrm{on}^{s}_{k,t}
\]
Line-ext-s_nom-lower#
Line_ext_s_nom_lower
Line_ext_s_nom_lower:
description: "`Line-ext-s_nom-lower` — the chosen build is at least its floor in every scenario"
dims: [scenario, line]
where: Line_s_nom_extendable AND Line_active_asset
expression: Line_s_nom_ext >= Line_s_nom_min
\[
S_{k} \ge \underline{\mathrm{s}}^{\mathrm{nom}}_{k,\xi} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi \,:\, \mathrm{ext}^{s}_{k} \wedge \mathrm{act}^{s}_{k}
\]
Line-ext-s_nom-upper#
Line_ext_s_nom_upper
Line_ext_s_nom_upper:
description: "`Line-ext-s_nom-upper` — the chosen build is at most its cap in every scenario. A cap of infinity is no row"
dims: [scenario, line]
where: Line_s_nom_extendable AND Line_active_asset AND Line_s_nom_max
expression: Line_s_nom_ext <= Line_s_nom_max
\[
S_{k} \le \overline{\mathrm{s}}^{\mathrm{nom}}_{k,\xi} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi \,:\, \mathrm{ext}^{s}_{k} \wedge \mathrm{act}^{s}_{k} \wedge \overline{\mathrm{s}}^{\mathrm{nom}}_{k,\xi} \text{ is defined}
\]
Line-ext-s-upper#
Line_ext_s_upper
Line_ext_s_upper:
description: "`Line-ext-s-upper` — an extendable line carries at most its rating of the chosen build, the loss included"
dims: [scenario, snapshot, line]
where: Line_s_nom_extendable AND Line_active
expression: Line_s + Line_loss <= Line_s_max_pu * Line_s_nom_ext
\[
s_{k,\xi,t} + \ell_{k,\xi,t} \le \overline{\mathrm{s}}_{k,\xi,t} \cdot S_{k} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{ext}^{s}_{k} \wedge \mathrm{on}^{s}_{k,t}
\]
Line-ext-s-upper-security-for-{c}-outage-in-sub-network-{n}#
Line_ext_s_upper_security
Line_ext_s_upper_security:
description: >-
`Line-ext-s-upper-security-for-{c}-outage-in-sub-network-{n}` — after
any one outage, an extendable line carries at most its rating of the
chosen build. Its flow takes on its share of the outaged branch's flow.
PyPSA names one row per outaged component `c` and sub-network `n`. Under
`multi_investment_periods` it appends `-period-{p}` for each period `p`.
This entry states them all over the outage dimension. A line inactive in
the snapshot's period has no row
dims: [scenario, snapshot, line, outage]
where: Line_s_nom_extendable AND at(Line_BODF, by=snapshot_period[period]) AND Line_active
expression: Line_s_monitored + at(Line_BODF, by=snapshot_period[period]) * Outage_s <= Line_s_max_pu * Line_s_nom_ext
\[
\check{s}_{k,\xi,t} + \beta_{k,\kappa,\mathrm{snapshot\_period}(t)} \cdot \hat{s}_{\kappa,\xi,t} \le \overline{\mathrm{s}}_{k,\xi,t} \cdot S_{k} \qquad \forall\, k \in \mathcal{K},\ \kappa \in \mathcal{K}^{\mathrm{out}},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{ext}^{s}_{k} \wedge \beta_{k,\kappa,\mathrm{snapshot\_period}(t)} \text{ is defined} \wedge \mathrm{on}^{s}_{k,t}
\]
Line-fix-s-lower#
Line_fix_s_lower
Line_fix_s_lower:
description: "`Line-fix-s-lower` — a fixed line carries at least the negative of its rating, the loss counted against it"
dims: [scenario, snapshot, line]
where: not Line_s_nom_extendable AND Line_active
expression: Line_s - Line_loss >= -Line_s_max_pu * Line_s_nom
\[
s_{k,\xi,t} - \ell_{k,\xi,t} \ge -\overline{\mathrm{s}}_{k,\xi,t} \cdot \mathrm{s}^{\mathrm{nom}}_{k,\xi} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \neg \mathrm{ext}^{s}_{k} \wedge \mathrm{on}^{s}_{k,t}
\]
Line-fix-s-lower-security-for-{c}-outage-in-sub-network-{n}#
Line_fix_s_lower_security
Line_fix_s_lower_security:
description: >-
`Line-fix-s-lower-security-for-{c}-outage-in-sub-network-{n}` — after
any one outage, a fixed line carries at least the negative of its
rating. Its flow takes on its share of the outaged branch's flow. PyPSA
names one row per outaged component `c` and sub-network `n`. Under
`multi_investment_periods` it appends `-period-{p}` for each period `p`.
This entry states them all over the outage dimension. A line inactive in
the snapshot's period has no row
dims: [scenario, snapshot, line, outage]
where: not Line_s_nom_extendable AND at(Line_BODF, by=snapshot_period[period]) AND Line_active
expression: Line_s_monitored + at(Line_BODF, by=snapshot_period[period]) * Outage_s >= -Line_s_max_pu * Line_s_nom
\[
\check{s}_{k,\xi,t} + \beta_{k,\kappa,\mathrm{snapshot\_period}(t)} \cdot \hat{s}_{\kappa,\xi,t} \ge -\overline{\mathrm{s}}_{k,\xi,t} \cdot \mathrm{s}^{\mathrm{nom}}_{k,\xi} \qquad \forall\, k \in \mathcal{K},\ \kappa \in \mathcal{K}^{\mathrm{out}},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \neg \mathrm{ext}^{s}_{k} \wedge \beta_{k,\kappa,\mathrm{snapshot\_period}(t)} \text{ is defined} \wedge \mathrm{on}^{s}_{k,t}
\]
Line-fix-s-upper#
Line_fix_s_upper
Line_fix_s_upper:
description: "`Line-fix-s-upper` — a fixed line carries at most its rating, the loss included"
dims: [scenario, snapshot, line]
where: not Line_s_nom_extendable AND Line_active
expression: Line_s + Line_loss <= Line_s_max_pu * Line_s_nom
\[
s_{k,\xi,t} + \ell_{k,\xi,t} \le \overline{\mathrm{s}}_{k,\xi,t} \cdot \mathrm{s}^{\mathrm{nom}}_{k,\xi} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \neg \mathrm{ext}^{s}_{k} \wedge \mathrm{on}^{s}_{k,t}
\]
Line-fix-s-upper-security-for-{c}-outage-in-sub-network-{n}#
Line_fix_s_upper_security
Line_fix_s_upper_security:
description: >-
`Line-fix-s-upper-security-for-{c}-outage-in-sub-network-{n}` — after
any one outage, a fixed line carries at most its rating. Its flow takes
on its share of the outaged branch's flow. PyPSA names one row per
outaged component `c` and sub-network `n`. Under
`multi_investment_periods` it appends `-period-{p}` for each period `p`.
This entry states them all over the outage dimension. A line inactive in
the snapshot's period has no row
dims: [scenario, snapshot, line, outage]
where: not Line_s_nom_extendable AND at(Line_BODF, by=snapshot_period[period]) AND Line_active
expression: Line_s_monitored + at(Line_BODF, by=snapshot_period[period]) * Outage_s <= Line_s_max_pu * Line_s_nom
\[
\check{s}_{k,\xi,t} + \beta_{k,\kappa,\mathrm{snapshot\_period}(t)} \cdot \hat{s}_{\kappa,\xi,t} \le \overline{\mathrm{s}}_{k,\xi,t} \cdot \mathrm{s}^{\mathrm{nom}}_{k,\xi} \qquad \forall\, k \in \mathcal{K},\ \kappa \in \mathcal{K}^{\mathrm{out}},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \neg \mathrm{ext}^{s}_{k} \wedge \beta_{k,\kappa,\mathrm{snapshot\_period}(t)} \text{ is defined} \wedge \mathrm{on}^{s}_{k,t}
\]
Line-loss_tangents-{k}-1#
Line_loss_tangents_forward
Line_loss_tangents_forward:
description: >-
`Line-loss_tangents-{k}-1`, `Line-loss_secants-pos` — the loss sits
above every cut to its curve for flow one way. PyPSA names one row per
tangent `k`, or one row stacked over its `secant` axis, and this entry
states them all over the segment dimension
dims: [scenario, snapshot, line, segment]
where: Line_lossy
expression: Line_loss + Line_loss_slope * Line_s >= Line_loss_offset
\[
\ell_{k,\xi,t} + \mathrm{a}_{k,\xi,m,t} \cdot s_{k,\xi,t} \ge \mathrm{b}_{k,\xi,m,t} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi,\ m \in \mathcal{M},\ t \in \mathcal{T} \,:\, \mathrm{on}^{s,\mathrm{lossy}}_{k,t}
\]
Line-loss_tangents-{k}--1#
Line_loss_tangents_reverse
Line_loss_tangents_reverse:
description: >-
`Line-loss_tangents-{k}--1`, `Line-loss_secants-neg` — the same fan
mirrored, the loss depending on the flow's magnitude
dims: [scenario, snapshot, line, segment]
where: Line_lossy
expression: Line_loss - Line_loss_slope * Line_s >= Line_loss_offset
\[
\ell_{k,\xi,t} - \mathrm{a}_{k,\xi,m,t} \cdot s_{k,\xi,t} \ge \mathrm{b}_{k,\xi,m,t} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi,\ m \in \mathcal{M},\ t \in \mathcal{T} \,:\, \mathrm{on}^{s,\mathrm{lossy}}_{k,t}
\]
Line-loss_upper#
Line_loss_upper
Line_loss_upper:
description: "`Line-loss_upper` — a line dissipates at most the loss at its rating"
dims: [scenario, snapshot, line]
where: Line_lossy
expression: Line_loss <= Line_loss_max
\[
\ell_{k,\xi,t} \le \overline{\ell}_{k,\xi,t} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{s,\mathrm{lossy}}_{k,t}
\]
Line-s_nom_modularity#
Line_s_nom_modularity
Line_s_nom_modularity:
description: "`Line-s_nom_modularity` — the chosen build is a whole number of modules"
dims: [line]
where: Line_s_nom_extendable AND Line_s_nom_mod > 0 AND Line_active_asset
expression: Line_s_nom_ext == Line_s_nom_mod * Line_n_mod
\[
S_{k} = \mathrm{s}^{\mathrm{mod}}_{k} \cdot N^{s}_{k} \qquad \forall\, k \in \mathcal{K} \,:\, \mathrm{ext}^{s}_{k} \wedge \mathrm{s}^{\mathrm{mod}}_{k} > 0 \wedge \mathrm{act}^{s}_{k}
\]
Line-s_nom_set#
Line_s_nom_set
Line_s_nom_set:
description: "`Line-s_nom_set` — the chosen build pinned, wherever a value is given"
dims: [line]
where: Line_s_nom_extendable AND Line_active_asset AND Line_s_nom_set
expression: Line_s_nom_ext == Line_s_nom_set
\[
S_{k} = \mathrm{s}^{\mathrm{nom,set}}_{k} \qquad \forall\, k \in \mathcal{K} \,:\, \mathrm{ext}^{s}_{k} \wedge \mathrm{act}^{s}_{k} \wedge \mathrm{s}^{\mathrm{nom,set}}_{k} \text{ is defined}
\]
Line-s_set#
Line_s_set
Line_s_set:
description: "`Line-s_set` — flow pinned to the given schedule, wherever one is given"
dims: [scenario, snapshot, line]
where: Line_s_set AND Line_active
expression: Line_s == Line_s_set
\[
s_{k,\xi,t} = \mathrm{s}^{\mathrm{set}}_{k,\xi,t} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{s}^{\mathrm{set}}_{k,\xi,t} \text{ is defined} \wedge \mathrm{on}^{s}_{k,t}
\]
Line-v_ang-lower#
Line_v_ang_lower
Line_v_ang_lower:
description: >-
`Line-v_ang-lower` — an AC line carries at least the flow at which the
voltage angle difference across it is the negative of its limit. That
difference is `x_pu_eff` times the flow in radians
dims: [scenario, snapshot, line]
where: Line_v_ang_max AND Line_carrier == 'AC' AND Line_active
expression: Line_s >= -Line_v_ang_max * (3.141592653589793 / 180) / Line_x_pu_eff
\[
s_{k,\xi,t} \ge \frac{-\overline{\delta}_{k,\xi} \cdot \frac{3.141592653589793}{180}}{\mathrm{x}^{\mathrm{eff}}_{k,\xi}} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \overline{\delta}_{k,\xi} \text{ is defined} \wedge \mathrm{Line\_carrier}(k) = \text{'}\mathrm{AC}\text{'} \wedge \mathrm{on}^{s}_{k,t}
\]
Line-v_ang-upper#
Line_v_ang_upper
Line_v_ang_upper:
description: >-
`Line-v_ang-upper` — an AC line carries at most the flow at which the
voltage angle difference across it reaches its limit
dims: [scenario, snapshot, line]
where: Line_v_ang_max AND Line_carrier == 'AC' AND Line_active
expression: Line_s <= Line_v_ang_max * (3.141592653589793 / 180) / Line_x_pu_eff
\[
s_{k,\xi,t} \le \frac{\overline{\delta}_{k,\xi} \cdot \frac{3.141592653589793}{180}}{\mathrm{x}^{\mathrm{eff}}_{k,\xi}} \qquad \forall\, k \in \mathcal{K},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \overline{\delta}_{k,\xi} \text{ is defined} \wedge \mathrm{Line\_carrier}(k) = \text{'}\mathrm{AC}\text{'} \wedge \mathrm{on}^{s}_{k,t}
\]
Link-com-down-time#
Link_com_down_time
Link_com_down_time:
description: >-
`Link-com-down-time` — a link stopped within its own minimum down time
is still off, and so are the modules of a modular link that stopped.
The down time it brought into the horizon counts as one stop before the
first snapshot, also where the link is modular and all its modules
stopped (`constraints.py:656-683`, PyPSA/PyPSA#2114)
dims: [scenario, snapshot, link]
where: Link_committed AND Link_min_down_time > 0
expression: >-
sum_back(Link_shut_down, along=snapshot, window=Link_min_down_time)
+ Link_down_time_brought_in <= Link_modules - Link_status
\[
\sum_{t' \in \mathcal{T} \,:\, 0 \le t - t' < \mathrm{DT}^{f}} \mathit{dn}^{f}_{l,\xi,t'} + \mathrm{rest}^{f}_{l,\xi,t} \le \mathrm{M}^{f}_{l,\xi} - u^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \mathrm{DT}^{f}_{l,\xi} > 0
\]
Link-com-ext-p-lower#
Link_com_ext_p_lower
Link_com_ext_p_lower:
description: >-
`Link-com-ext-p-lower` — a committed extendable link flows at least its
minimum of the chosen build. Off, the big M releases the row
dims: [scenario, snapshot, link]
where: Link_com_ext
expression: >-
Link_p >=
Link_p_min_pu * (Link_p_nom_ext - Link_maintenance_pu * Link_maintenance_capacity)
+ Link_big_m * Link_status - Link_big_m
\[
f_{l,\xi,t} \ge \underline{\mathrm{f}}_{l,\xi,t} \cdot \left( F_{l} - \gamma^{f}_{l,\xi} \cdot \mu^{f,\mathrm{nom}}_{l,\xi,t} \right) + \mathrm{M}^{f}_{l,\xi} \cdot u^{f}_{l,\xi,t} - \mathrm{M}^{f}_{l,\xi} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com,ext}}_{l,t}
\]
Link-com-ext-p-lower-nonneg#
Link_com_ext_p_lower_nonneg
Link_com_ext_p_lower_nonneg:
description: >-
`Link-com-ext-p-lower-nonneg` — where no minimum-per-unit is
negative, flow is also plainly non-negative, a row the big-M lower
cannot assert while the link is off
dims: [scenario, snapshot, link]
where: Link_com_ext AND Link_p_min_pu_nonneg
expression: Link_p >= 0
\[
f_{l,\xi,t} \ge 0 \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com,ext}}_{l,t} \wedge \mathrm{nonneg}^{f}_{l}
\]
Link-com-ext-p-upper-bigM#
Link_com_ext_p_upper_big_m
Link_com_ext_p_upper_big_m:
description: "`Link-com-ext-p-upper-bigM` — off, a link flows nothing. On, the big M is no bound"
dims: [scenario, snapshot, link]
where: Link_com_ext
expression: Link_p <= Link_big_m * Link_status
\[
f_{l,\xi,t} \le \mathrm{M}^{f}_{l,\xi} \cdot u^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com,ext}}_{l,t}
\]
Link-com-ext-p-upper-cap#
Link_com_ext_p_upper_cap
Link_com_ext_p_upper_cap:
description: >-
`Link-com-ext-p-upper-cap` — a committed extendable link flows
at most what is available of the chosen build, whatever its status
dims: [scenario, snapshot, link]
where: Link_com_ext
expression: Link_p <= Link_p_max_pu * (Link_p_nom_ext - Link_maintenance_pu * Link_maintenance_capacity)
\[
f_{l,\xi,t} \le \overline{\mathrm{f}}_{l,\xi,t} \cdot \left( F_{l} - \gamma^{f}_{l,\xi} \cdot \mu^{f,\mathrm{nom}}_{l,\xi,t} \right) \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com,ext}}_{l,t}
\]
Link-com-mod-p-lower#
Link_com_mod_p_lower
Link_com_mod_p_lower:
description: >-
`Link-com-mod-p-lower` — a committed modular link flows at least
its minimum of one module, whether the build is fixed or a decision
dims: [scenario, snapshot, link]
where: Link_committed AND Link_p_nom_mod > 0
expression: Link_p >= Link_p_min_pu * Link_p_nom_mod * (Link_status - Link_maintenance_pu * Link_maintenance_status)
\[
f_{l,\xi,t} \ge \underline{\mathrm{f}}_{l,\xi,t} \cdot \mathrm{f}^{\mathrm{mod}}_{l} \cdot \left( u^{f}_{l,\xi,t} - \gamma^{f}_{l,\xi} \cdot \mu^{f,u}_{l,\xi,t} \right) \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0
\]
Link-com-mod-p-upper#
Link_com_mod_p_upper
Link_com_mod_p_upper:
description: >-
`Link-com-mod-p-upper` — a committed modular link flows at most
one module's share, whether the build is fixed or a decision
dims: [scenario, snapshot, link]
where: Link_committed AND Link_p_nom_mod > 0
expression: Link_p <= Link_p_max_pu * Link_p_nom_mod * (Link_status - Link_maintenance_pu * Link_maintenance_status)
\[
f_{l,\xi,t} \le \overline{\mathrm{f}}_{l,\xi,t} \cdot \mathrm{f}^{\mathrm{mod}}_{l} \cdot \left( u^{f}_{l,\xi,t} - \gamma^{f}_{l,\xi} \cdot \mu^{f,u}_{l,\xi,t} \right) \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0
\]
Link-com-p-lower#
Link_com_p_lower
Link_com_p_lower:
description: "`Link-com-p-lower` — a committed link flows at least its minimum. Off, it flows at least nothing"
dims: [scenario, snapshot, link]
where: Link_committed AND NOT Link_p_nom_extendable AND NOT (Link_p_nom_mod > 0)
expression: Link_p >= Link_p_min_pu * Link_p_nom * (Link_status - Link_maintenance_pu * Link_maintenance_status)
\[
f_{l,\xi,t} \ge \underline{\mathrm{f}}_{l,\xi,t} \cdot \mathrm{f}^{\mathrm{nom}}_{l,\xi} \cdot \left( u^{f}_{l,\xi,t} - \gamma^{f}_{l,\xi} \cdot \mu^{f,u}_{l,\xi,t} \right) \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \neg \mathrm{ext}^{f}_{l} \wedge \neg \left( \mathrm{f}^{\mathrm{mod}}_{l} > 0 \right)
\]
Link-com-p-upper#
Link_com_p_upper
Link_com_p_upper:
description: "`Link-com-p-upper` — a committed link flows at most what is available. Off, it flows at most nothing"
dims: [scenario, snapshot, link]
where: Link_committed AND NOT Link_p_nom_extendable AND NOT (Link_p_nom_mod > 0)
expression: Link_p <= Link_p_max_pu * Link_p_nom * (Link_status - Link_maintenance_pu * Link_maintenance_status)
\[
f_{l,\xi,t} \le \overline{\mathrm{f}}_{l,\xi,t} \cdot \mathrm{f}^{\mathrm{nom}}_{l,\xi} \cdot \left( u^{f}_{l,\xi,t} - \gamma^{f}_{l,\xi} \cdot \mu^{f,u}_{l,\xi,t} \right) \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \neg \mathrm{ext}^{f}_{l} \wedge \neg \left( \mathrm{f}^{\mathrm{mod}}_{l} > 0 \right)
\]
Link-com-transition-shut-down#
Link_com_transition_shut_down
Link_com_transition_shut_down:
description: "`Link-com-transition-shut-down` — turning off is a stop, counted against the state the link carried into the snapshot"
dims: [scenario, snapshot, link]
where: Link_committed
expression: Link_shut_down >= Link_previous_status - Link_status
\[
\mathit{dn}^{f}_{l,\xi,t} \ge \overleftarrow{u}^{f}_{l,\xi,t} - u^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t}
\]
Link-com-transition-start-up#
Link_com_transition_start_up
Link_com_transition_start_up:
description: "`Link-com-transition-start-up` — turning on is a start, counted against the state the link carried into the snapshot"
dims: [scenario, snapshot, link]
where: Link_committed
expression: Link_start_up >= Link_status - Link_previous_status
\[
\mathit{up}^{f}_{l,\xi,t} \ge u^{f}_{l,\xi,t} - \overleftarrow{u}^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t}
\]
Link-com-up-time#
Link_com_up_time
Link_com_up_time:
description: >-
