Reference networks¶

monee ships fixed multi-energy reference networks for testing, benchmarking, and tutorials. Import them from monee.network. Every factory is deterministic (hand-built or internally seeded), so repeated calls return the same grid. They couple electricity, gas, and (in most cases) heat through conversion units (CHP, P2H, P2G, G2P).

To generate a multi-energy overlay on top of an arbitrary power network instead (gas and heat grids derived from the electrical topology, with configurable coupling-point density), see Generate synthetic multi-energy networks.

Available networks¶

Function	Scale	Description
`create_urban_district_net()`	Small	20 kV residential district. 5 buses · 5 gas junctions · 4 heat junctions. CHP-centred topology with high coupling-point density. Best for load-shedding and demand-side flexibility studies.
`create_large_urban_mes_net()`	Scalable	`n_districts` replicated urban districts under a shared HV slack and a shared gas feeder. ~96 nodes at the default 6 districts, ~320 at 20.
`create_monee_benchmark_net()`	Small	Seeded 7-bus 120 kV MES with gas and heat overlay, CHP, P2G, two G2P units, a grid-forming generator, and two normally-open backup lines (`backup=True`, `on_off=0`).
`create_mv_multi_cigre()`	Medium	Seeded MES built on the pandapower CIGRE MV grid (with PV and wind DER), overlaid with gas and heat plus P2G and CHP, and one open backup line. Requires the optional `pandapower` dependency.
`create_restoration_benchmark()`	Large	Curated multi-energy restoration benchmark, see the dedicated section below.

Quick start¶

Load and run an energy-flow calculation on the urban district network:

import monee.model as mm
from monee import run_energy_flow
from monee.network import create_urban_district_net

net = create_urban_district_net()
result = run_energy_flow(net)

print(result.get(mm.Bus)[["vm_pu"]].shape)
print(result.get(mm.GasPipe)[["mass_flow_kgs"]].shape)

(5, 1)