`Link-com-up-time` — a link started within its own minimum up time is
still on. The up time it brought into the horizon counts as a start
before the first snapshot
dims: [scenario, snapshot, link]
where: Link_committed AND Link_min_up_time > 0
expression: >-
sum_back(Link_start_up, along=snapshot, window=Link_min_up_time)
+ Link_up_time_brought_in <= Link_status
\[
\sum_{t' \in \mathcal{T} \,:\, 0 \le t - t' < \mathrm{UT}^{f}} \mathit{up}^{f}_{l,\xi,t'} + \mathrm{hold}^{f}_{l,\xi,t} \le u^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \mathrm{UT}^{f}_{l,\xi} > 0
\]
Link-e_sum_max#
Link_e_sum_max
Link_e_sum_max:
description: "`Link-e_sum_max` — energy withdrawn at `bus0` over the horizon, each period weighted by its years, is at most its budget. A budget of infinity is no row, and so is a link active in no snapshot"
dims: [scenario, link]
where: Link_e_sum_max AND count(Link_active, over=snapshot) > 0
expression: sum(Link_p * energy_limit_weight, over=snapshot) <= Link_e_sum_max
\[
\sum_{t \in \mathcal{T}} f_{l,\xi,t} \cdot \mathit{w}^{E}_{t} \le \overline{\mathrm{E}}^{f}_{l,\xi} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi \,:\, \overline{\mathrm{E}}^{f}_{l,\xi} \text{ is defined} \wedge \lvert \{ t \in \mathcal{T} \,:\, \mathrm{on}^{f}_{l,t} \} \rvert > 0
\]
Link-e_sum_min#
Link_e_sum_min
Link_e_sum_min:
description: "`Link-e_sum_min` — energy withdrawn at `bus0` over the horizon, each period weighted by its years, is at least its floor. A floor of minus infinity is no row, and so is a link active in no snapshot"
dims: [scenario, link]
where: Link_e_sum_min AND count(Link_active, over=snapshot) > 0
expression: sum(Link_p * energy_limit_weight, over=snapshot) >= Link_e_sum_min
\[
\sum_{t \in \mathcal{T}} f_{l,\xi,t} \cdot \mathit{w}^{E}_{t} \ge \underline{\mathrm{E}}^{f}_{l,\xi} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi \,:\, \underline{\mathrm{E}}^{f}_{l,\xi} \text{ is defined} \wedge \lvert \{ t \in \mathcal{T} \,:\, \mathrm{on}^{f}_{l,t} \} \rvert > 0
\]
Link-ext-p-lower#
Link_ext_p_lower
Link_ext_p_lower:
description: "`Link-ext-p-lower` — an extendable link carries at least its minimum of the chosen build, negative for the other way"
dims: [scenario, snapshot, link]
where: Link_p_nom_extendable AND not Link_committable AND Link_active
expression: Link_p >= Link_p_min_pu * (Link_p_nom_ext - Link_maintenance_pu * Link_maintenance_capacity)
\[
f_{l,\xi,t} \ge \underline{\mathrm{f}}_{l,\xi,t} \cdot \left( F_{l} - \gamma^{f}_{l,\xi} \cdot \mu^{f,\mathrm{nom}}_{l,\xi,t} \right) \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{ext}^{f}_{l} \wedge \neg \mathrm{com}^{f}_{l} \wedge \mathrm{on}^{f}_{l,t}
\]
Link-ext-p_nom-lower#
Link_ext_p_nom_lower
Link_ext_p_nom_lower:
description: "`Link-ext-p_nom-lower` — the chosen build is at least its floor in every scenario"
dims: [scenario, link]
where: Link_p_nom_extendable AND Link_active_asset
expression: Link_p_nom_ext >= Link_p_nom_min
\[
F_{l} \ge \underline{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi \,:\, \mathrm{ext}^{f}_{l} \wedge \mathrm{act}^{f}_{l}
\]
Link-ext-p_nom-upper#
Link_ext_p_nom_upper
Link_ext_p_nom_upper:
description: "`Link-ext-p_nom-upper` — the chosen build is at most its cap in every scenario. A cap of infinity is no row"
dims: [scenario, link]
where: Link_p_nom_extendable AND Link_active_asset AND Link_p_nom_max
expression: Link_p_nom_ext <= Link_p_nom_max
\[
F_{l} \le \overline{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi \,:\, \mathrm{ext}^{f}_{l} \wedge \mathrm{act}^{f}_{l} \wedge \overline{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} \text{ is defined}
\]
Link-ext-p-upper#
Link_ext_p_upper
Link_ext_p_upper:
description: "`Link-ext-p-upper` — an extendable link carries at most the chosen build"
dims: [scenario, snapshot, link]
where: Link_p_nom_extendable AND not Link_committable AND Link_active
expression: Link_p <= Link_p_max_pu * (Link_p_nom_ext - Link_maintenance_pu * Link_maintenance_capacity)
\[
f_{l,\xi,t} \le \overline{\mathrm{f}}_{l,\xi,t} \cdot \left( F_{l} - \gamma^{f}_{l,\xi} \cdot \mu^{f,\mathrm{nom}}_{l,\xi,t} \right) \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{ext}^{f}_{l} \wedge \neg \mathrm{com}^{f}_{l} \wedge \mathrm{on}^{f}_{l,t}
\]
Link-fix-p-lower#
Link_fix_p_lower
Link_fix_p_lower:
description: "`Link-fix-p-lower` — a fixed link carries at least its minimum, negative for the other way"
dims: [scenario, snapshot, link]
where: not Link_p_nom_extendable AND not Link_committable AND Link_active
expression: Link_p >= Link_p_min_pu * Link_p_nom * (1 - Link_maintenance_pu * Link_maintenance)
\[
f_{l,\xi,t} \ge \underline{\mathrm{f}}_{l,\xi,t} \cdot \mathrm{f}^{\mathrm{nom}}_{l,\xi} \cdot \left( 1 - \gamma^{f}_{l,\xi} \cdot \mu^{f}_{l,\xi,t} \right) \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \neg \mathrm{ext}^{f}_{l} \wedge \neg \mathrm{com}^{f}_{l} \wedge \mathrm{on}^{f}_{l,t}
\]
Link-fix-p-upper#
Link_fix_p_upper
Link_fix_p_upper:
description: "`Link-fix-p-upper` — a fixed link carries at most its nominal power"
dims: [scenario, snapshot, link]
where: not Link_p_nom_extendable AND not Link_committable AND Link_active
expression: Link_p <= Link_p_max_pu * Link_p_nom * (1 - Link_maintenance_pu * Link_maintenance)
\[
f_{l,\xi,t} \le \overline{\mathrm{f}}_{l,\xi,t} \cdot \mathrm{f}^{\mathrm{nom}}_{l,\xi} \cdot \left( 1 - \gamma^{f}_{l,\xi} \cdot \mu^{f}_{l,\xi,t} \right) \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \neg \mathrm{ext}^{f}_{l} \wedge \neg \mathrm{com}^{f}_{l} \wedge \mathrm{on}^{f}_{l,t}
\]
Link-maint-event-count#
Link_maint_event_count
Link_maint_event_count:
description: "`Link-maint-event-count` — a maintainable link holds its number of maintenance events over the horizon"
dims: [scenario, link]
where: Link_maintainable
expression: sum(Link_maintenance_start, over=snapshot) == Link_maintenance_events
\[
\sum_{t \in \mathcal{T}} \mu^{f,\mathrm{up}}_{l,\xi,t} = \mathrm{n}^{f,\mathrm{mnt}}_{l,\xi} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi \,:\, \mathrm{mnt}^{f}_{l}
\]
Link-maint-modstatus-lb#
Link_maint_modstatus_lb
Link_maint_modstatus_lb:
description: "`Link-maint-modstatus-lb` — in maintenance, every module on is on in maintenance"
dims: [scenario, snapshot, link]
where: Link_maintainable AND Link_committed AND Link_p_nom_mod > 0
expression: Link_maintenance_status >= Link_status - Link_maintenance_modules * (1 - Link_maintenance)
\[
\mu^{f,u}_{l,\xi,t} \ge u^{f}_{l,\xi,t} - \mathrm{Link\_maintenance\_modules}_{l,\xi} \cdot \left( 1 - \mu^{f}_{l,\xi,t} \right) \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{f}_{l} \wedge \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0
\]
Link-maint-modstatus-le-maint#
Link_maint_modstatus_le_maint
Link_maint_modstatus_le_maint:
description: >-
`Link-maint-modstatus-le-maint` — out of maintenance, no module is on in
maintenance. In it, at most its modules in maintenance
dims: [scenario, snapshot, link]
where: Link_maintainable AND Link_committed AND Link_p_nom_mod > 0
expression: Link_maintenance_status <= Link_maintenance_modules * Link_maintenance
\[
\mu^{f,u}_{l,\xi,t} \le \mathrm{Link\_maintenance\_modules}_{l,\xi} \cdot \mu^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{f}_{l} \wedge \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0
\]
Link-maint-modstatus-le-status#
Link_maint_modstatus_le_status
Link_maint_modstatus_le_status:
description: "`Link-maint-modstatus-le-status` — the modules on in maintenance are at most the modules on"
dims: [scenario, snapshot, link]
where: Link_maintainable AND Link_committed AND Link_p_nom_mod > 0
expression: Link_maintenance_status <= Link_status
\[
\mu^{f,u}_{l,\xi,t} \le u^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{f}_{l} \wedge \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0
\]
Link-maint-start-horizon#
Link_maint_start_horizon
Link_maint_start_horizon:
description: "`Link-maint-start-horizon` — no event starts where it could not run its whole duration"
dims: [scenario, snapshot, link]
where: Link_maintainable AND Link_active AND Link_maintenance_start_blocked
expression: Link_maintenance_start == 0
\[
\mu^{f,\mathrm{up}}_{l,\xi,t} = 0 \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{f}_{l} \wedge \mathrm{on}^{f}_{l,t} \wedge \mathrm{blk}^{f}_{l,\xi,t}
\]
Link-maint-status-lb#
Link_maint_status_lb
Link_maint_status_lb:
description: "`Link-maint-status-lb` — on and in maintenance, the status in maintenance is one"
dims: [scenario, snapshot, link]
where: Link_maintainable AND Link_committed AND NOT Link_p_nom_extendable AND NOT (Link_p_nom_mod > 0)
expression: Link_maintenance_status >= Link_status + Link_maintenance - 1
\[
\mu^{f,u}_{l,\xi,t} \ge u^{f}_{l,\xi,t} + \mu^{f}_{l,\xi,t} - 1 \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{f}_{l} \wedge \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \neg \mathrm{ext}^{f}_{l} \wedge \neg \left( \mathrm{f}^{\mathrm{mod}}_{l} > 0 \right)
\]
Link-maint-status-le-maint#
Link_maint_status_le_maint
Link_maint_status_le_maint:
description: "`Link-maint-status-le-maint` — out of maintenance, the status in maintenance is zero"
dims: [scenario, snapshot, link]
where: Link_maintainable AND Link_committed AND NOT Link_p_nom_extendable AND NOT (Link_p_nom_mod > 0)
expression: Link_maintenance_status <= Link_maintenance
\[
\mu^{f,u}_{l,\xi,t} \le \mu^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{f}_{l} \wedge \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \neg \mathrm{ext}^{f}_{l} \wedge \neg \left( \mathrm{f}^{\mathrm{mod}}_{l} > 0 \right)
\]
Link-maint-status-le-status#
Link_maint_status_le_status
Link_maint_status_le_status:
description: "`Link-maint-status-le-status` — the status in maintenance is at most the status"
dims: [scenario, snapshot, link]
where: Link_maintainable AND Link_committed AND NOT Link_p_nom_extendable AND NOT (Link_p_nom_mod > 0)
expression: Link_maintenance_status <= Link_status
\[
\mu^{f,u}_{l,\xi,t} \le u^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{f}_{l} \wedge \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \neg \mathrm{ext}^{f}_{l} \wedge \neg \left( \mathrm{f}^{\mathrm{mod}}_{l} > 0 \right)
\]
Link-maint-window#
Link_maint_window
Link_maint_window:
description: >-
`Link-maint-window` — a link is in maintenance exactly where an event it
started covers the snapshot. Two events do not overlap, since the
maintenance status is at most one
dims: [scenario, snapshot, link]
where: Link_maintainable AND Link_active
expression: Link_maintenance == sum(Link_maintenance_start, over=start, by=Link_maintenance_cover[covered])
\[
\mu^{f}_{l,\xi,t} = \sum_{t' \in \mathcal{T} \,:\, \left( t,\ l,\ \xi,\ t' \right) \in \mathrm{Link\_maintenance\_cover}} \mu^{f,\mathrm{up}}_{l,\xi,t'} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{f}_{l} \wedge \mathrm{on}^{f}_{l,t}
\]
Link-maintcap_lower_nommax#
Link_maintcap_lower_nommax
Link_maintcap_lower_nommax:
description: "`Link-maintcap_lower_nommax` — in maintenance, the whole chosen build is taken off"
dims: [scenario, snapshot, link]
where: Link_maint_ext
expression: Link_maintenance_capacity >= Link_p_nom_ext - Link_p_nom_max * (1 - Link_maintenance)
\[
\mu^{f,\mathrm{nom}}_{l,\xi,t} \ge F_{l} - \overline{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} \cdot \left( 1 - \mu^{f}_{l,\xi,t} \right) \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{mnt,ext}}_{l,t}
\]
Link-maintcap_lower_nommin#
Link_maintcap_lower_nommin
Link_maintcap_lower_nommin:
description: "`Link-maintcap_lower_nommin` — in maintenance, at least the floor of the build is taken off"
dims: [scenario, snapshot, link]
where: Link_maint_ext AND Link_p_nom_min > 0
expression: Link_maintenance_capacity >= Link_p_nom_min * Link_maintenance
\[
\mu^{f,\mathrm{nom}}_{l,\xi,t} \ge \underline{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} \cdot \mu^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{mnt,ext}}_{l,t} \wedge \underline{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} > 0
\]
Link-maintcap_upper#
Link_maintcap_upper
Link_maintcap_upper:
description: >-
`Link-maintcap_upper` — the build taken off is at most the chosen build in
maintenance, and at most the build less its floor out of it
dims: [scenario, snapshot, link]
where: Link_maint_ext
expression: Link_maintenance_capacity <= Link_p_nom_ext - Link_p_nom_min * (1 - Link_maintenance)
\[
\mu^{f,\mathrm{nom}}_{l,\xi,t} \le F_{l} - \underline{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} \cdot \left( 1 - \mu^{f}_{l,\xi,t} \right) \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{mnt,ext}}_{l,t}
\]
Link-maintcap_upper_nommax#
Link_maintcap_upper_nommax
Link_maintcap_upper_nommax:
description: "`Link-maintcap_upper_nommax` — out of maintenance, no build is taken off"
dims: [scenario, snapshot, link]
where: Link_maint_ext
expression: Link_maintenance_capacity <= Link_p_nom_max * Link_maintenance
\[
\mu^{f,\mathrm{nom}}_{l,\xi,t} \le \overline{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} \cdot \mu^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{mnt,ext}}_{l,t}
\]
Link-p_nom_modularity#
Link_p_nom_modularity
Link_p_nom_modularity:
description: "`Link-p_nom_modularity` — the chosen build is a whole number of modules"
dims: [link]
where: Link_p_nom_extendable AND Link_p_nom_mod > 0 AND Link_active_asset
expression: Link_p_nom_ext == Link_p_nom_mod * Link_n_mod
\[
F_{l} = \mathrm{f}^{\mathrm{mod}}_{l} \cdot N^{f}_{l} \qquad \forall\, l \in \mathcal{L} \,:\, \mathrm{ext}^{f}_{l} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0 \wedge \mathrm{act}^{f}_{l}
\]
Link-p_nom_set#
Link_p_nom_set
Link_p_nom_set:
description: "`Link-p_nom_set` — the chosen build pinned, wherever a value is given"
dims: [link]
where: Link_p_nom_extendable AND Link_active_asset AND Link_p_nom_set
expression: Link_p_nom_ext == Link_p_nom_set
\[
F_{l} = \mathrm{f}^{\mathrm{nom,set}}_{l} \qquad \forall\, l \in \mathcal{L} \,:\, \mathrm{ext}^{f}_{l} \wedge \mathrm{act}^{f}_{l} \wedge \mathrm{f}^{\mathrm{nom,set}}_{l} \text{ is defined}
\]
Link-p-ramp_limit_down#
Link_p_ramp_limit_down
Link_p_ramp_limit_down:
description: >-
`Link-p-ramp_limit_down` — a link lowers flow no faster than its ramp
limit of the build. A committed one lowers it no further than its
shut-down ramp in the snapshot it turns off. A link that came into the
horizon running has a row at the first snapshot only where its `p_init`
gives the flow it brought in. No link has one at the start of a later
investment period, nor does any link a big M releases instead
dims: [scenario, snapshot, link]
where: >-
(Link_ramp_limit_down OR Link_ramp_limit_shut_down)
AND NOT (Link_committable AND Link_p_nom_extendable AND NOT (Link_p_nom_mod > 0))
AND Link_ramps_from_previous
AND Link_active
expression: Link_previous_p - Link_p <= Link_ramp_down_allowance
\[
\overleftarrow{f}_{l,\xi,t} - f_{l,\xi,t} \le \Delta^{f,-}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \left( \mathrm{rd}^{f}_{l,\xi,t} \text{ is defined} \vee \mathrm{rd}^{f,\mathrm{dn}}_{l,\xi} \text{ is defined} \right) \wedge \neg \left( \mathrm{com}^{f}_{l} \wedge \mathrm{ext}^{f}_{l} \wedge \neg \left( \mathrm{f}^{\mathrm{mod}}_{l} > 0 \right) \right) \wedge \mathrm{prev}^{f}_{l,\xi,t} \wedge \mathrm{on}^{f}_{l,t}
\]
Link-p-ramp_limit_down-run-bigM#
Link_p_ramp_limit_down_run_big_m
Link_p_ramp_limit_down_run_big_m:
description: >-
`Link-p-ramp_limit_down-run-bigM` — a committed extendable link lowers
flow no faster than its limit of the chosen build. The big M releases
the row in the snapshot it turns off
dims: [scenario, snapshot, link]
where: Link_com_ext AND (Link_ramp_limit_down OR Link_ramp_limit_shut_down) AND Link_ramps_from_previous
expression: >-
Link_p - Link_previous_p >=
-Link_ramp_down_rate * Link_p_nom_ext
- Link_big_m + Link_big_m * Link_status
\[
f_{l,\xi,t} - \overleftarrow{f}_{l,\xi,t} \ge -\widetilde{\mathrm{rd}}^{f}_{l,\xi,t} \cdot F_{l} - \mathrm{M}^{f}_{l,\xi} + \mathrm{M}^{f}_{l,\xi} \cdot u^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com,ext}}_{l,t} \wedge \left( \mathrm{rd}^{f}_{l,\xi,t} \text{ is defined} \vee \mathrm{rd}^{f,\mathrm{dn}}_{l,\xi} \text{ is defined} \right) \wedge \mathrm{prev}^{f}_{l,\xi,t}
\]
Link-p-ramp_limit_down-shut-bigM#
Link_p_ramp_limit_down_shut_big_m
Link_p_ramp_limit_down_shut_big_m:
description: >-
`Link-p-ramp_limit_down-shut-bigM` — in the snapshot it turns off, a
committed extendable link ramps no further than its shut-down ramp of
the chosen build. The big M releases the row everywhere else
dims: [scenario, snapshot, link]
where: Link_com_ext AND (Link_ramp_limit_down OR Link_ramp_limit_shut_down) AND Link_ramps_from_previous
expression: >-
Link_previous_p - Link_p <=
Link_shut_down_rate * Link_p_nom_ext
+ Link_big_m - Link_big_m * Link_shut_down
\[
\overleftarrow{f}_{l,\xi,t} - f_{l,\xi,t} \le \widetilde{\mathrm{rd}}^{f,\mathrm{dn}}_{l,\xi} \cdot F_{l} + \mathrm{M}^{f}_{l,\xi} - \mathrm{M}^{f}_{l,\xi} \cdot \mathit{dn}^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com,ext}}_{l,t} \wedge \left( \mathrm{rd}^{f}_{l,\xi,t} \text{ is defined} \vee \mathrm{rd}^{f,\mathrm{dn}}_{l,\xi} \text{ is defined} \right) \wedge \mathrm{prev}^{f}_{l,\xi,t}
\]
Link-p-ramp_limit_up#
Link_p_ramp_limit_up
Link_p_ramp_limit_up:
description: >-
`Link-p-ramp_limit_up` — a link raises flow no faster than its ramp
limit of the build. A committed one raises it no further than its
start-up ramp in the snapshot it turns on. A link that came into the
horizon running has a row at the first snapshot only where its `p_init`
gives the flow it brought in. No link has one at the start of a later
investment period, nor does any link a big M releases instead
dims: [scenario, snapshot, link]
where: >-
(Link_ramp_limit_up OR Link_ramp_limit_start_up)
AND NOT (Link_committable AND Link_p_nom_extendable AND NOT (Link_p_nom_mod > 0))
AND Link_ramps_from_previous
AND Link_active
expression: Link_p - Link_previous_p <= Link_ramp_up_allowance
\[
f_{l,\xi,t} - \overleftarrow{f}_{l,\xi,t} \le \Delta^{f,+}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \left( \mathrm{ru}^{f}_{l,\xi,t} \text{ is defined} \vee \mathrm{ru}^{f,\mathrm{up}}_{l,\xi} \text{ is defined} \right) \wedge \neg \left( \mathrm{com}^{f}_{l} \wedge \mathrm{ext}^{f}_{l} \wedge \neg \left( \mathrm{f}^{\mathrm{mod}}_{l} > 0 \right) \right) \wedge \mathrm{prev}^{f}_{l,\xi,t} \wedge \mathrm{on}^{f}_{l,t}
\]
Link-p-ramp_limit_up-run-bigM#
Link_p_ramp_limit_up_run_big_m
Link_p_ramp_limit_up_run_big_m:
description: >-
`Link-p-ramp_limit_up-run-bigM` — a committed extendable link raises
flow no faster than its limit of the chosen build. The big M releases
the row in the snapshot it turns on
dims: [scenario, snapshot, link]
where: Link_com_ext AND (Link_ramp_limit_up OR Link_ramp_limit_start_up) AND Link_ramps_from_previous
expression: >-
Link_p - Link_previous_p <=
Link_ramp_up_rate * Link_p_nom_ext
+ Link_big_m - Link_big_m * Link_previous_status
\[
f_{l,\xi,t} - \overleftarrow{f}_{l,\xi,t} \le \widetilde{\mathrm{ru}}^{f}_{l,\xi,t} \cdot F_{l} + \mathrm{M}^{f}_{l,\xi} - \mathrm{M}^{f}_{l,\xi} \cdot \overleftarrow{u}^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com,ext}}_{l,t} \wedge \left( \mathrm{ru}^{f}_{l,\xi,t} \text{ is defined} \vee \mathrm{ru}^{f,\mathrm{up}}_{l,\xi} \text{ is defined} \right) \wedge \mathrm{prev}^{f}_{l,\xi,t}
\]
Link-p-ramp_limit_up-start-bigM#
Link_p_ramp_limit_up_start_big_m
Link_p_ramp_limit_up_start_big_m:
description: >-
`Link-p-ramp_limit_up-start-bigM` — in the snapshot it turns on, a
committed extendable link ramps no further than its start-up ramp of the
chosen build. The big M releases the row everywhere else
dims: [scenario, snapshot, link]
where: Link_com_ext AND (Link_ramp_limit_up OR Link_ramp_limit_start_up) AND Link_ramps_from_previous
expression: >-
Link_p - Link_previous_p <=
Link_start_up_rate * Link_p_nom_ext
+ Link_big_m - Link_big_m * Link_start_up
\[
f_{l,\xi,t} - \overleftarrow{f}_{l,\xi,t} \le \widetilde{\mathrm{ru}}^{f,\mathrm{up}}_{l,\xi} \cdot F_{l} + \mathrm{M}^{f}_{l,\xi} - \mathrm{M}^{f}_{l,\xi} \cdot \mathit{up}^{f}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com,ext}}_{l,t} \wedge \left( \mathrm{ru}^{f}_{l,\xi,t} \text{ is defined} \vee \mathrm{ru}^{f,\mathrm{up}}_{l,\xi} \text{ is defined} \right) \wedge \mathrm{prev}^{f}_{l,\xi,t}
\]
Link-p_set#
Link_p_set
Link_p_set:
description: "`Link-p_set` — flow pinned to the given schedule, wherever one is given"
dims: [scenario, snapshot, link]
where: Link_p_set AND Link_active
expression: Link_p == Link_p_set
\[
f_{l,\xi,t} = \mathrm{f}^{\mathrm{set}}_{l,\xi,t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{f}^{\mathrm{set}}_{l,\xi,t} \text{ is defined} \wedge \mathrm{on}^{f}_{l,t}
\]
Link-shut_down-p-fixed-upper#
Link_shut_down_p_fixed_upper
Link_shut_down_p_fixed_upper:
description: >-
`Link-shut_down-p-fixed-upper` — a stop is at most the modules in place,
an explicit row as PyPSA writes it. The bound is one where the build is
not modular, and the fixed build's whole count of modules where it is
dims: [scenario, snapshot, link]
where: Link_committed AND NOT (Link_p_nom_extendable AND Link_p_nom_mod > 0)
expression: Link_shut_down <= Link_modules_installed
\[
\mathit{dn}^{f}_{l,\xi,t} \le \mathrm{N}^{f,\mathrm{fix}}_{l,\xi} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \neg \left( \mathrm{ext}^{f}_{l} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0 \right)
\]
Link-shut_down-p_nom-variable-upper#
Link_shut_down_p_nom_variable_upper
Link_shut_down_p_nom_variable_upper:
description: "`Link-shut_down-p_nom-variable-upper` — a modular link stops only where a module is built"
dims: [scenario, snapshot, link]
where: Link_committed AND Link_p_nom_extendable AND Link_p_nom_mod > 0
expression: Link_shut_down <= Link_n_mod
\[
\mathit{dn}^{f}_{l,\xi,t} \le N^{f}_{l} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \mathrm{ext}^{f}_{l} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0
\]
Link-start_up-p-fixed-upper#
Link_start_up_p_fixed_upper
Link_start_up_p_fixed_upper:
description: >-
`Link-start_up-p-fixed-upper` — a start is at most the modules in place,
an explicit row as PyPSA writes it. The bound is one where the build is
not modular, and the fixed build's whole count of modules where it is
dims: [scenario, snapshot, link]
where: Link_committed AND NOT (Link_p_nom_extendable AND Link_p_nom_mod > 0)
expression: Link_start_up <= Link_modules_installed
\[
\mathit{up}^{f}_{l,\xi,t} \le \mathrm{N}^{f,\mathrm{fix}}_{l,\xi} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \neg \left( \mathrm{ext}^{f}_{l} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0 \right)
\]
Link-start_up-p_nom-variable-upper#
Link_start_up_p_nom_variable_upper
Link_start_up_p_nom_variable_upper:
description: "`Link-start_up-p_nom-variable-upper` — a modular link starts only where a module is built"
dims: [scenario, snapshot, link]
where: Link_committed AND Link_p_nom_extendable AND Link_p_nom_mod > 0
expression: Link_start_up <= Link_n_mod
\[
\mathit{up}^{f}_{l,\xi,t} \le N^{f}_{l} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \mathrm{ext}^{f}_{l} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0
\]
Link-status-p-fixed-upper#
Link_status_p_fixed_upper
Link_status_p_fixed_upper:
description: >-
`Link-status-p-fixed-upper` — a status is at most the modules in place,
an explicit row as PyPSA writes it. The bound is one where the build is
not modular, and the fixed build's whole count of modules where it is
dims: [scenario, snapshot, link]
where: Link_committed AND NOT (Link_p_nom_extendable AND Link_p_nom_mod > 0)
expression: Link_status <= Link_modules_installed
\[
u^{f}_{l,\xi,t} \le \mathrm{N}^{f,\mathrm{fix}}_{l,\xi} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \neg \left( \mathrm{ext}^{f}_{l} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0 \right)
\]
Link-status-p_nom-variable-upper#
Link_status_p_nom_variable_upper
Link_status_p_nom_variable_upper:
description: "`Link-status-p_nom-variable-upper` — a modular link is on only where a module is built"
dims: [scenario, snapshot, link]
where: Link_committed AND Link_p_nom_extendable AND Link_p_nom_mod > 0
expression: Link_status <= Link_n_mod
\[
u^{f}_{l,\xi,t} \le N^{f}_{l} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com}}_{l,t} \wedge \mathrm{ext}^{f}_{l} \wedge \mathrm{f}^{\mathrm{mod}}_{l} > 0
\]
Process-com-down-time#
Process_com_down_time
Process_com_down_time:
description: >-
`Process-com-down-time` — a process stopped within its own minimum down time
is still off, and so are the modules of a modular process that stopped.
The down time it brought into the horizon counts as one stop before the
first snapshot, also where the process is modular and all its modules
stopped (`constraints.py:656-683`, PyPSA/PyPSA#2114)
dims: [scenario, snapshot, process]
where: Process_committed AND Process_min_down_time > 0
expression: >-
sum_back(Process_shut_down, along=snapshot, window=Process_min_down_time)
+ Process_down_time_brought_in <= Process_modules - Process_status
\[
\sum_{t' \in \mathcal{T} \,:\, 0 \le t - t' < \mathrm{DT}^{z}} \mathit{dn}^{z}_{j,\xi,t'} + \mathrm{rest}^{z}_{j,\xi,t} \le \mathrm{M}^{z}_{j,\xi} - u^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \mathrm{DT}^{z}_{j,\xi} > 0
\]
Process-com-ext-p-lower#
Process_com_ext_p_lower
Process_com_ext_p_lower:
description: >-
`Process-com-ext-p-lower` — a committed extendable process runs at least
its minimum of the chosen build. Off, the big M releases the row
dims: [scenario, snapshot, process]
where: Process_com_ext
expression: >-
Process_p >=
Process_p_min_pu * (Process_p_nom_ext - Process_maintenance_pu * Process_maintenance_capacity)
+ Process_big_m * Process_status - Process_big_m
\[
z_{j,\xi,t} \ge \underline{\mathrm{z}}_{j,\xi,t} \cdot \left( Z_{j} - \gamma^{z}_{j,\xi} \cdot \mu^{z,\mathrm{nom}}_{j,\xi,t} \right) + \mathrm{M}^{z}_{j,\xi} \cdot u^{z}_{j,\xi,t} - \mathrm{M}^{z}_{j,\xi} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com,ext}}_{j,t}
\]
Process-com-ext-p-lower-nonneg#
Process_com_ext_p_lower_nonneg
Process_com_ext_p_lower_nonneg:
description: >-
`Process-com-ext-p-lower-nonneg` — where no minimum-per-unit is
negative, internal power is also plainly non-negative, a row the big-M lower
cannot assert while the process is off
dims: [scenario, snapshot, process]
where: Process_com_ext AND Process_p_min_pu_nonneg
expression: Process_p >= 0
\[
z_{j,\xi,t} \ge 0 \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com,ext}}_{j,t} \wedge \mathrm{nonneg}^{z}_{j}
\]
Process-com-ext-p-upper-bigM#
Process_com_ext_p_upper_big_m
Process_com_ext_p_upper_big_m:
description: "`Process-com-ext-p-upper-bigM` — off, a process does not run. On, the big M is no bound"
dims: [scenario, snapshot, process]
where: Process_com_ext
expression: Process_p <= Process_big_m * Process_status
\[
z_{j,\xi,t} \le \mathrm{M}^{z}_{j,\xi} \cdot u^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com,ext}}_{j,t}
\]
Process-com-ext-p-upper-cap#
Process_com_ext_p_upper_cap
Process_com_ext_p_upper_cap:
description: >-
`Process-com-ext-p-upper-cap` — a committed extendable process runs
at most what is available of the chosen build, whatever its status
dims: [scenario, snapshot, process]
where: Process_com_ext
expression: Process_p <= Process_p_max_pu * (Process_p_nom_ext - Process_maintenance_pu * Process_maintenance_capacity)
\[
z_{j,\xi,t} \le \overline{\mathrm{z}}_{j,\xi,t} \cdot \left( Z_{j} - \gamma^{z}_{j,\xi} \cdot \mu^{z,\mathrm{nom}}_{j,\xi,t} \right) \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com,ext}}_{j,t}
\]
Process-com-mod-p-lower#
Process_com_mod_p_lower
Process_com_mod_p_lower:
description: >-
`Process-com-mod-p-lower` — a committed modular process runs at least
its minimum of one module, whether the build is fixed or a decision
dims: [scenario, snapshot, process]
where: Process_committed AND Process_p_nom_mod > 0
expression: Process_p >= Process_p_min_pu * Process_p_nom_mod * (Process_status - Process_maintenance_pu * Process_maintenance_status)
\[
z_{j,\xi,t} \ge \underline{\mathrm{z}}_{j,\xi,t} \cdot \mathrm{z}^{\mathrm{mod}}_{j} \cdot \left( u^{z}_{j,\xi,t} - \gamma^{z}_{j,\xi} \cdot \mu^{z,u}_{j,\xi,t} \right) \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0
\]
Process-com-mod-p-upper#
Process_com_mod_p_upper
Process_com_mod_p_upper:
description: >-
`Process-com-mod-p-upper` — a committed modular process runs at most
one module's share, whether the build is fixed or a decision
dims: [scenario, snapshot, process]
where: Process_committed AND Process_p_nom_mod > 0
expression: Process_p <= Process_p_max_pu * Process_p_nom_mod * (Process_status - Process_maintenance_pu * Process_maintenance_status)
\[
z_{j,\xi,t} \le \overline{\mathrm{z}}_{j,\xi,t} \cdot \mathrm{z}^{\mathrm{mod}}_{j} \cdot \left( u^{z}_{j,\xi,t} - \gamma^{z}_{j,\xi} \cdot \mu^{z,u}_{j,\xi,t} \right) \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0
\]
Process-com-p-lower#
Process_com_p_lower
Process_com_p_lower:
description: "`Process-com-p-lower` — a committed process runs at least its minimum. Off, it runs at least nothing"
dims: [scenario, snapshot, process]
where: Process_committed AND NOT Process_p_nom_extendable AND NOT (Process_p_nom_mod > 0)
expression: Process_p >= Process_p_min_pu * Process_p_nom * (Process_status - Process_maintenance_pu * Process_maintenance_status)
\[
z_{j,\xi,t} \ge \underline{\mathrm{z}}_{j,\xi,t} \cdot \mathrm{z}^{\mathrm{nom}}_{j,\xi} \cdot \left( u^{z}_{j,\xi,t} - \gamma^{z}_{j,\xi} \cdot \mu^{z,u}_{j,\xi,t} \right) \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \neg \mathrm{ext}^{z}_{j} \wedge \neg \left( \mathrm{z}^{\mathrm{mod}}_{j} > 0 \right)
\]
Process-com-p-upper#
Process_com_p_upper
Process_com_p_upper:
description: "`Process-com-p-upper` — a committed process runs at most what is available. Off, it runs at most nothing"
dims: [scenario, snapshot, process]
where: Process_committed AND NOT Process_p_nom_extendable AND NOT (Process_p_nom_mod > 0)
expression: Process_p <= Process_p_max_pu * Process_p_nom * (Process_status - Process_maintenance_pu * Process_maintenance_status)
\[
z_{j,\xi,t} \le \overline{\mathrm{z}}_{j,\xi,t} \cdot \mathrm{z}^{\mathrm{nom}}_{j,\xi} \cdot \left( u^{z}_{j,\xi,t} - \gamma^{z}_{j,\xi} \cdot \mu^{z,u}_{j,\xi,t} \right) \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \neg \mathrm{ext}^{z}_{j} \wedge \neg \left( \mathrm{z}^{\mathrm{mod}}_{j} > 0 \right)
\]
Process-com-transition-shut-down#
Process_com_transition_shut_down
Process_com_transition_shut_down:
description: "`Process-com-transition-shut-down` — turning off is a stop, counted against the state the process carried into the snapshot"
dims: [scenario, snapshot, process]
where: Process_committed
expression: Process_shut_down >= Process_previous_status - Process_status
\[
\mathit{dn}^{z}_{j,\xi,t} \ge \overleftarrow{u}^{z}_{j,\xi,t} - u^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t}
\]
Process-com-transition-start-up#
Process_com_transition_start_up
Process_com_transition_start_up:
description: "`Process-com-transition-start-up` — turning on is a start, counted against the state the process carried into the snapshot"
dims: [scenario, snapshot, process]
where: Process_committed
expression: Process_start_up >= Process_status - Process_previous_status
\[
\mathit{up}^{z}_{j,\xi,t} \ge u^{z}_{j,\xi,t} - \overleftarrow{u}^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t}
\]
Process-com-up-time#
Process_com_up_time
Process_com_up_time:
description: >-
`Process-com-up-time` — a process started within its own minimum up time is
still on. The up time it brought into the horizon counts as a start
before the first snapshot
dims: [scenario, snapshot, process]
where: Process_committed AND Process_min_up_time > 0
expression: >-
sum_back(Process_start_up, along=snapshot, window=Process_min_up_time)
+ Process_up_time_brought_in <= Process_status
\[
\sum_{t' \in \mathcal{T} \,:\, 0 \le t - t' < \mathrm{UT}^{z}} \mathit{up}^{z}_{j,\xi,t'} + \mathrm{hold}^{z}_{j,\xi,t} \le u^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \mathrm{UT}^{z}_{j,\xi} > 0
\]
Process-e_sum_max#
Process_e_sum_max
Process_e_sum_max:
description: "`Process-e_sum_max` — energy of the internal power over the horizon, each period weighted by its years, is at most its budget. A budget of infinity is no row, and so is a process active in no snapshot"
dims: [scenario, process]
where: Process_e_sum_max AND count(Process_active, over=snapshot) > 0
expression: sum(Process_p * energy_limit_weight, over=snapshot) <= Process_e_sum_max
\[
\sum_{t \in \mathcal{T}} z_{j,\xi,t} \cdot \mathit{w}^{E}_{t} \le \overline{\mathrm{E}}^{z}_{j,\xi} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi \,:\, \overline{\mathrm{E}}^{z}_{j,\xi} \text{ is defined} \wedge \lvert \{ t \in \mathcal{T} \,:\, \mathrm{on}^{z}_{j,t} \} \rvert > 0
\]
Process-e_sum_min#
Process_e_sum_min
Process_e_sum_min:
description: "`Process-e_sum_min` — energy of the internal power over the horizon, each period weighted by its years, is at least its floor. A floor of minus infinity is no row, and so is a process active in no snapshot"
dims: [scenario, process]
where: Process_e_sum_min AND count(Process_active, over=snapshot) > 0
expression: sum(Process_p * energy_limit_weight, over=snapshot) >= Process_e_sum_min
\[
\sum_{t \in \mathcal{T}} z_{j,\xi,t} \cdot \mathit{w}^{E}_{t} \ge \underline{\mathrm{E}}^{z}_{j,\xi} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi \,:\, \underline{\mathrm{E}}^{z}_{j,\xi} \text{ is defined} \wedge \lvert \{ t \in \mathcal{T} \,:\, \mathrm{on}^{z}_{j,t} \} \rvert > 0
\]
Process-ext-p-lower#
Process_ext_p_lower
Process_ext_p_lower:
description: "`Process-ext-p-lower` — an extendable process runs at least its minimum of the chosen build, negative for the other way"
dims: [scenario, snapshot, process]
where: Process_p_nom_extendable AND not Process_committable AND Process_active
expression: Process_p >= Process_p_min_pu * (Process_p_nom_ext - Process_maintenance_pu * Process_maintenance_capacity)
\[
z_{j,\xi,t} \ge \underline{\mathrm{z}}_{j,\xi,t} \cdot \left( Z_{j} - \gamma^{z}_{j,\xi} \cdot \mu^{z,\mathrm{nom}}_{j,\xi,t} \right) \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{ext}^{z}_{j} \wedge \neg \mathrm{com}^{z}_{j} \wedge \mathrm{on}^{z}_{j,t}
\]
Process-ext-p_nom-lower#
Process_ext_p_nom_lower
Process_ext_p_nom_lower:
description: "`Process-ext-p_nom-lower` — the chosen build is at least its floor in every scenario"
dims: [scenario, process]
where: Process_p_nom_extendable AND Process_active_asset
expression: Process_p_nom_ext >= Process_p_nom_min
\[
Z_{j} \ge \underline{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi \,:\, \mathrm{ext}^{z}_{j} \wedge \mathrm{act}^{z}_{j}
\]
Process-ext-p_nom-upper#
Process_ext_p_nom_upper
Process_ext_p_nom_upper:
description: "`Process-ext-p_nom-upper` — the chosen build is at most its cap in every scenario. A cap of infinity is no row"
dims: [scenario, process]
where: Process_p_nom_extendable AND Process_active_asset AND Process_p_nom_max
expression: Process_p_nom_ext <= Process_p_nom_max
\[
Z_{j} \le \overline{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi \,:\, \mathrm{ext}^{z}_{j} \wedge \mathrm{act}^{z}_{j} \wedge \overline{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} \text{ is defined}
\]
Process-ext-p-upper#
Process_ext_p_upper
Process_ext_p_upper:
description: "`Process-ext-p-upper` — an extendable process runs at most the chosen build"
dims: [scenario, snapshot, process]
where: Process_p_nom_extendable AND not Process_committable AND Process_active
expression: Process_p <= Process_p_max_pu * (Process_p_nom_ext - Process_maintenance_pu * Process_maintenance_capacity)
\[
z_{j,\xi,t} \le \overline{\mathrm{z}}_{j,\xi,t} \cdot \left( Z_{j} - \gamma^{z}_{j,\xi} \cdot \mu^{z,\mathrm{nom}}_{j,\xi,t} \right) \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{ext}^{z}_{j} \wedge \neg \mathrm{com}^{z}_{j} \wedge \mathrm{on}^{z}_{j,t}
\]
Process-fix-p-lower#
Process_fix_p_lower
Process_fix_p_lower:
description: "`Process-fix-p-lower` — a fixed process runs at least its minimum, negative for the other way"
dims: [scenario, snapshot, process]
where: not Process_p_nom_extendable AND not Process_committable AND Process_active
expression: Process_p >= Process_p_min_pu * Process_p_nom * (1 - Process_maintenance_pu * Process_maintenance)
\[
z_{j,\xi,t} \ge \underline{\mathrm{z}}_{j,\xi,t} \cdot \mathrm{z}^{\mathrm{nom}}_{j,\xi} \cdot \left( 1 - \gamma^{z}_{j,\xi} \cdot \mu^{z}_{j,\xi,t} \right) \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \neg \mathrm{ext}^{z}_{j} \wedge \neg \mathrm{com}^{z}_{j} \wedge \mathrm{on}^{z}_{j,t}
\]
Process-fix-p-upper#
Process_fix_p_upper
Process_fix_p_upper:
description: "`Process-fix-p-upper` — a fixed process runs at most its nominal power"
dims: [scenario, snapshot, process]
where: not Process_p_nom_extendable AND not Process_committable AND Process_active
expression: Process_p <= Process_p_max_pu * Process_p_nom * (1 - Process_maintenance_pu * Process_maintenance)
\[
z_{j,\xi,t} \le \overline{\mathrm{z}}_{j,\xi,t} \cdot \mathrm{z}^{\mathrm{nom}}_{j,\xi} \cdot \left( 1 - \gamma^{z}_{j,\xi} \cdot \mu^{z}_{j,\xi,t} \right) \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \neg \mathrm{ext}^{z}_{j} \wedge \neg \mathrm{com}^{z}_{j} \wedge \mathrm{on}^{z}_{j,t}
\]
Process-maint-event-count#
Process_maint_event_count
Process_maint_event_count:
description: "`Process-maint-event-count` — a maintainable process holds its number of maintenance events over the horizon"
dims: [scenario, process]
where: Process_maintainable
expression: sum(Process_maintenance_start, over=snapshot) == Process_maintenance_events
\[
\sum_{t \in \mathcal{T}} \mu^{z,\mathrm{up}}_{j,\xi,t} = \mathrm{n}^{z,\mathrm{mnt}}_{j,\xi} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi \,:\, \mathrm{mnt}^{z}_{j}
\]
Process-maint-modstatus-lb#
Process_maint_modstatus_lb
Process_maint_modstatus_lb:
description: "`Process-maint-modstatus-lb` — in maintenance, every module on is on in maintenance"
dims: [scenario, snapshot, process]
where: Process_maintainable AND Process_committed AND Process_p_nom_mod > 0
expression: Process_maintenance_status >= Process_status - Process_maintenance_modules * (1 - Process_maintenance)
\[
\mu^{z,u}_{j,\xi,t} \ge u^{z}_{j,\xi,t} - \mathrm{Process\_maintenance\_modules}_{j,\xi} \cdot \left( 1 - \mu^{z}_{j,\xi,t} \right) \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{z}_{j} \wedge \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0
\]
Process-maint-modstatus-le-maint#
Process_maint_modstatus_le_maint
Process_maint_modstatus_le_maint:
description: >-
`Process-maint-modstatus-le-maint` — out of maintenance, no module is on
in maintenance. In it, at most its modules in maintenance
dims: [scenario, snapshot, process]
where: Process_maintainable AND Process_committed AND Process_p_nom_mod > 0
expression: Process_maintenance_status <= Process_maintenance_modules * Process_maintenance
\[
\mu^{z,u}_{j,\xi,t} \le \mathrm{Process\_maintenance\_modules}_{j,\xi} \cdot \mu^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{z}_{j} \wedge \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0
\]
Process-maint-modstatus-le-status#
Process_maint_modstatus_le_status
Process_maint_modstatus_le_status:
description: "`Process-maint-modstatus-le-status` — the modules on in maintenance are at most the modules on"
dims: [scenario, snapshot, process]
where: Process_maintainable AND Process_committed AND Process_p_nom_mod > 0
expression: Process_maintenance_status <= Process_status
\[
\mu^{z,u}_{j,\xi,t} \le u^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{z}_{j} \wedge \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0
\]
Process-maint-start-horizon#
Process_maint_start_horizon
Process_maint_start_horizon:
description: "`Process-maint-start-horizon` — no event starts where it could not run its whole duration"
dims: [scenario, snapshot, process]
where: Process_maintainable AND Process_active AND Process_maintenance_start_blocked
expression: Process_maintenance_start == 0
\[
\mu^{z,\mathrm{up}}_{j,\xi,t} = 0 \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{z}_{j} \wedge \mathrm{on}^{z}_{j,t} \wedge \mathrm{blk}^{z}_{j,\xi,t}
\]
Process-maint-status-lb#
Process_maint_status_lb
Process_maint_status_lb:
description: "`Process-maint-status-lb` — on and in maintenance, the status in maintenance is one"
dims: [scenario, snapshot, process]
where: Process_maintainable AND Process_committed AND NOT Process_p_nom_extendable AND NOT (Process_p_nom_mod > 0)
expression: Process_maintenance_status >= Process_status + Process_maintenance - 1
\[
\mu^{z,u}_{j,\xi,t} \ge u^{z}_{j,\xi,t} + \mu^{z}_{j,\xi,t} - 1 \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{z}_{j} \wedge \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \neg \mathrm{ext}^{z}_{j} \wedge \neg \left( \mathrm{z}^{\mathrm{mod}}_{j} > 0 \right)
\]
Process-maint-status-le-maint#
Process_maint_status_le_maint
Process_maint_status_le_maint:
description: "`Process-maint-status-le-maint` — out of maintenance, the status in maintenance is zero"
dims: [scenario, snapshot, process]
where: Process_maintainable AND Process_committed AND NOT Process_p_nom_extendable AND NOT (Process_p_nom_mod > 0)
expression: Process_maintenance_status <= Process_maintenance
\[
\mu^{z,u}_{j,\xi,t} \le \mu^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{z}_{j} \wedge \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \neg \mathrm{ext}^{z}_{j} \wedge \neg \left( \mathrm{z}^{\mathrm{mod}}_{j} > 0 \right)
\]
Process-maint-status-le-status#
Process_maint_status_le_status
Process_maint_status_le_status:
description: "`Process-maint-status-le-status` — the status in maintenance is at most the status"
dims: [scenario, snapshot, process]
where: Process_maintainable AND Process_committed AND NOT Process_p_nom_extendable AND NOT (Process_p_nom_mod > 0)
expression: Process_maintenance_status <= Process_status
\[
\mu^{z,u}_{j,\xi,t} \le u^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{z}_{j} \wedge \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \neg \mathrm{ext}^{z}_{j} \wedge \neg \left( \mathrm{z}^{\mathrm{mod}}_{j} > 0 \right)
\]
Process-maint-window#
Process_maint_window
Process_maint_window:
description: >-
`Process-maint-window` — a process is in maintenance exactly where an
event it started covers the snapshot. Two events do not overlap, since
the maintenance status is at most one
dims: [scenario, snapshot, process]
where: Process_maintainable AND Process_active
expression: Process_maintenance == sum(Process_maintenance_start, over=start, by=Process_maintenance_cover[covered])
\[
\mu^{z}_{j,\xi,t} = \sum_{t' \in \mathcal{T} \,:\, \left( t,\ j,\ \xi,\ t' \right) \in \mathrm{Process\_maintenance\_cover}} \mu^{z,\mathrm{up}}_{j,\xi,t'} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{mnt}^{z}_{j} \wedge \mathrm{on}^{z}_{j,t}
\]
Process-maintcap_lower_nommax#
Process_maintcap_lower_nommax
Process_maintcap_lower_nommax:
description: "`Process-maintcap_lower_nommax` — in maintenance, the whole chosen build is taken off"
dims: [scenario, snapshot, process]
where: Process_maint_ext
expression: Process_maintenance_capacity >= Process_p_nom_ext - Process_p_nom_max * (1 - Process_maintenance)
\[
\mu^{z,\mathrm{nom}}_{j,\xi,t} \ge Z_{j} - \overline{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} \cdot \left( 1 - \mu^{z}_{j,\xi,t} \right) \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{mnt,ext}}_{j,t}
\]
Process-maintcap_lower_nommin#
Process_maintcap_lower_nommin
Process_maintcap_lower_nommin:
description: "`Process-maintcap_lower_nommin` — in maintenance, at least the floor of the build is taken off"
dims: [scenario, snapshot, process]
where: Process_maint_ext AND Process_p_nom_min > 0
expression: Process_maintenance_capacity >= Process_p_nom_min * Process_maintenance
\[
\mu^{z,\mathrm{nom}}_{j,\xi,t} \ge \underline{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} \cdot \mu^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{mnt,ext}}_{j,t} \wedge \underline{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} > 0
\]
Process-maintcap_upper#
Process_maintcap_upper
Process_maintcap_upper:
description: >-
`Process-maintcap_upper` — the build taken off is at most the chosen build in
maintenance, and at most the build less its floor out of it
dims: [scenario, snapshot, process]
where: Process_maint_ext
expression: Process_maintenance_capacity <= Process_p_nom_ext - Process_p_nom_min * (1 - Process_maintenance)
\[
\mu^{z,\mathrm{nom}}_{j,\xi,t} \le Z_{j} - \underline{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} \cdot \left( 1 - \mu^{z}_{j,\xi,t} \right) \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{mnt,ext}}_{j,t}
\]
Process-maintcap_upper_nommax#
Process_maintcap_upper_nommax
Process_maintcap_upper_nommax:
description: "`Process-maintcap_upper_nommax` — out of maintenance, no build is taken off"
dims: [scenario, snapshot, process]
where: Process_maint_ext
expression: Process_maintenance_capacity <= Process_p_nom_max * Process_maintenance
\[
\mu^{z,\mathrm{nom}}_{j,\xi,t} \le \overline{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} \cdot \mu^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{mnt,ext}}_{j,t}
\]
Process-p_nom_modularity#
Process_p_nom_modularity
Process_p_nom_modularity:
description: "`Process-p_nom_modularity` — the chosen build is a whole number of modules"
dims: [process]
where: Process_p_nom_extendable AND Process_p_nom_mod > 0 AND Process_active_asset
expression: Process_p_nom_ext == Process_p_nom_mod * Process_n_mod
\[
Z_{j} = \mathrm{z}^{\mathrm{mod}}_{j} \cdot N^{z}_{j} \qquad \forall\, j \in \mathcal{J} \,:\, \mathrm{ext}^{z}_{j} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0 \wedge \mathrm{act}^{z}_{j}
\]
Process-p_nom_set#
Process_p_nom_set
Process_p_nom_set:
description: "`Process-p_nom_set` — the chosen build pinned, wherever a value is given"
dims: [process]
where: Process_p_nom_extendable AND Process_active_asset AND Process_p_nom_set
expression: Process_p_nom_ext == Process_p_nom_set
\[
Z_{j} = \mathrm{z}^{\mathrm{nom,set}}_{j} \qquad \forall\, j \in \mathcal{J} \,:\, \mathrm{ext}^{z}_{j} \wedge \mathrm{act}^{z}_{j} \wedge \mathrm{z}^{\mathrm{nom,set}}_{j} \text{ is defined}
\]
Process-p-ramp_limit_down#
Process_p_ramp_limit_down
Process_p_ramp_limit_down:
description: >-
`Process-p-ramp_limit_down` — a process lowers internal power no faster
than its ramp limit of the build. A committed one lowers it no further
than its shut-down ramp in the snapshot it turns off. A process that
came into the horizon running has a row at the first snapshot only where
`p_init` gives the internal power it brought in. No process has one at
the start of a later investment period, nor does any process a big M
releases instead
dims: [scenario, snapshot, process]
where: >-
(Process_ramp_limit_down OR Process_ramp_limit_shut_down)
AND NOT (Process_committable AND Process_p_nom_extendable AND NOT (Process_p_nom_mod > 0))
AND Process_ramps_from_previous
AND Process_active
expression: Process_previous_p - Process_p <= Process_ramp_down_allowance
\[
\overleftarrow{z}_{j,\xi,t} - z_{j,\xi,t} \le \Delta^{z,-}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \left( \mathrm{rd}^{z}_{j,\xi,t} \text{ is defined} \vee \mathrm{rd}^{z,\mathrm{dn}}_{j,\xi} \text{ is defined} \right) \wedge \neg \left( \mathrm{com}^{z}_{j} \wedge \mathrm{ext}^{z}_{j} \wedge \neg \left( \mathrm{z}^{\mathrm{mod}}_{j} > 0 \right) \right) \wedge \mathrm{prev}^{z}_{j,\xi,t} \wedge \mathrm{on}^{z}_{j,t}
\]
Process-p-ramp_limit_down-run-bigM#
Process_p_ramp_limit_down_run_big_m
Process_p_ramp_limit_down_run_big_m:
description: >-
`Process-p-ramp_limit_down-run-bigM` — a committed extendable process
lowers internal power no faster than its limit of the chosen build. The
big M releases the row in the snapshot it turns off
dims: [scenario, snapshot, process]
where: >-
Process_com_ext
AND (Process_ramp_limit_down OR Process_ramp_limit_shut_down)
AND Process_ramps_from_previous
expression: >-
Process_p - Process_previous_p >=
-Process_ramp_down_rate * Process_p_nom_ext
- Process_big_m + Process_big_m * Process_status
\[
z_{j,\xi,t} - \overleftarrow{z}_{j,\xi,t} \ge -\widetilde{\mathrm{rd}}^{z}_{j,\xi,t} \cdot Z_{j} - \mathrm{M}^{z}_{j,\xi} + \mathrm{M}^{z}_{j,\xi} \cdot u^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com,ext}}_{j,t} \wedge \left( \mathrm{rd}^{z}_{j,\xi,t} \text{ is defined} \vee \mathrm{rd}^{z,\mathrm{dn}}_{j,\xi} \text{ is defined} \right) \wedge \mathrm{prev}^{z}_{j,\xi,t}
\]
Process-p-ramp_limit_down-shut-bigM#
Process_p_ramp_limit_down_shut_big_m
Process_p_ramp_limit_down_shut_big_m:
description: >-
`Process-p-ramp_limit_down-shut-bigM` — in the snapshot it turns off, a
committed extendable process ramps no further than its shut-down ramp of
the chosen build. The big M releases the row everywhere else
dims: [scenario, snapshot, process]
where: >-
Process_com_ext
AND (Process_ramp_limit_down OR Process_ramp_limit_shut_down)
AND Process_ramps_from_previous
expression: >-
Process_previous_p - Process_p <=
Process_shut_down_rate * Process_p_nom_ext
+ Process_big_m - Process_big_m * Process_shut_down
\[
\overleftarrow{z}_{j,\xi,t} - z_{j,\xi,t} \le \widetilde{\mathrm{rd}}^{z,\mathrm{dn}}_{j,\xi} \cdot Z_{j} + \mathrm{M}^{z}_{j,\xi} - \mathrm{M}^{z}_{j,\xi} \cdot \mathit{dn}^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com,ext}}_{j,t} \wedge \left( \mathrm{rd}^{z}_{j,\xi,t} \text{ is defined} \vee \mathrm{rd}^{z,\mathrm{dn}}_{j,\xi} \text{ is defined} \right) \wedge \mathrm{prev}^{z}_{j,\xi,t}
\]
Process-p-ramp_limit_up#
Process_p_ramp_limit_up
Process_p_ramp_limit_up:
description: >-
`Process-p-ramp_limit_up` — a process raises internal power no faster
than its ramp limit of the build. A committed one raises it no further
than its start-up ramp in the snapshot it turns on. A process that came
into the horizon running has a row at the first snapshot only where
`p_init` gives the internal power it brought in. No process has one at
the start of a later investment period, nor does any process a big M
releases instead
dims: [scenario, snapshot, process]
where: >-
(Process_ramp_limit_up OR Process_ramp_limit_start_up)
AND NOT (Process_committable AND Process_p_nom_extendable AND NOT (Process_p_nom_mod > 0))
AND Process_ramps_from_previous
AND Process_active
expression: Process_p - Process_previous_p <= Process_ramp_up_allowance
\[
z_{j,\xi,t} - \overleftarrow{z}_{j,\xi,t} \le \Delta^{z,+}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \left( \mathrm{ru}^{z}_{j,\xi,t} \text{ is defined} \vee \mathrm{ru}^{z,\mathrm{up}}_{j,\xi} \text{ is defined} \right) \wedge \neg \left( \mathrm{com}^{z}_{j} \wedge \mathrm{ext}^{z}_{j} \wedge \neg \left( \mathrm{z}^{\mathrm{mod}}_{j} > 0 \right) \right) \wedge \mathrm{prev}^{z}_{j,\xi,t} \wedge \mathrm{on}^{z}_{j,t}
\]
Process-p-ramp_limit_up-run-bigM#
Process_p_ramp_limit_up_run_big_m
Process_p_ramp_limit_up_run_big_m:
description: >-
`Process-p-ramp_limit_up-run-bigM` — a committed extendable process
raises internal power no faster than its limit of the chosen build. The
big M releases the row in the snapshot it turns on
dims: [scenario, snapshot, process]
where: Process_com_ext AND (Process_ramp_limit_up OR Process_ramp_limit_start_up) AND Process_ramps_from_previous
expression: >-
Process_p - Process_previous_p <=
Process_ramp_up_rate * Process_p_nom_ext
+ Process_big_m - Process_big_m * Process_previous_status
\[
z_{j,\xi,t} - \overleftarrow{z}_{j,\xi,t} \le \widetilde{\mathrm{ru}}^{z}_{j,\xi,t} \cdot Z_{j} + \mathrm{M}^{z}_{j,\xi} - \mathrm{M}^{z}_{j,\xi} \cdot \overleftarrow{u}^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com,ext}}_{j,t} \wedge \left( \mathrm{ru}^{z}_{j,\xi,t} \text{ is defined} \vee \mathrm{ru}^{z,\mathrm{up}}_{j,\xi} \text{ is defined} \right) \wedge \mathrm{prev}^{z}_{j,\xi,t}
\]
Process-p-ramp_limit_up-start-bigM#
Process_p_ramp_limit_up_start_big_m
Process_p_ramp_limit_up_start_big_m:
description: >-
`Process-p-ramp_limit_up-start-bigM` — in the snapshot it turns on, a
committed extendable process ramps no further than its start-up ramp of
the chosen build. The big M releases the row everywhere else
dims: [scenario, snapshot, process]
where: Process_com_ext AND (Process_ramp_limit_up OR Process_ramp_limit_start_up) AND Process_ramps_from_previous
expression: >-
Process_p - Process_previous_p <=
Process_start_up_rate * Process_p_nom_ext
+ Process_big_m - Process_big_m * Process_start_up
\[
z_{j,\xi,t} - \overleftarrow{z}_{j,\xi,t} \le \widetilde{\mathrm{ru}}^{z,\mathrm{up}}_{j,\xi} \cdot Z_{j} + \mathrm{M}^{z}_{j,\xi} - \mathrm{M}^{z}_{j,\xi} \cdot \mathit{up}^{z}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com,ext}}_{j,t} \wedge \left( \mathrm{ru}^{z}_{j,\xi,t} \text{ is defined} \vee \mathrm{ru}^{z,\mathrm{up}}_{j,\xi} \text{ is defined} \right) \wedge \mathrm{prev}^{z}_{j,\xi,t}
\]
Process-p_set#
Process_p_set
Process_p_set:
description: "`Process-p_set` — internal power pinned to the given schedule, wherever one is given"
dims: [scenario, snapshot, process]
where: Process_p_set AND Process_active
expression: Process_p == Process_p_set
\[
z_{j,\xi,t} = \mathrm{z}^{\mathrm{set}}_{j,\xi,t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{z}^{\mathrm{set}}_{j,\xi,t} \text{ is defined} \wedge \mathrm{on}^{z}_{j,t}
\]
Process-shut_down-p-fixed-upper#
Process_shut_down_p_fixed_upper
Process_shut_down_p_fixed_upper:
description: >-
`Process-shut_down-p-fixed-upper` — a stop is at most the modules in
place, an explicit row as PyPSA writes it. The bound is one where the
build is not modular, and the fixed build's whole count of modules where
it is
dims: [scenario, snapshot, process]
where: Process_committed AND NOT (Process_p_nom_extendable AND Process_p_nom_mod > 0)
expression: Process_shut_down <= Process_modules_installed
\[
\mathit{dn}^{z}_{j,\xi,t} \le \mathrm{N}^{z,\mathrm{fix}}_{j,\xi} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \neg \left( \mathrm{ext}^{z}_{j} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0 \right)
\]
Process-shut_down-p_nom-variable-upper#
Process_shut_down_p_nom_variable_upper
Process_shut_down_p_nom_variable_upper:
description: "`Process-shut_down-p_nom-variable-upper` — a modular process stops only where a module is built"
dims: [scenario, snapshot, process]
where: Process_committed AND Process_p_nom_extendable AND Process_p_nom_mod > 0
expression: Process_shut_down <= Process_n_mod
\[
\mathit{dn}^{z}_{j,\xi,t} \le N^{z}_{j} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \mathrm{ext}^{z}_{j} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0
\]
Process-start_up-p-fixed-upper#
Process_start_up_p_fixed_upper
Process_start_up_p_fixed_upper:
description: >-
`Process-start_up-p-fixed-upper` — a start is at most the modules in
place, an explicit row as PyPSA writes it. The bound is one where the
build is not modular, and the fixed build's whole count of modules where
it is
dims: [scenario, snapshot, process]
where: Process_committed AND NOT (Process_p_nom_extendable AND Process_p_nom_mod > 0)
expression: Process_start_up <= Process_modules_installed
\[
\mathit{up}^{z}_{j,\xi,t} \le \mathrm{N}^{z,\mathrm{fix}}_{j,\xi} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \neg \left( \mathrm{ext}^{z}_{j} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0 \right)
\]
Process-start_up-p_nom-variable-upper#
Process_start_up_p_nom_variable_upper
Process_start_up_p_nom_variable_upper:
description: "`Process-start_up-p_nom-variable-upper` — a modular process starts only where a module is built"
dims: [scenario, snapshot, process]
where: Process_committed AND Process_p_nom_extendable AND Process_p_nom_mod > 0
expression: Process_start_up <= Process_n_mod
\[
\mathit{up}^{z}_{j,\xi,t} \le N^{z}_{j} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \mathrm{ext}^{z}_{j} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0
\]
Process-status-p-fixed-upper#
Process_status_p_fixed_upper
Process_status_p_fixed_upper:
description: >-
`Process-status-p-fixed-upper` — a status is at most the modules in
place, an explicit row as PyPSA writes it. The bound is one where the
build is not modular, and the fixed build's whole count of modules where
it is
dims: [scenario, snapshot, process]
where: Process_committed AND NOT (Process_p_nom_extendable AND Process_p_nom_mod > 0)
expression: Process_status <= Process_modules_installed
\[
u^{z}_{j,\xi,t} \le \mathrm{N}^{z,\mathrm{fix}}_{j,\xi} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \neg \left( \mathrm{ext}^{z}_{j} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0 \right)
\]
Process-status-p_nom-variable-upper#
Process_status_p_nom_variable_upper
Process_status_p_nom_variable_upper:
description: "`Process-status-p_nom-variable-upper` — a modular process is on only where a module is built"
dims: [scenario, snapshot, process]
where: Process_committed AND Process_p_nom_extendable AND Process_p_nom_mod > 0
expression: Process_status <= Process_n_mod
\[
u^{z}_{j,\xi,t} \le N^{z}_{j} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com}}_{j,t} \wedge \mathrm{ext}^{z}_{j} \wedge \mathrm{z}^{\mathrm{mod}}_{j} > 0
\]
StorageUnit-energy_balance#
StorageUnit_energy_balance
StorageUnit_energy_balance:
description: >-
`StorageUnit-energy_balance` — the charge carried in, plus what is
stored after its efficiency, less what dispatch draws down before its
own, plus inflow not spilled
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_active
expression: >-
StorageUnit_charge_carried_in
+ StorageUnit_efficiency_store * StorageUnit_p_store * snapshot_weightings_stores
- StorageUnit_p_dispatch * snapshot_weightings_stores / StorageUnit_efficiency_dispatch
+ (StorageUnit_inflow - StorageUnit_spill) * snapshot_weightings_stores
== StorageUnit_state_of_charge
\[
\overleftarrow{\mathit{soc}}_{\xi,t,s} + \eta^{-}_{\xi,t,s} \cdot h^{-}_{\xi,t,s} \cdot \mathrm{w}^{\mathrm{sto}}_{t} - \frac{h^{+}_{\xi,t,s} \cdot \mathrm{w}^{\mathrm{sto}}_{t}}{\eta^{+}_{\xi,t,s}} + \left( \mathrm{inflow}_{\xi,t,s} - \mathit{spill}_{\xi,t,s} \right) \cdot \mathrm{w}^{\mathrm{sto}}_{t} = \mathit{soc}_{\xi,t,s} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h}_{t,s}
\]
StorageUnit-ext-p_dispatch-lower#
StorageUnit_ext_p_dispatch_lower
StorageUnit_ext_p_dispatch_lower:
description: "`StorageUnit-ext-p_dispatch-lower` — dispatch is non-negative"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_ext
expression: StorageUnit_p_dispatch >= 0
\[
h^{+}_{\xi,t,s} \ge 0 \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h,\mathrm{ext}}_{t,s}
\]
StorageUnit-ext-p_dispatch-upper#
StorageUnit_ext_p_dispatch_upper
StorageUnit_ext_p_dispatch_upper:
description: "`StorageUnit-ext-p_dispatch-upper` — an extendable unit dispatches at most the chosen build"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_ext
expression: StorageUnit_p_dispatch <= StorageUnit_p_max_pu * StorageUnit_p_nom_ext
\[
h^{+}_{\xi,t,s} \le \overline{\mathrm{h}}_{\xi,t,s} \cdot H_{s} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h,\mathrm{ext}}_{t,s}
\]
StorageUnit-ext-p_nom-lower#
StorageUnit_ext_p_nom_lower
StorageUnit_ext_p_nom_lower:
description: "`StorageUnit-ext-p_nom-lower` — the chosen build is at least its floor in every scenario"
dims: [scenario, storage_unit]
where: StorageUnit_p_nom_extendable AND StorageUnit_active_asset
expression: StorageUnit_p_nom_ext >= StorageUnit_p_nom_min
\[
H_{s} \ge \underline{\mathrm{h}}^{\mathrm{nom}}_{\xi,s} \qquad \forall\, \xi \in \Xi,\ s \in \mathcal{S} \,:\, \mathrm{ext}^{h}_{s} \wedge \mathrm{act}^{h}_{s}
\]
StorageUnit-ext-p_nom-upper#
StorageUnit_ext_p_nom_upper
StorageUnit_ext_p_nom_upper:
description: "`StorageUnit-ext-p_nom-upper` — the chosen build is at most its cap in every scenario. A cap of infinity is no row"
dims: [scenario, storage_unit]
where: StorageUnit_p_nom_extendable AND StorageUnit_active_asset AND StorageUnit_p_nom_max
expression: StorageUnit_p_nom_ext <= StorageUnit_p_nom_max
\[
H_{s} \le \overline{\mathrm{h}}^{\mathrm{nom}}_{\xi,s} \qquad \forall\, \xi \in \Xi,\ s \in \mathcal{S} \,:\, \mathrm{ext}^{h}_{s} \wedge \mathrm{act}^{h}_{s} \wedge \overline{\mathrm{h}}^{\mathrm{nom}}_{\xi,s} \text{ is defined}
\]
StorageUnit-ext-p_store-lower#
StorageUnit_ext_p_store_lower
StorageUnit_ext_p_store_lower:
description: "`StorageUnit-ext-p_store-lower` — storing is non-negative"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_ext
expression: StorageUnit_p_store >= 0
\[
h^{-}_{\xi,t,s} \ge 0 \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h,\mathrm{ext}}_{t,s}
\]
StorageUnit-ext-p_store-upper#
StorageUnit_ext_p_store_upper
StorageUnit_ext_p_store_upper:
description: >-
`StorageUnit-ext-p_store-upper` — an extendable unit stores at most the
chosen build, the minimum-per-unit column carrying that cap negated
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_ext
expression: StorageUnit_p_store <= -StorageUnit_p_min_pu * StorageUnit_p_nom_ext
\[
h^{-}_{\xi,t,s} \le -\underline{\mathrm{h}}_{\xi,t,s} \cdot H_{s} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h,\mathrm{ext}}_{t,s}
\]
StorageUnit-ext-state_of_charge-lower#
StorageUnit_ext_state_of_charge_lower
StorageUnit_ext_state_of_charge_lower:
description: "`StorageUnit-ext-state_of_charge-lower` — charge is non-negative"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_ext
expression: StorageUnit_state_of_charge >= 0
\[
\mathit{soc}_{\xi,t,s} \ge 0 \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h,\mathrm{ext}}_{t,s}
\]
StorageUnit-ext-state_of_charge-upper#
StorageUnit_ext_state_of_charge_upper
StorageUnit_ext_state_of_charge_upper:
description: "`StorageUnit-ext-state_of_charge-upper` — an extendable unit holds at most its hours at the chosen build"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_ext
expression: StorageUnit_state_of_charge <= StorageUnit_max_hours * StorageUnit_p_nom_ext
\[
\mathit{soc}_{\xi,t,s} \le \mathrm{T}^{h}_{\xi,s} \cdot H_{s} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h,\mathrm{ext}}_{t,s}
\]
StorageUnit-fix-p_dispatch-lower#
StorageUnit_fix_p_dispatch_lower
StorageUnit_fix_p_dispatch_lower:
description: "`StorageUnit-fix-p_dispatch-lower` — dispatch is non-negative"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_fix
expression: StorageUnit_p_dispatch >= 0
\[
h^{+}_{\xi,t,s} \ge 0 \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h,\mathrm{fix}}_{t,s}
\]
StorageUnit-fix-p_dispatch-upper#
StorageUnit_fix_p_dispatch_upper
StorageUnit_fix_p_dispatch_upper:
description: "`StorageUnit-fix-p_dispatch-upper` — a fixed unit dispatches at most its nominal power"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_fix
expression: StorageUnit_p_dispatch <= StorageUnit_p_max_pu * StorageUnit_p_nom
\[
h^{+}_{\xi,t,s} \le \overline{\mathrm{h}}_{\xi,t,s} \cdot \mathrm{h}^{\mathrm{nom}}_{\xi,s} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h,\mathrm{fix}}_{t,s}
\]
StorageUnit-fix-p_store-lower#
StorageUnit_fix_p_store_lower
StorageUnit_fix_p_store_lower:
description: "`StorageUnit-fix-p_store-lower` — storing is non-negative"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_fix
expression: StorageUnit_p_store >= 0
\[
h^{-}_{\xi,t,s} \ge 0 \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h,\mathrm{fix}}_{t,s}
\]
StorageUnit-fix-p_store-upper#
StorageUnit_fix_p_store_upper
StorageUnit_fix_p_store_upper:
description: >-
`StorageUnit-fix-p_store-upper` — a fixed unit stores at most its
nominal power, the minimum-per-unit column carrying that cap negated
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_fix
expression: StorageUnit_p_store <= -StorageUnit_p_min_pu * StorageUnit_p_nom
\[
h^{-}_{\xi,t,s} \le -\underline{\mathrm{h}}_{\xi,t,s} \cdot \mathrm{h}^{\mathrm{nom}}_{\xi,s} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h,\mathrm{fix}}_{t,s}
\]
StorageUnit-fix-state_of_charge-lower#
StorageUnit_fix_state_of_charge_lower
StorageUnit_fix_state_of_charge_lower:
description: "`StorageUnit-fix-state_of_charge-lower` — charge is non-negative"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_fix
expression: StorageUnit_state_of_charge >= 0
\[
\mathit{soc}_{\xi,t,s} \ge 0 \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h,\mathrm{fix}}_{t,s}
\]
StorageUnit-fix-state_of_charge-upper#
StorageUnit_fix_state_of_charge_upper
StorageUnit_fix_state_of_charge_upper:
description: "`StorageUnit-fix-state_of_charge-upper` — a fixed unit holds at most its hours at nominal power"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_fix
expression: StorageUnit_state_of_charge <= StorageUnit_max_hours * StorageUnit_p_nom
\[
\mathit{soc}_{\xi,t,s} \le \mathrm{T}^{h}_{\xi,s} \cdot \mathrm{h}^{\mathrm{nom}}_{\xi,s} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h,\mathrm{fix}}_{t,s}
\]
StorageUnit-p_dispatch_set#
StorageUnit_p_dispatch_set
StorageUnit_p_dispatch_set:
description: "`StorageUnit-p_dispatch_set` — dispatch pinned to the given schedule, wherever one is given"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_p_dispatch_set AND StorageUnit_active
expression: StorageUnit_p_dispatch == StorageUnit_p_dispatch_set
\[
h^{+}_{\xi,t,s} = \mathrm{h}^{+,\mathrm{set}}_{\xi,t,s} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{h}^{+,\mathrm{set}}_{\xi,t,s} \text{ is defined} \wedge \mathrm{on}^{h}_{t,s}
\]
StorageUnit-p_nom_modularity#
StorageUnit_p_nom_modularity
StorageUnit_p_nom_modularity:
description: "`StorageUnit-p_nom_modularity` — the chosen build is a whole number of modules"
dims: [storage_unit]
where: StorageUnit_p_nom_extendable AND StorageUnit_p_nom_mod > 0 AND StorageUnit_active_asset
expression: StorageUnit_p_nom_ext == StorageUnit_p_nom_mod * StorageUnit_n_mod
\[
H_{s} = \mathrm{h}^{\mathrm{mod}}_{s} \cdot N^{h}_{s} \qquad \forall\, s \in \mathcal{S} \,:\, \mathrm{ext}^{h}_{s} \wedge \mathrm{h}^{\mathrm{mod}}_{s} > 0 \wedge \mathrm{act}^{h}_{s}
\]
StorageUnit-p_nom_set#
StorageUnit_p_nom_set
StorageUnit_p_nom_set:
description: "`StorageUnit-p_nom_set` — the chosen build pinned, wherever a value is given"
dims: [storage_unit]
where: StorageUnit_p_nom_extendable AND StorageUnit_active_asset AND StorageUnit_p_nom_set
expression: StorageUnit_p_nom_ext == StorageUnit_p_nom_set
\[
H_{s} = \mathrm{h}^{\mathrm{nom,set}}_{s} \qquad \forall\, s \in \mathcal{S} \,:\, \mathrm{ext}^{h}_{s} \wedge \mathrm{act}^{h}_{s} \wedge \mathrm{h}^{\mathrm{nom,set}}_{s} \text{ is defined}
\]
StorageUnit-p_set#
StorageUnit_p_set
StorageUnit_p_set:
description: "`StorageUnit-p_set` — net dispatch pinned to the given schedule, wherever one is given"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_p_set AND StorageUnit_active
expression: StorageUnit_p_dispatch - StorageUnit_p_store == StorageUnit_p_set
\[
h^{+}_{\xi,t,s} - h^{-}_{\xi,t,s} = \mathrm{h}^{\mathrm{set}}_{\xi,t,s} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{h}^{\mathrm{set}}_{\xi,t,s} \text{ is defined} \wedge \mathrm{on}^{h}_{t,s}
\]
StorageUnit-p_store_set#
StorageUnit_p_store_set
StorageUnit_p_store_set:
description: "`StorageUnit-p_store_set` — charging pinned to the given schedule, wherever one is given"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_p_store_set AND StorageUnit_active
expression: StorageUnit_p_store == StorageUnit_p_store_set
\[
h^{-}_{\xi,t,s} = \mathrm{h}^{-,\mathrm{set}}_{\xi,t,s} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{h}^{-,\mathrm{set}}_{\xi,t,s} \text{ is defined} \wedge \mathrm{on}^{h}_{t,s}
\]
StorageUnit-state_of_charge_set#
StorageUnit_state_of_charge_set
StorageUnit_state_of_charge_set:
description: "`StorageUnit-state_of_charge_set` — charge pinned to the given schedule, wherever one is given"
dims: [scenario, snapshot, storage_unit]
where: StorageUnit_state_of_charge_set AND StorageUnit_active
expression: StorageUnit_state_of_charge == StorageUnit_state_of_charge_set
\[
\mathit{soc}_{\xi,t,s} = \mathrm{soc}^{\mathrm{set}}_{\xi,t,s} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{soc}^{\mathrm{set}}_{\xi,t,s} \text{ is defined} \wedge \mathrm{on}^{h}_{t,s}
\]
Store-cycles_max#
Store_cycles_max
Store_cycles_max:
description: >-
`Store-cycles_max` — the discharge over the horizon, `p + p_store`
weighted by the stores weighting, is at most `cycles_max` times the
build. A store without `cycles_max` has no row
(`constraints.py:2607-2633`)
dims: [scenario, store]
where: Store_cycles_max AND Store_active_asset
expression: sum((Store_p + Store_p_store) * snapshot_weightings_stores, over=snapshot) <= Store_cycles_max * Store_e_nom_effective
\[
\sum_{t \in \mathcal{T}} \left( q_{\xi,t,v} + q^{-}_{\xi,t,v} \right) \cdot \mathrm{w}^{\mathrm{sto}}_{t} \le \mathrm{cyc}^{\max}_{\xi,v} \cdot \widetilde{\mathrm{e}}^{\mathrm{nom}}_{\xi,v} \qquad \forall\, \xi \in \Xi,\ v \in \mathcal{V} \,:\, \mathrm{cyc}^{\max}_{\xi,v} \text{ is defined} \wedge \mathrm{act}^{e}_{v}
\]
Store-e_nom_modularity#
Store_e_nom_modularity
Store_e_nom_modularity:
description: "`Store-e_nom_modularity` — the chosen build is a whole number of modules"
dims: [store]
where: Store_e_nom_extendable AND Store_e_nom_mod > 0 AND Store_active_asset
expression: Store_e_nom_ext == Store_e_nom_mod * Store_n_mod
\[
E_{v} = \mathrm{e}^{\mathrm{mod}}_{v} \cdot N^{e}_{v} \qquad \forall\, v \in \mathcal{V} \,:\, \mathrm{ext}^{e}_{v} \wedge \mathrm{e}^{\mathrm{mod}}_{v} > 0 \wedge \mathrm{act}^{e}_{v}
\]
Store-e_nom_set#
Store_e_nom_set
Store_e_nom_set:
description: "`Store-e_nom_set` — the chosen build pinned, wherever a value is given"
dims: [store]
where: Store_e_nom_extendable AND Store_active_asset AND Store_e_nom_set
expression: Store_e_nom_ext == Store_e_nom_set
\[
E_{v} = \mathrm{e}^{\mathrm{nom,set}}_{v} \qquad \forall\, v \in \mathcal{V} \,:\, \mathrm{ext}^{e}_{v} \wedge \mathrm{act}^{e}_{v} \wedge \mathrm{e}^{\mathrm{nom,set}}_{v} \text{ is defined}
\]
Store-e_set#
Store_e_set
Store_e_set:
description: "`Store-e_set` — energy pinned to the given schedule, wherever one is given"
dims: [scenario, snapshot, store]
where: Store_e_set AND Store_active
expression: Store_e == Store_e_set
\[
e_{\xi,t,v} = \mathrm{e}^{\mathrm{set}}_{\xi,t,v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{e}^{\mathrm{set}}_{\xi,t,v} \text{ is defined} \wedge \mathrm{on}^{e}_{t,v}
\]
Store-e_sum_max#
Store_e_sum_max
Store_e_sum_max:
description: "`Store-e_sum_max` — net energy delivered over the horizon, each period weighted by its years, is at most its budget. A budget of infinity is no row, and so is a store active in no snapshot"
dims: [scenario, store]
where: Store_e_sum_max AND count(Store_active, over=snapshot) > 0
expression: sum(Store_p * energy_limit_weight, over=snapshot) <= Store_e_sum_max
\[
\sum_{t \in \mathcal{T}} q_{\xi,t,v} \cdot \mathit{w}^{E}_{t} \le \overline{\mathrm{E}}^{q}_{\xi,v} \qquad \forall\, \xi \in \Xi,\ v \in \mathcal{V} \,:\, \overline{\mathrm{E}}^{q}_{\xi,v} \text{ is defined} \wedge \lvert \{ t \in \mathcal{T} \,:\, \mathrm{on}^{e}_{t,v} \} \rvert > 0
\]
Store-e_sum_min#
Store_e_sum_min
Store_e_sum_min:
description: "`Store-e_sum_min` — net energy delivered over the horizon, each period weighted by its years, is at least its floor. A floor of minus infinity is no row, and so is a store active in no snapshot"
dims: [scenario, store]
where: Store_e_sum_min AND count(Store_active, over=snapshot) > 0
expression: sum(Store_p * energy_limit_weight, over=snapshot) >= Store_e_sum_min
\[
\sum_{t \in \mathcal{T}} q_{\xi,t,v} \cdot \mathit{w}^{E}_{t} \ge \underline{\mathrm{E}}^{q}_{\xi,v} \qquad \forall\, \xi \in \Xi,\ v \in \mathcal{V} \,:\, \underline{\mathrm{E}}^{q}_{\xi,v} \text{ is defined} \wedge \lvert \{ t \in \mathcal{T} \,:\, \mathrm{on}^{e}_{t,v} \} \rvert > 0
\]
Store-energy_balance#
Store_energy_balance
Store_energy_balance:
description: >-
`Store-energy_balance` — the energy carried in, less the discharge drawn
down before its efficiency, plus the charge after its own, plus inflow
not spilled. The net dispatch `p` is the discharge less `p_store`, so
`p_store` enters with both efficiencies (`constraints.py:2472-2476`)
dims: [scenario, snapshot, store]
where: Store_active
expression: >-
Store_energy_carried_in
- Store_p * snapshot_weightings_stores / Store_efficiency_dispatch
+ (Store_efficiency_store - 1 / Store_efficiency_dispatch) * Store_p_store * snapshot_weightings_stores
+ (Store_inflow - Store_spill) * snapshot_weightings_stores
== Store_e
\[
\overleftarrow{e}_{\xi,t,v} - \frac{q_{\xi,t,v} \cdot \mathrm{w}^{\mathrm{sto}}_{t}}{\eta^{e,+}_{\xi,t,v}} + \left( \eta^{e,-}_{\xi,t,v} - \frac{1}{\eta^{e,+}_{\xi,t,v}} \right) \cdot q^{-}_{\xi,t,v} \cdot \mathrm{w}^{\mathrm{sto}}_{t} + \left( \mathrm{inflow}^{e}_{\xi,t,v} - \mathit{spill}^{e}_{\xi,t,v} \right) \cdot \mathrm{w}^{\mathrm{sto}}_{t} = e_{\xi,t,v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{on}^{e}_{t,v}
\]
Store-ext-e-lower#
Store_ext_e_lower
Store_ext_e_lower:
description: "`Store-ext-e-lower` — an extendable store holds at least its floor of the chosen build"
dims: [scenario, snapshot, store]
where: Store_e_nom_extendable AND Store_active
expression: Store_e >= Store_e_min_pu * Store_e_nom_ext
\[
e_{\xi,t,v} \ge \underline{\mathrm{e}}_{\xi,t,v} \cdot E_{v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{ext}^{e}_{v} \wedge \mathrm{on}^{e}_{t,v}
\]
Store-ext-e_nom-lower#
Store_ext_e_nom_lower
Store_ext_e_nom_lower:
description: "`Store-ext-e_nom-lower` — the chosen build is at least its floor in every scenario"
dims: [scenario, store]
where: Store_e_nom_extendable AND Store_active_asset
expression: Store_e_nom_ext >= Store_e_nom_min
\[
E_{v} \ge \underline{\mathrm{e}}^{\mathrm{nom}}_{\xi,v} \qquad \forall\, \xi \in \Xi,\ v \in \mathcal{V} \,:\, \mathrm{ext}^{e}_{v} \wedge \mathrm{act}^{e}_{v}
\]
Store-ext-e_nom-upper#
Store_ext_e_nom_upper
Store_ext_e_nom_upper:
description: "`Store-ext-e_nom-upper` — the chosen build is at most its cap in every scenario. A cap of infinity is no row"
dims: [scenario, store]
where: Store_e_nom_extendable AND Store_active_asset AND Store_e_nom_max
expression: Store_e_nom_ext <= Store_e_nom_max
\[
E_{v} \le \overline{\mathrm{e}}^{\mathrm{nom}}_{\xi,v} \qquad \forall\, \xi \in \Xi,\ v \in \mathcal{V} \,:\, \mathrm{ext}^{e}_{v} \wedge \mathrm{act}^{e}_{v} \wedge \overline{\mathrm{e}}^{\mathrm{nom}}_{\xi,v} \text{ is defined}
\]
Store-ext-e-upper#
Store_ext_e_upper
Store_ext_e_upper:
description: "`Store-ext-e-upper` — an extendable store holds at most the chosen build"
dims: [scenario, snapshot, store]
where: Store_e_nom_extendable AND Store_active
expression: Store_e <= Store_e_max_pu * Store_e_nom_ext
\[
e_{\xi,t,v} \le \overline{\mathrm{e}}_{\xi,t,v} \cdot E_{v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{ext}^{e}_{v} \wedge \mathrm{on}^{e}_{t,v}
\]
Store-ext-p-lower#
Store_ext_p_lower
Store_ext_p_lower:
description: >-
`Store-ext-p-lower` — an extendable store discharges at least its floor, which is
zero where it splits its dispatch and its most charging otherwise. A
store that does not split and has no `max_hours` has no row
dims: [scenario, snapshot, store]
where: Store_e_nom_extendable AND Store_active AND (Store_split_dispatch OR Store_max_hours)
expression: Store_p + Store_p_store >= Store_discharge_floor_pu * Store_e_nom_ext
\[
q_{\xi,t,v} + q^{-}_{\xi,t,v} \ge \underline{\mathrm{q}}^{+}_{\xi,t,v} \cdot E_{v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{ext}^{e}_{v} \wedge \mathrm{on}^{e}_{t,v} \wedge \left( \mathrm{split}^{e}_{\xi,v} \vee \mathrm{T}^{e}_{\xi,v} \text{ is defined} \right)
\]
Store-ext-p_store-lower#
Store_ext_p_store_lower
Store_ext_p_store_lower:
description: "`Store-ext-p_store-lower` — charging is non-negative"
dims: [scenario, snapshot, store]
where: Store_e_nom_extendable AND Store_active
expression: Store_p_store >= 0
\[
q^{-}_{\xi,t,v} \ge 0 \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{ext}^{e}_{v} \wedge \mathrm{on}^{e}_{t,v}
\]
Store-ext-p_store-upper#
Store_ext_p_store_upper
Store_ext_p_store_upper:
description: >-
`Store-ext-p_store-upper` — an extendable store charges at most the power rating of the chosen build where it
splits its dispatch, and not at all where it does not. A store that splits
and has no `max_hours` has no row
dims: [scenario, snapshot, store]
where: Store_e_nom_extendable AND Store_active AND (NOT Store_split_dispatch OR Store_max_hours)
expression: Store_p_store <= Store_charge_cap_pu * Store_e_nom_ext
\[
q^{-}_{\xi,t,v} \le \overline{\mathrm{q}}^{-}_{\xi,t,v} \cdot E_{v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{ext}^{e}_{v} \wedge \mathrm{on}^{e}_{t,v} \wedge \left( \neg \mathrm{split}^{e}_{\xi,v} \vee \mathrm{T}^{e}_{\xi,v} \text{ is defined} \right)
\]
Store-ext-p-upper#
Store_ext_p_upper
Store_ext_p_upper:
description: "`Store-ext-p-upper` — an extendable store discharges at most the power rating of the chosen build. A store without `max_hours` has no row"
dims: [scenario, snapshot, store]
where: Store_e_nom_extendable AND Store_active AND Store_max_hours
expression: Store_p + Store_p_store <= Store_p_max_pu / Store_max_hours * Store_e_nom_ext
\[
q_{\xi,t,v} + q^{-}_{\xi,t,v} \le \frac{\overline{\mathrm{q}}_{\xi,t,v}}{\mathrm{T}^{e}_{\xi,v}} \cdot E_{v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{ext}^{e}_{v} \wedge \mathrm{on}^{e}_{t,v} \wedge \mathrm{T}^{e}_{\xi,v} \text{ is defined}
\]
Store-fix-e-lower#
Store_fix_e_lower
Store_fix_e_lower:
description: "`Store-fix-e-lower` — a fixed store holds at least its floor"
dims: [scenario, snapshot, store]
where: not Store_e_nom_extendable AND Store_active
expression: Store_e >= Store_e_min_pu * Store_e_nom
\[
e_{\xi,t,v} \ge \underline{\mathrm{e}}_{\xi,t,v} \cdot \mathrm{e}^{\mathrm{nom}}_{\xi,v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \neg \mathrm{ext}^{e}_{v} \wedge \mathrm{on}^{e}_{t,v}
\]
Store-fix-e-upper#
Store_fix_e_upper
Store_fix_e_upper:
description: "`Store-fix-e-upper` — a fixed store holds at most its nominal capacity"
dims: [scenario, snapshot, store]
where: not Store_e_nom_extendable AND Store_active
expression: Store_e <= Store_e_max_pu * Store_e_nom
\[
e_{\xi,t,v} \le \overline{\mathrm{e}}_{\xi,t,v} \cdot \mathrm{e}^{\mathrm{nom}}_{\xi,v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \neg \mathrm{ext}^{e}_{v} \wedge \mathrm{on}^{e}_{t,v}
\]
Store-fix-p-lower#
Store_fix_p_lower
Store_fix_p_lower:
description: >-
`Store-fix-p-lower` — a fixed store discharges at least its floor, which is
zero where it splits its dispatch and its most charging otherwise. A
store that does not split and has no `max_hours` has no row
dims: [scenario, snapshot, store]
where: NOT Store_e_nom_extendable AND Store_active AND (Store_split_dispatch OR Store_max_hours)
expression: Store_p + Store_p_store >= Store_discharge_floor_pu * Store_e_nom
\[
q_{\xi,t,v} + q^{-}_{\xi,t,v} \ge \underline{\mathrm{q}}^{+}_{\xi,t,v} \cdot \mathrm{e}^{\mathrm{nom}}_{\xi,v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \neg \mathrm{ext}^{e}_{v} \wedge \mathrm{on}^{e}_{t,v} \wedge \left( \mathrm{split}^{e}_{\xi,v} \vee \mathrm{T}^{e}_{\xi,v} \text{ is defined} \right)
\]
Store-fix-p_store-lower#
Store_fix_p_store_lower
Store_fix_p_store_lower:
description: "`Store-fix-p_store-lower` — charging is non-negative"
dims: [scenario, snapshot, store]
where: NOT Store_e_nom_extendable AND Store_active
expression: Store_p_store >= 0
\[
q^{-}_{\xi,t,v} \ge 0 \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \neg \mathrm{ext}^{e}_{v} \wedge \mathrm{on}^{e}_{t,v}
\]
Store-fix-p_store-upper#
Store_fix_p_store_upper
Store_fix_p_store_upper:
description: >-
`Store-fix-p_store-upper` — a fixed store charges at most its power rating where it
splits its dispatch, and not at all where it does not. A store that splits
and has no `max_hours` has no row
dims: [scenario, snapshot, store]
where: NOT Store_e_nom_extendable AND Store_active AND (NOT Store_split_dispatch OR Store_max_hours)
expression: Store_p_store <= Store_charge_cap_pu * Store_e_nom
\[
q^{-}_{\xi,t,v} \le \overline{\mathrm{q}}^{-}_{\xi,t,v} \cdot \mathrm{e}^{\mathrm{nom}}_{\xi,v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \neg \mathrm{ext}^{e}_{v} \wedge \mathrm{on}^{e}_{t,v} \wedge \left( \neg \mathrm{split}^{e}_{\xi,v} \vee \mathrm{T}^{e}_{\xi,v} \text{ is defined} \right)
\]
Store-fix-p-upper#
Store_fix_p_upper
Store_fix_p_upper:
description: "`Store-fix-p-upper` — a fixed store discharges at most its power rating. A store without `max_hours` has no row"
dims: [scenario, snapshot, store]
where: NOT Store_e_nom_extendable AND Store_active AND Store_max_hours
expression: Store_p + Store_p_store <= Store_p_max_pu / Store_max_hours * Store_e_nom
\[
q_{\xi,t,v} + q^{-}_{\xi,t,v} \le \frac{\overline{\mathrm{q}}_{\xi,t,v}}{\mathrm{T}^{e}_{\xi,v}} \cdot \mathrm{e}^{\mathrm{nom}}_{\xi,v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \neg \mathrm{ext}^{e}_{v} \wedge \mathrm{on}^{e}_{t,v} \wedge \mathrm{T}^{e}_{\xi,v} \text{ is defined}
\]
Store-p_set#
Store_p_set
Store_p_set:
description: "`Store-p_set` — power delivered pinned to the given schedule, wherever one is given"
dims: [scenario, snapshot, store]
where: Store_p_set AND Store_active
expression: Store_p == Store_p_set
\[
q_{\xi,t,v} = \mathrm{q}^{\mathrm{set}}_{\xi,t,v} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{q}^{\mathrm{set}}_{\xi,t,v} \text{ is defined} \wedge \mathrm{on}^{e}_{t,v}
\]
Transformer-ext-s-lower#
Transformer_ext_s_lower
Transformer_ext_s_lower:
description: "`Transformer-ext-s-lower` — an extendable transformer carries at least the negative of its rating of the chosen build, the loss counted against it"
dims: [scenario, snapshot, transformer]
where: Transformer_s_nom_extendable AND Transformer_active
expression: Transformer_s - Transformer_loss >= -Transformer_s_max_pu * Transformer_s_nom_ext
\[
\sigma_{\xi,t,m} - \ell^{\sigma}_{\xi,t,m} \ge -\overline{\sigma}_{\xi,t,m} \cdot \Sigma_{m} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \mathrm{ext}^{\sigma}_{m} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Transformer-ext-s-lower-security-for-{c}-outage-in-sub-network-{n}#
Transformer_ext_s_lower_security
Transformer_ext_s_lower_security:
description: >-
`Transformer-ext-s-lower-security-for-{c}-outage-in-sub-network-{n}` —
after any one outage, an extendable transformer carries at least the
negative of its rating of the chosen build. Its flow takes on its share
of the outaged branch's flow. PyPSA names one row per outaged component
`c` and sub-network `n`. Under `multi_investment_periods` it appends
`-period-{p}` for each period `p`. This entry states them all over the
outage dimension. A transformer inactive in the snapshot's period has no
row
dims: [scenario, snapshot, transformer, outage]
where: Transformer_s_nom_extendable AND at(Transformer_BODF, by=snapshot_period[period]) AND Transformer_active
expression: Transformer_s_monitored + at(Transformer_BODF, by=snapshot_period[period]) * Outage_s >= -Transformer_s_max_pu * Transformer_s_nom_ext
\[
\check{\sigma}_{\xi,t,m} + \beta^{\sigma}_{\kappa,\mathrm{snapshot\_period}(t),m} \cdot \hat{s}_{\kappa,\xi,t} \ge -\overline{\sigma}_{\xi,t,m} \cdot \Sigma_{m} \qquad \forall\, \kappa \in \mathcal{K}^{\mathrm{out}},\ \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \mathrm{ext}^{\sigma}_{m} \wedge \beta^{\sigma}_{\kappa,\mathrm{snapshot\_period}(t),m} \text{ is defined} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Transformer-ext-s_nom-lower#
Transformer_ext_s_nom_lower
Transformer_ext_s_nom_lower:
description: "`Transformer-ext-s_nom-lower` — the chosen build is at least its floor in every scenario"
dims: [scenario, transformer]
where: Transformer_s_nom_extendable AND Transformer_active_asset
expression: Transformer_s_nom_ext >= Transformer_s_nom_min
\[
\Sigma_{m} \ge \underline{\sigma}^{\mathrm{nom}}_{\xi,m} \qquad \forall\, \xi \in \Xi,\ m \in \mathcal{M} \,:\, \mathrm{ext}^{\sigma}_{m} \wedge \mathrm{act}^{\sigma}_{m}
\]
Transformer-ext-s_nom-upper#
Transformer_ext_s_nom_upper
Transformer_ext_s_nom_upper:
description: "`Transformer-ext-s_nom-upper` — the chosen build is at most its cap in every scenario. A cap of infinity is no row"
dims: [scenario, transformer]
where: Transformer_s_nom_extendable AND Transformer_active_asset AND Transformer_s_nom_max
expression: Transformer_s_nom_ext <= Transformer_s_nom_max
\[
\Sigma_{m} \le \overline{\sigma}^{\mathrm{nom}}_{\xi,m} \qquad \forall\, \xi \in \Xi,\ m \in \mathcal{M} \,:\, \mathrm{ext}^{\sigma}_{m} \wedge \mathrm{act}^{\sigma}_{m} \wedge \overline{\sigma}^{\mathrm{nom}}_{\xi,m} \text{ is defined}
\]
Transformer-ext-s-upper#
Transformer_ext_s_upper
Transformer_ext_s_upper:
description: "`Transformer-ext-s-upper` — an extendable transformer carries at most its rating of the chosen build, the loss included"
dims: [scenario, snapshot, transformer]
where: Transformer_s_nom_extendable AND Transformer_active
expression: Transformer_s + Transformer_loss <= Transformer_s_max_pu * Transformer_s_nom_ext
\[
\sigma_{\xi,t,m} + \ell^{\sigma}_{\xi,t,m} \le \overline{\sigma}_{\xi,t,m} \cdot \Sigma_{m} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \mathrm{ext}^{\sigma}_{m} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Transformer-ext-s-upper-security-for-{c}-outage-in-sub-network-{n}#
Transformer_ext_s_upper_security
Transformer_ext_s_upper_security:
description: >-
`Transformer-ext-s-upper-security-for-{c}-outage-in-sub-network-{n}` —
after any one outage, an extendable transformer carries at most its
rating of the chosen build. Its flow takes on its share of the outaged
branch's flow. PyPSA names one row per outaged component `c` and
sub-network `n`. Under `multi_investment_periods` it appends
`-period-{p}` for each period `p`. This entry states them all over the
outage dimension. A transformer inactive in the snapshot's period has no
row
dims: [scenario, snapshot, transformer, outage]
where: Transformer_s_nom_extendable AND at(Transformer_BODF, by=snapshot_period[period]) AND Transformer_active
expression: Transformer_s_monitored + at(Transformer_BODF, by=snapshot_period[period]) * Outage_s <= Transformer_s_max_pu * Transformer_s_nom_ext
\[
\check{\sigma}_{\xi,t,m} + \beta^{\sigma}_{\kappa,\mathrm{snapshot\_period}(t),m} \cdot \hat{s}_{\kappa,\xi,t} \le \overline{\sigma}_{\xi,t,m} \cdot \Sigma_{m} \qquad \forall\, \kappa \in \mathcal{K}^{\mathrm{out}},\ \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \mathrm{ext}^{\sigma}_{m} \wedge \beta^{\sigma}_{\kappa,\mathrm{snapshot\_period}(t),m} \text{ is defined} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Transformer-fix-s-lower#
Transformer_fix_s_lower
Transformer_fix_s_lower:
description: "`Transformer-fix-s-lower` — a fixed transformer carries at least the negative of its rating, the loss counted against it"
dims: [scenario, snapshot, transformer]
where: not Transformer_s_nom_extendable AND Transformer_active
expression: Transformer_s - Transformer_loss >= -Transformer_s_max_pu * Transformer_s_nom
\[
\sigma_{\xi,t,m} - \ell^{\sigma}_{\xi,t,m} \ge -\overline{\sigma}_{\xi,t,m} \cdot \sigma^{\mathrm{nom}}_{\xi,m} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \neg \mathrm{ext}^{\sigma}_{m} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Transformer-fix-s-lower-security-for-{c}-outage-in-sub-network-{n}#
Transformer_fix_s_lower_security
Transformer_fix_s_lower_security:
description: >-
`Transformer-fix-s-lower-security-for-{c}-outage-in-sub-network-{n}` —
after any one outage, a fixed transformer carries at least the negative
of its rating. Its flow takes on its share of the outaged branch's flow.
PyPSA names one row per outaged component `c` and sub-network `n`. Under
`multi_investment_periods` it appends `-period-{p}` for each period `p`.
This entry states them all over the outage dimension. A transformer
inactive in the snapshot's period has no row
dims: [scenario, snapshot, transformer, outage]
where: not Transformer_s_nom_extendable AND at(Transformer_BODF, by=snapshot_period[period]) AND Transformer_active
expression: Transformer_s_monitored + at(Transformer_BODF, by=snapshot_period[period]) * Outage_s >= -Transformer_s_max_pu * Transformer_s_nom
\[
\check{\sigma}_{\xi,t,m} + \beta^{\sigma}_{\kappa,\mathrm{snapshot\_period}(t),m} \cdot \hat{s}_{\kappa,\xi,t} \ge -\overline{\sigma}_{\xi,t,m} \cdot \sigma^{\mathrm{nom}}_{\xi,m} \qquad \forall\, \kappa \in \mathcal{K}^{\mathrm{out}},\ \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \neg \mathrm{ext}^{\sigma}_{m} \wedge \beta^{\sigma}_{\kappa,\mathrm{snapshot\_period}(t),m} \text{ is defined} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Transformer-fix-s-upper#
Transformer_fix_s_upper
Transformer_fix_s_upper:
description: "`Transformer-fix-s-upper` — a fixed transformer carries at most its rating, the loss included"
dims: [scenario, snapshot, transformer]
where: not Transformer_s_nom_extendable AND Transformer_active
expression: Transformer_s + Transformer_loss <= Transformer_s_max_pu * Transformer_s_nom
\[
\sigma_{\xi,t,m} + \ell^{\sigma}_{\xi,t,m} \le \overline{\sigma}_{\xi,t,m} \cdot \sigma^{\mathrm{nom}}_{\xi,m} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \neg \mathrm{ext}^{\sigma}_{m} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Transformer-fix-s-upper-security-for-{c}-outage-in-sub-network-{n}#
Transformer_fix_s_upper_security
Transformer_fix_s_upper_security:
description: >-
`Transformer-fix-s-upper-security-for-{c}-outage-in-sub-network-{n}` —
after any one outage, a fixed transformer carries at most its rating.
Its flow takes on its share of the outaged branch's flow. PyPSA names
one row per outaged component `c` and sub-network `n`. Under
`multi_investment_periods` it appends `-period-{p}` for each period `p`.
This entry states them all over the outage dimension. A transformer
inactive in the snapshot's period has no row
dims: [scenario, snapshot, transformer, outage]
where: not Transformer_s_nom_extendable AND at(Transformer_BODF, by=snapshot_period[period]) AND Transformer_active
expression: Transformer_s_monitored + at(Transformer_BODF, by=snapshot_period[period]) * Outage_s <= Transformer_s_max_pu * Transformer_s_nom
\[
\check{\sigma}_{\xi,t,m} + \beta^{\sigma}_{\kappa,\mathrm{snapshot\_period}(t),m} \cdot \hat{s}_{\kappa,\xi,t} \le \overline{\sigma}_{\xi,t,m} \cdot \sigma^{\mathrm{nom}}_{\xi,m} \qquad \forall\, \kappa \in \mathcal{K}^{\mathrm{out}},\ \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \neg \mathrm{ext}^{\sigma}_{m} \wedge \beta^{\sigma}_{\kappa,\mathrm{snapshot\_period}(t),m} \text{ is defined} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Transformer-loss_tangents-{k}-1#
Transformer_loss_tangents_forward
Transformer_loss_tangents_forward:
description: >-
`Transformer-loss_tangents-{k}-1`, `Transformer-loss_secants-pos` — the
loss sits above every cut to its curve for flow one way, as a line's
does, over the segment dimension
dims: [scenario, snapshot, transformer, segment]
where: Transformer_lossy
expression: Transformer_loss + Transformer_loss_slope * Transformer_s >= Transformer_loss_offset
\[
\ell^{\sigma}_{\xi,t,m} + \mathrm{a}^{\sigma}_{\xi,m,t,m} \cdot \sigma_{\xi,t,m} \ge \mathrm{b}^{\sigma}_{\xi,m,t,m} \qquad \forall\, \xi \in \Xi,\ m \in \mathcal{M},\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \mathrm{on}^{\sigma,\mathrm{lossy}}_{t,m}
\]
Transformer-loss_tangents-{k}--1#
Transformer_loss_tangents_reverse
Transformer_loss_tangents_reverse:
description: >-
`Transformer-loss_tangents-{k}--1`, `Transformer-loss_secants-neg` — the
same fan mirrored, the loss depending on the flow's magnitude
dims: [scenario, snapshot, transformer, segment]
where: Transformer_lossy
expression: Transformer_loss - Transformer_loss_slope * Transformer_s >= Transformer_loss_offset
\[
\ell^{\sigma}_{\xi,t,m} - \mathrm{a}^{\sigma}_{\xi,m,t,m} \cdot \sigma_{\xi,t,m} \ge \mathrm{b}^{\sigma}_{\xi,m,t,m} \qquad \forall\, \xi \in \Xi,\ m \in \mathcal{M},\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \mathrm{on}^{\sigma,\mathrm{lossy}}_{t,m}
\]
Transformer-loss_upper#
Transformer_loss_upper
Transformer_loss_upper:
description: "`Transformer-loss_upper` — a transformer dissipates at most the loss at its rating"
dims: [scenario, snapshot, transformer]
where: Transformer_lossy
expression: Transformer_loss <= Transformer_loss_max
\[
\ell^{\sigma}_{\xi,t,m} \le \overline{\ell}^{\sigma}_{\xi,t,m} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \mathrm{on}^{\sigma,\mathrm{lossy}}_{t,m}
\]
Transformer-s_nom_modularity#
Transformer_s_nom_modularity
Transformer_s_nom_modularity:
description: "`Transformer-s_nom_modularity` — the chosen build is a whole number of modules"
dims: [transformer]
where: Transformer_s_nom_extendable AND Transformer_s_nom_mod > 0 AND Transformer_active_asset
expression: Transformer_s_nom_ext == Transformer_s_nom_mod * Transformer_n_mod
\[
\Sigma_{m} = \sigma^{\mathrm{mod}}_{m} \cdot N^{\sigma}_{m} \qquad \forall\, m \in \mathcal{M} \,:\, \mathrm{ext}^{\sigma}_{m} \wedge \sigma^{\mathrm{mod}}_{m} > 0 \wedge \mathrm{act}^{\sigma}_{m}
\]
Transformer-s_nom_set#
Transformer_s_nom_set
Transformer_s_nom_set:
description: "`Transformer-s_nom_set` — the chosen build pinned, wherever a value is given"
dims: [transformer]
where: Transformer_s_nom_extendable AND Transformer_active_asset AND Transformer_s_nom_set
expression: Transformer_s_nom_ext == Transformer_s_nom_set
\[
\Sigma_{m} = \sigma^{\mathrm{nom,set}}_{m} \qquad \forall\, m \in \mathcal{M} \,:\, \mathrm{ext}^{\sigma}_{m} \wedge \mathrm{act}^{\sigma}_{m} \wedge \sigma^{\mathrm{nom,set}}_{m} \text{ is defined}
\]
Transformer-s_set#
Transformer_s_set
Transformer_s_set:
description: "`Transformer-s_set` — flow pinned to the given schedule, wherever one is given"
dims: [scenario, snapshot, transformer]
where: Transformer_s_set AND Transformer_active
expression: Transformer_s == Transformer_s_set
\[
\sigma_{\xi,t,m} = \sigma^{\mathrm{set}}_{\xi,t,m} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \sigma^{\mathrm{set}}_{\xi,t,m} \text{ is defined} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Transformer-v_ang-lower#
Transformer_v_ang_lower
Transformer_v_ang_lower:
description: >-
`Transformer-v_ang-lower` — a transformer with a fixed shift carries at
least the flow at which the voltage angle difference across it is minus
its limit. That difference is `x_pu_eff` times the flow plus the shift
dims: [scenario, snapshot, transformer]
where: Transformer_v_ang_max AND NOT Transformer_phase_shift_varying AND Transformer_active
expression: Transformer_s >= -(Transformer_v_ang_max + Transformer_phase_shift_fixed) * (3.141592653589793 / 180) / Transformer_x_pu_eff
\[
\sigma_{\xi,t,m} \ge \frac{-\left( \overline{\delta}^{\sigma}_{\xi,m} + \varphi^{\sigma}_{\xi,t,m} \right) \cdot \frac{3.141592653589793}{180}}{\mathrm{x}^{\mathrm{eff},\sigma}_{\xi,m}} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \overline{\delta}^{\sigma}_{\xi,m} \text{ is defined} \wedge \neg \mathrm{Transformer\_phase\_shift\_varying}_{\xi,m} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Transformer-v_ang-upper#
Transformer_v_ang_upper
Transformer_v_ang_upper:
description: >-
`Transformer-v_ang-upper` — a transformer with a fixed shift carries at
most the flow at which the voltage angle difference across it reaches
its limit. That difference includes the shift
dims: [scenario, snapshot, transformer]
where: Transformer_v_ang_max AND NOT Transformer_phase_shift_varying AND Transformer_active
expression: Transformer_s <= (Transformer_v_ang_max - Transformer_phase_shift_fixed) * (3.141592653589793 / 180) / Transformer_x_pu_eff
\[
\sigma_{\xi,t,m} \le \frac{\left( \overline{\delta}^{\sigma}_{\xi,m} - \varphi^{\sigma}_{\xi,t,m} \right) \cdot \frac{3.141592653589793}{180}}{\mathrm{x}^{\mathrm{eff},\sigma}_{\xi,m}} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \overline{\delta}^{\sigma}_{\xi,m} \text{ is defined} \wedge \neg \mathrm{Transformer\_phase\_shift\_varying}_{\xi,m} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Transformer-v_ang-var-lower#
Transformer_v_ang_var_lower
Transformer_v_ang_var_lower:
description: >-
`Transformer-v_ang-var-lower` — for a transformer whose shift is a
decision, the flow plus the shift over `x_pu_eff` is at least minus the
limit over `x_pu_eff`
dims: [scenario, snapshot, transformer]
where: Transformer_v_ang_max AND Transformer_phase_shift_varying AND Transformer_active
expression: Transformer_s + Transformer_phase_shift * (3.141592653589793 / 180) / Transformer_x_pu_eff >= -Transformer_v_ang_max * (3.141592653589793 / 180) / Transformer_x_pu_eff
\[
\sigma_{\xi,t,m} + \frac{\mathit{Transformer\_phase\_shift}_{\xi,t,m} \cdot \frac{3.141592653589793}{180}}{\mathrm{x}^{\mathrm{eff},\sigma}_{\xi,m}} \ge \frac{-\overline{\delta}^{\sigma}_{\xi,m} \cdot \frac{3.141592653589793}{180}}{\mathrm{x}^{\mathrm{eff},\sigma}_{\xi,m}} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \overline{\delta}^{\sigma}_{\xi,m} \text{ is defined} \wedge \mathrm{Transformer\_phase\_shift\_varying}_{\xi,m} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Transformer-v_ang-var-upper#
Transformer_v_ang_var_upper
Transformer_v_ang_var_upper:
description: >-
`Transformer-v_ang-var-upper` — for a transformer whose shift is a
decision, the flow plus the shift over `x_pu_eff` is at most the limit
over `x_pu_eff`
dims: [scenario, snapshot, transformer]
where: Transformer_v_ang_max AND Transformer_phase_shift_varying AND Transformer_active
expression: Transformer_s + Transformer_phase_shift * (3.141592653589793 / 180) / Transformer_x_pu_eff <= Transformer_v_ang_max * (3.141592653589793 / 180) / Transformer_x_pu_eff
\[
\sigma_{\xi,t,m} + \frac{\mathit{Transformer\_phase\_shift}_{\xi,t,m} \cdot \frac{3.141592653589793}{180}}{\mathrm{x}^{\mathrm{eff},\sigma}_{\xi,m}} \le \frac{\overline{\delta}^{\sigma}_{\xi,m} \cdot \frac{3.141592653589793}{180}}{\mathrm{x}^{\mathrm{eff},\sigma}_{\xi,m}} \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ m \in \mathcal{M} \,:\, \overline{\delta}^{\sigma}_{\xi,m} \text{ is defined} \wedge \mathrm{Transformer\_phase\_shift\_varying}_{\xi,m} \wedge \mathrm{on}^{\sigma}_{t,m}
\]
Generator_marginal_cost_incremental#
Generator_marginal_cost_incremental:
description: >-
`Generator-marginal_cost_piecewise` — the running cost of a generator
with a curve is at least the curve at its output, stated by segments
filled in order and switched by the status of a committable generator,
where linopy chooses the incremental rows
along: breakpoint
dims: [scenario, snapshot, generator]
where: NOT Generator_marginal_cost_as_lp AND Generator_marginal_cost_curve AND Generator_active AND Generator_marginal_cost_curve_p
method: incremental
activity: Generator_status
links:
p: [Generator_p, Generator_marginal_cost_curve_p]
cost: [Generator_marginal_cost_piecewise, Generator_marginal_cost_curve_cost, ">="]
\[
C^{\mathrm{curve}}_{g,\xi,t} \ge \left( \begin{cases} u_{g,\xi,t} & \text{if } \mathrm{on}^{\mathrm{com}}_{g,t} \\ 1 & \text{otherwise} \end{cases} \cdot \left( \mathrm{p}^{\mathrm{curve}}_{\min \{ b \in \mathcal{B} \,:\, \neg \mathrm{lp} \wedge \mathrm{curve}_{g} \wedge \mathrm{on}_{g,t} \wedge \mathrm{p}^{\mathrm{curve}}_{b,g} \text{ is defined} \},g},\ \mathrm{C}^{\mathrm{curve}}_{\min \{ b \in \mathcal{B} \,:\, \neg \mathrm{lp} \wedge \mathrm{curve}_{g} \wedge \mathrm{on}_{g,t} \wedge \mathrm{p}^{\mathrm{curve}}_{b,g} \text{ is defined} \},g} \right) + \min(\begin{cases} u_{g,\xi,t} & \text{if } \mathrm{on}^{\mathrm{com}}_{g,t} \\ 1 & \text{otherwise} \end{cases},\ 1) \cdot \left( \mathrm{pwl}_{b \in \mathcal{B} \,:\, \neg \mathrm{lp} \wedge \mathrm{curve}_{g} \wedge \mathrm{on}_{g,t} \wedge \mathrm{p}^{\mathrm{curve}}_{b,g} \text{ is defined}}(\mathrm{p}^{\mathrm{curve}}_{b,g},\ \mathrm{C}^{\mathrm{curve}}_{b,g}) - \left( \mathrm{p}^{\mathrm{curve}}_{\min \{ b \in \mathcal{B} \,:\, \neg \mathrm{lp} \wedge \mathrm{curve}_{g} \wedge \mathrm{on}_{g,t} \wedge \mathrm{p}^{\mathrm{curve}}_{b,g} \text{ is defined} \},g},\ \mathrm{C}^{\mathrm{curve}}_{\min \{ b \in \mathcal{B} \,:\, \neg \mathrm{lp} \wedge \mathrm{curve}_{g} \wedge \mathrm{on}_{g,t} \wedge \mathrm{p}^{\mathrm{curve}}_{b,g} \text{ is defined} \},g} \right) \right) \right)(p_{g,\xi,t}) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_marginal_cost_lp#
Generator_marginal_cost_lp:
description: >-
`Generator-marginal_cost_piecewise` — the running cost of a generator
with a curve is at least the curve at its output, stated as the line of
each segment, where linopy chooses the segment lines
along: breakpoint
dims: [scenario, snapshot, generator]
where: Generator_marginal_cost_as_lp AND Generator_marginal_cost_curve AND Generator_active AND Generator_marginal_cost_curve_p
method: lp
links:
p: [Generator_p, Generator_marginal_cost_curve_p]
cost: [Generator_marginal_cost_piecewise, Generator_marginal_cost_curve_cost, ">="]
\[
C^{\mathrm{curve}}_{g,\xi,t} \ge \mathrm{pwl}_{b \in \mathcal{B} \,:\, \mathrm{lp} \wedge \mathrm{curve}_{g} \wedge \mathrm{on}_{g,t} \wedge \mathrm{p}^{\mathrm{curve}}_{b,g} \text{ is defined}}(\mathrm{p}^{\mathrm{curve}}_{b,g},\ \mathrm{C}^{\mathrm{curve}}_{b,g})(p_{g,\xi,t}) \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T}
\]
Generator_big_m_where_committed_and_extendable#
Generator_big_m_where_committed_and_extendable:
holds: "Generator_big_m"
where: "Generator_com_ext"
description: >-
a committable generator with an extendable, non-modular build has
the big-M its rows release it by — PyPSA computes one for each
(`get_committable_big_m_values`)
\[
\mathrm{M}_{g,\xi} \text{ is defined} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{\mathrm{com,ext}}_{g,t}
\]
Generator_maintenance_build_cap_is_finite#
Generator_maintenance_build_cap_is_finite:
holds: "Generator_p_nom_max < inf"
where: "Generator_maintainable AND Generator_p_nom_extendable"
description: >-
the `maintcap` rows bound the chosen build in maintenance by
`p_nom_max`, so an infinite cap is an infinite coefficient. PyPSA
refuses it (`consistency.py:1576`)
\[
\overline{\mathrm{p}}^{\mathrm{nom}}_{g,\xi} < \infty \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi \,:\, \mathrm{mnt}_{g} \wedge \mathrm{ext}_{g}
\]
Generator_maintenance_duration_fits_the_horizon#
Generator_maintenance_duration_fits_the_horizon:
holds: "Generator_maintenance_duration <= sum(snapshot_weightings_generators, over=snapshot)"
where: "Generator_maintainable"
description: >-
one event longer than the horizon, in generator weightings, blocks every
start and makes the event count infeasible. PyPSA refuses it
(`consistency.py:1552`)
\[
\tau^{\mathrm{mnt}}_{g,\xi} \le \sum_{t \in \mathcal{T}} \mathrm{w}^{\mathrm{gen}}_{t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi \,:\, \mathrm{mnt}_{g}
\]
Generator_maintenance_duration_positive#
Generator_maintenance_duration_positive:
holds: "Generator_maintenance_duration > 0"
where: "Generator_maintainable"
description: >-
an event that covers no hours is no maintenance window. PyPSA refuses it
(`consistency.py:1531`)
\[
\tau^{\mathrm{mnt}}_{g,\xi} > 0 \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi \,:\, \mathrm{mnt}_{g}
\]
Generator_maintenance_events_fit_the_horizon#
Generator_maintenance_events_fit_the_horizon:
holds: "Generator_maintenance_duration * Generator_maintenance_events <= sum(snapshot_weightings_generators, over=snapshot)"
where: "Generator_maintainable"
description: >-
the events together longer than the horizon, in generator weightings,
cannot all be scheduled. PyPSA refuses it (`consistency.py:1564`)
\[
\tau^{\mathrm{mnt}}_{g,\xi} \cdot \mathrm{n}^{\mathrm{mnt}}_{g,\xi} \le \sum_{t \in \mathcal{T}} \mathrm{w}^{\mathrm{gen}}_{t} \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi \,:\, \mathrm{mnt}_{g}
\]
Generator_maintenance_events_positive#
Generator_maintenance_events_positive:
holds: "Generator_maintenance_events > 0"
where: "Generator_maintainable"
description: >-
a maintainable generator with no event schedules no maintenance. PyPSA
refuses it (`consistency.py:1541`)
\[
\mathrm{n}^{\mathrm{mnt}}_{g,\xi} > 0 \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi \,:\, \mathrm{mnt}_{g}
\]
Generator_marginal_cost_curve_on_a_fixed_build#
Generator_marginal_cost_curve_on_a_fixed_build:
holds: "NOT Generator_p_nom_extendable AND Generator_p_nom > 0 AND Generator_p_nom < inf"
where: "Generator_marginal_cost_curve"
description: >-
a generator's cost curve is given per unit of a fixed, positive, finite
`p_nom`, which scales it to output. PyPSA refuses an extendable unit and
a zero, missing or infinite `p_nom` (`piecewise.py:284-304`), and takes
a negative one, which the file does not state
\[
\neg \mathrm{ext}_{g} \wedge \mathrm{p}^{\mathrm{nom}}_{g,\xi} > 0 \wedge \mathrm{p}^{\mathrm{nom}}_{g,\xi} < \infty \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi \,:\, \mathrm{curve}_{g}
\]
Generator_marginal_cost_quadratic_without_risk_preference#
Generator_marginal_cost_quadratic_without_risk_preference:
holds: "Generator_marginal_cost_quadratic == 0"
where: "CVaR_omega"
description: >-
a quadratic cost puts a square into every `CVaR-excess` row. PyPSA
refuses quadratic costs under any risk preference, `omega = 0` included
(`optimize.py:506-513`). In data prep you write a `CVaR_omega` row only
where a risk preference is set
\[
\mathrm{c}^{(2)}_{g,\xi,t} = 0 \qquad \forall\, g \in \mathcal{G},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \omega \text{ is defined}
\]
GlobalConstraint_flow_based_sense_is_le#
GlobalConstraint_flow_based_sense_is_le:
holds: "GlobalConstraint_sense == '<='"
where: "GlobalConstraint_type == 'flow_based'"
description: >-
PyPSA refuses a `flow_based` row whose sense is not `<=`
(`flow_based.py:109-114`)
\[
\mathrm{sense}_{o,\xi} = \text{'}\mathrm{<=}\text{'} \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi \,:\, \mathrm{type}_{o} = \text{'}\mathrm{flow\_based}\text{'}
\]
GlobalConstraint_tech_capacity_expansion_limit_without_scenarios#
GlobalConstraint_tech_capacity_expansion_limit_without_scenarios:
holds: "GlobalConstraint_type != 'tech_capacity_expansion_limit'"
where: "count(scenario_weight, over=scenario) > 1"
description: >-
PyPSA does not build a `tech_capacity_expansion_limit` row on a network
with scenarios and refuses it (`global_constraints.py:68-70`). The spec
cannot tell a network with one scenario from one with none. A tool thus
refuses the row only where there is more than one scenario
\[
\mathrm{type}_{o} \neq \text{'}\mathrm{tech\_capacity\_expansion\_limit}\text{'} \qquad \forall\, o \in \mathcal{O} \,:\, \lvert \{ \xi \in \Xi \,:\, \pi_{\xi} \text{ is defined} \} \rvert > 1
\]
Link_big_m_where_committed_and_extendable#
Link_big_m_where_committed_and_extendable:
holds: "Link_big_m"
where: "Link_com_ext"
description: >-
a committable link with an extendable, non-modular build has
the big-M its rows release it by — PyPSA computes one for each
(`get_committable_big_m_values`)
\[
\mathrm{M}^{f}_{l,\xi} \text{ is defined} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{f,\mathrm{com,ext}}_{l,t}
\]
Link_maintenance_build_cap_is_finite#
Link_maintenance_build_cap_is_finite:
holds: "Link_p_nom_max < inf"
where: "Link_maintainable AND Link_p_nom_extendable"
description: >-
the `maintcap` rows bound the chosen build in maintenance by
`p_nom_max`, so an infinite cap is an infinite coefficient. PyPSA
refuses it (`consistency.py:1576`)
\[
\overline{\mathrm{f}}^{\mathrm{nom}}_{l,\xi} < \infty \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi \,:\, \mathrm{mnt}^{f}_{l} \wedge \mathrm{ext}^{f}_{l}
\]
Link_maintenance_duration_fits_the_horizon#
Link_maintenance_duration_fits_the_horizon:
holds: "Link_maintenance_duration <= sum(snapshot_weightings_generators, over=snapshot)"
where: "Link_maintainable"
description: >-
one event longer than the horizon, in generator weightings, blocks every
start and makes the event count infeasible. PyPSA refuses it
(`consistency.py:1552`)
\[
\tau^{f,\mathrm{mnt}}_{l,\xi} \le \sum_{t \in \mathcal{T}} \mathrm{w}^{\mathrm{gen}}_{t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi \,:\, \mathrm{mnt}^{f}_{l}
\]
Link_maintenance_duration_positive#
Link_maintenance_duration_positive:
holds: "Link_maintenance_duration > 0"
where: "Link_maintainable"
description: >-
an event that covers no hours is no maintenance window. PyPSA refuses it
(`consistency.py:1531`)
\[
\tau^{f,\mathrm{mnt}}_{l,\xi} > 0 \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi \,:\, \mathrm{mnt}^{f}_{l}
\]
Link_maintenance_events_fit_the_horizon#
Link_maintenance_events_fit_the_horizon:
holds: "Link_maintenance_duration * Link_maintenance_events <= sum(snapshot_weightings_generators, over=snapshot)"
where: "Link_maintainable"
description: >-
the events together longer than the horizon, in generator weightings,
cannot all be scheduled. PyPSA refuses it (`consistency.py:1564`)
\[
\tau^{f,\mathrm{mnt}}_{l,\xi} \cdot \mathrm{n}^{f,\mathrm{mnt}}_{l,\xi} \le \sum_{t \in \mathcal{T}} \mathrm{w}^{\mathrm{gen}}_{t} \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi \,:\, \mathrm{mnt}^{f}_{l}
\]
Link_maintenance_events_positive#
Link_maintenance_events_positive:
holds: "Link_maintenance_events > 0"
where: "Link_maintainable"
description: >-
a maintainable link with no event schedules no maintenance. PyPSA
refuses it (`consistency.py:1541`)
\[
\mathrm{n}^{f,\mathrm{mnt}}_{l,\xi} > 0 \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi \,:\, \mathrm{mnt}^{f}_{l}
\]
Link_marginal_cost_quadratic_without_risk_preference#
Link_marginal_cost_quadratic_without_risk_preference:
holds: "Link_marginal_cost_quadratic == 0"
where: "CVaR_omega"
description: >-
a quadratic cost puts a square into every `CVaR-excess` row. PyPSA
refuses quadratic costs under any risk preference, `omega = 0` included
(`optimize.py:506-513`). In data prep you write a `CVaR_omega` row only
where a risk preference is set
\[
\mathrm{c}^{f,(2)}_{l,\xi,t} = 0 \qquad \forall\, l \in \mathcal{L},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \omega \text{ is defined}
\]
Link_output_delay_under_uniform_weighting#
Link_output_delay_under_uniform_weighting:
holds: "snapshot_weightings_generators == shift(snapshot_weightings_generators, along=snapshot, offset=1, edge='wrap')"
where: "Link_output_delay > 0"
description: >-
a delayed port lags by the same whole number of snapshots in every
period. PyPSA lags by elapsed time in `generators` weighting and rounds
the source down to a snapshot start (`multiports.py:100-133`). The two
agree only where that weighting is one value over the horizon. PyPSA
does not refuse an uneven weighting, so a tool refuses it through this
assumption when data is attached (#299)
\[
\mathrm{w}^{\mathrm{gen}}_{t} = \mathrm{w}^{\mathrm{gen}}_{t \ominus 1} \qquad \forall\, o \in \mathcal{O},\ t \in \mathcal{T} \,:\, \mathrm{d}^{f}_{o} > 0
\]
Process_big_m_where_committed_and_extendable#
Process_big_m_where_committed_and_extendable:
holds: "Process_big_m"
where: "Process_com_ext"
description: >-
a committable process with an extendable, non-modular build has
the big-M its rows release it by — PyPSA computes one for each
(`get_committable_big_m_values`)
\[
\mathrm{M}^{z}_{j,\xi} \text{ is defined} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \mathrm{on}^{z,\mathrm{com,ext}}_{j,t}
\]
Process_maintenance_build_cap_is_finite#
Process_maintenance_build_cap_is_finite:
holds: "Process_p_nom_max < inf"
where: "Process_maintainable AND Process_p_nom_extendable"
description: >-
the `maintcap` rows bound the chosen build in maintenance by
`p_nom_max`, so an infinite cap is an infinite coefficient. PyPSA
refuses it (`consistency.py:1576`)
\[
\overline{\mathrm{z}}^{\mathrm{nom}}_{j,\xi} < \infty \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi \,:\, \mathrm{mnt}^{z}_{j} \wedge \mathrm{ext}^{z}_{j}
\]
Process_maintenance_duration_fits_the_horizon#
Process_maintenance_duration_fits_the_horizon:
holds: "Process_maintenance_duration <= sum(snapshot_weightings_generators, over=snapshot)"
where: "Process_maintainable"
description: >-
one event longer than the horizon, in generator weightings, blocks every
start and makes the event count infeasible. PyPSA refuses it
(`consistency.py:1552`)
\[
\tau^{z,\mathrm{mnt}}_{j,\xi} \le \sum_{t \in \mathcal{T}} \mathrm{w}^{\mathrm{gen}}_{t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi \,:\, \mathrm{mnt}^{z}_{j}
\]
Process_maintenance_duration_positive#
Process_maintenance_duration_positive:
holds: "Process_maintenance_duration > 0"
where: "Process_maintainable"
description: >-
an event that covers no hours is no maintenance window. PyPSA refuses it
(`consistency.py:1531`)
\[
\tau^{z,\mathrm{mnt}}_{j,\xi} > 0 \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi \,:\, \mathrm{mnt}^{z}_{j}
\]
Process_maintenance_events_fit_the_horizon#
Process_maintenance_events_fit_the_horizon:
holds: "Process_maintenance_duration * Process_maintenance_events <= sum(snapshot_weightings_generators, over=snapshot)"
where: "Process_maintainable"
description: >-
the events together longer than the horizon, in generator weightings,
cannot all be scheduled. PyPSA refuses it (`consistency.py:1564`)
\[
\tau^{z,\mathrm{mnt}}_{j,\xi} \cdot \mathrm{n}^{z,\mathrm{mnt}}_{j,\xi} \le \sum_{t \in \mathcal{T}} \mathrm{w}^{\mathrm{gen}}_{t} \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi \,:\, \mathrm{mnt}^{z}_{j}
\]
Process_maintenance_events_positive#
Process_maintenance_events_positive:
holds: "Process_maintenance_events > 0"
where: "Process_maintainable"
description: >-
a maintainable process with no event schedules no maintenance. PyPSA
refuses it (`consistency.py:1541`)
\[
\mathrm{n}^{z,\mathrm{mnt}}_{j,\xi} > 0 \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi \,:\, \mathrm{mnt}^{z}_{j}
\]
Process_marginal_cost_quadratic_without_risk_preference#
Process_marginal_cost_quadratic_without_risk_preference:
holds: "Process_marginal_cost_quadratic == 0"
where: "CVaR_omega"
description: >-
a quadratic cost puts a square into every `CVaR-excess` row. PyPSA
refuses quadratic costs under any risk preference, `omega = 0` included
(`optimize.py:506-513`). In data prep you write a `CVaR_omega` row only
where a risk preference is set
\[
\mathrm{c}^{z,(2)}_{j,\xi,t} = 0 \qquad \forall\, j \in \mathcal{J},\ \xi \in \Xi,\ t \in \mathcal{T} \,:\, \omega \text{ is defined}
\]
Process_output_delay_under_uniform_weighting#
Process_output_delay_under_uniform_weighting:
holds: "snapshot_weightings_generators == shift(snapshot_weightings_generators, along=snapshot, offset=1, edge='wrap')"
where: "Process_output_delay > 0"
description: >-
the same for a delayed process port (#299)
\[
\mathrm{w}^{\mathrm{gen}}_{t} = \mathrm{w}^{\mathrm{gen}}_{t \ominus 1} \qquad \forall\, r \in \mathcal{R},\ t \in \mathcal{T} \,:\, \mathrm{d}^{z}_{r} > 0
\]
StorageUnit_marginal_cost_quadratic_without_risk_preference#
StorageUnit_marginal_cost_quadratic_without_risk_preference:
holds: "StorageUnit_marginal_cost_quadratic == 0"
where: "CVaR_omega"
description: >-
a quadratic cost puts a square into every `CVaR-excess` row. PyPSA
refuses quadratic costs under any risk preference, `omega = 0` included
(`optimize.py:506-513`). In data prep you write a `CVaR_omega` row only
where a risk preference is set
\[
\mathrm{c}^{h,(2)}_{\xi,t,s} = 0 \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \omega \text{ is defined}
\]
StorageUnit_opens_late_only_where_it_opens#
StorageUnit_opens_late_only_where_it_opens:
holds: "StorageUnit_active AND NOT shift(StorageUnit_active, along=snapshot, offset=1) AND position(snapshot) > 0"
where: "StorageUnit_opens_late"
description: >-
`StorageUnit_opens_late` marks the first snapshot the storage unit is active in, past
the first of the horizon, and no other
\[
\mathrm{on}^{h}_{t,s} \wedge \neg \mathrm{on}^{h}_{t - 1,s} \wedge \mathrm{pos}(t) > 0 \qquad \forall\, t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{open}_{t,s}
\]
StorageUnit_opens_late_where_it_opens#
StorageUnit_opens_late_where_it_opens:
holds: "StorageUnit_opens_late"
where: "StorageUnit_active AND NOT shift(StorageUnit_active, along=snapshot, offset=1) AND position(snapshot) > 0"
description: >-
a storage unit that opens past the first snapshot of the horizon opens
on its initial level, or its last level where it is cyclic. This holds
only where `StorageUnit_opens_late` is true at the snapshot
\[
\mathrm{open}_{t,s} \qquad \forall\, t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{on}^{h}_{t,s} \wedge \neg \mathrm{on}^{h}_{t - 1,s} \wedge \mathrm{pos}(t) > 0
\]
StorageUnit_operational_limit_carried_over_has_unit_years#
StorageUnit_operational_limit_carried_over_has_unit_years:
holds: "at(period_weight_years == 1, by=snapshot_period[period])"
where: "StorageUnit_operational_limit_weight AND NOT StorageUnit_state_of_charge_initial_per_period AND GlobalConstraint_counts_snapshot"
description: >-
the same for an `operational_limit` row, over the periods it counts.
PyPSA refuses it (`global_constraints.py:674`)
\[
\mathrm{w}^{\mathrm{yr}}_{\mathrm{snapshot\_period}(t)} = 1 \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi,\ t \in \mathcal{T},\ s \in \mathcal{S} \,:\, \mathrm{b}^{h}_{o,\xi,s} \text{ is defined} \wedge \neg \mathrm{reset}_{\xi,s} \wedge \mathrm{in}_{o,\xi,t}
\]
StorageUnit_primary_energy_carried_over_has_unit_years#
StorageUnit_primary_energy_carried_over_has_unit_years:
holds: "period_weight_years == 1"
where: "StorageUnit_primary_energy_weight AND NOT StorageUnit_state_of_charge_initial_per_period"
description: >-
a unit that carries its charge from one period to the next closes once,
at the last counted snapshot. No period's years weighs that level. PyPSA
refuses it where any period's years is not one
(`global_constraints.py:464`)
\[
\mathrm{w}^{\mathrm{yr}}_{y} = 1 \qquad \forall\, o \in \mathcal{O},\ y \in \mathcal{Y},\ \xi \in \Xi,\ s \in \mathcal{S} \,:\, \mathrm{a}^{h}_{o,\xi,s} \text{ is defined} \wedge \neg \mathrm{reset}_{\xi,s}
\]
StorageUnit_stands_in_one_run#
StorageUnit_stands_in_one_run:
holds: "count(StorageUnit_active AND NOT shift(StorageUnit_active, along=snapshot, offset=1), over=snapshot) <= 1"
description: >-
the snapshots a storage unit is active in are one unbroken run, as a build
year and a lifetime make them. The opening row holds at the first
of them only, and a cyclic storage unit reaches back
`StorageUnit_inactive_snapshots` further to the last of them
\[
\lvert \{ t \in \mathcal{T} \,:\, \mathrm{on}^{h}_{t,s} \wedge \neg \mathrm{on}^{h}_{t - 1,s} \} \rvert \le 1 \qquad \forall\, s \in \mathcal{S}
\]
Store_marginal_cost_quadratic_without_risk_preference#
Store_marginal_cost_quadratic_without_risk_preference:
holds: "Store_marginal_cost_quadratic == 0"
where: "CVaR_omega"
description: >-
a quadratic cost puts a square into every `CVaR-excess` row. PyPSA
refuses quadratic costs under any risk preference, `omega = 0` included
(`optimize.py:506-513`). In data prep you write a `CVaR_omega` row only
where a risk preference is set
\[
\mathrm{c}^{q,(2)}_{\xi,t,v} = 0 \qquad \forall\, \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \omega \text{ is defined}
\]
Store_opens_late_only_where_it_opens#
Store_opens_late_only_where_it_opens:
holds: "Store_active AND NOT shift(Store_active, along=snapshot, offset=1) AND position(snapshot) > 0"
where: "Store_opens_late"
description: >-
`Store_opens_late` marks the first snapshot the store is active in, past
the first of the horizon, and no other
\[
\mathrm{on}^{e}_{t,v} \wedge \neg \mathrm{on}^{e}_{t - 1,v} \wedge \mathrm{pos}(t) > 0 \qquad \forall\, t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{open}^{e}_{t,v}
\]
Store_opens_late_where_it_opens#
Store_opens_late_where_it_opens:
holds: "Store_opens_late"
where: "Store_active AND NOT shift(Store_active, along=snapshot, offset=1) AND position(snapshot) > 0"
description: >-
a store that opens past the first snapshot of the horizon opens on its
initial level, or its last level where it is cyclic. This holds only
where `Store_opens_late` is true at the snapshot
\[
\mathrm{open}^{e}_{t,v} \qquad \forall\, t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{on}^{e}_{t,v} \wedge \neg \mathrm{on}^{e}_{t - 1,v} \wedge \mathrm{pos}(t) > 0
\]
Store_operational_limit_carried_over_has_unit_years#
Store_operational_limit_carried_over_has_unit_years:
holds: "at(period_weight_years == 1, by=snapshot_period[period])"
where: "Store_operational_limit_weight AND NOT Store_e_initial_per_period AND GlobalConstraint_counts_snapshot"
description: >-
the same for an `operational_limit` row, over the periods it counts.
PyPSA refuses it (`global_constraints.py:724`)
\[
\mathrm{w}^{\mathrm{yr}}_{\mathrm{snapshot\_period}(t)} = 1 \qquad \forall\, o \in \mathcal{O},\ \xi \in \Xi,\ t \in \mathcal{T},\ v \in \mathcal{V} \,:\, \mathrm{b}^{e}_{o,\xi,v} \text{ is defined} \wedge \neg \mathrm{reset}^{e}_{\xi,v} \wedge \mathrm{in}_{o,\xi,t}
\]
Store_primary_energy_carried_over_has_unit_years#
Store_primary_energy_carried_over_has_unit_years:
holds: "period_weight_years == 1"
where: "Store_primary_energy_weight AND NOT Store_e_initial_per_period"
description: >-
a store that carries its energy from one period to the next closes once,
at the last counted snapshot. No period's years weighs that level. PyPSA
refuses it where any period's years is not one
(`global_constraints.py:518`)
\[
\mathrm{w}^{\mathrm{yr}}_{y} = 1 \qquad \forall\, o \in \mathcal{O},\ y \in \mathcal{Y},\ \xi \in \Xi,\ v \in \mathcal{V} \,:\, \mathrm{a}^{e}_{o,\xi,v} \text{ is defined} \wedge \neg \mathrm{reset}^{e}_{\xi,v}
\]
Store_stands_in_one_run#
Store_stands_in_one_run:
holds: "count(Store_active AND NOT shift(Store_active, along=snapshot, offset=1), over=snapshot) <= 1"
description: >-
the snapshots a store is active in are one unbroken run, as a build
year and a lifetime make them. The opening row holds at the first
of them only, and a cyclic store reaches back
`Store_inactive_snapshots` further to the last of them
\[
\lvert \{ t \in \mathcal{T} \,:\, \mathrm{on}^{e}_{t,v} \wedge \neg \mathrm{on}^{e}_{t - 1,v} \} \rvert \le 1 \qquad \forall\, v \in \mathcal{V}
\]
See also
- PyPSA in 25 files, for the same spec split into one file per component
- PyPSA, the relaxed commitment, for the patch that relaxes the commitment rows
- Composition, for the
mergeandoverridecalls that combine files - Assumptions, for the checks a tool runs when it attaches data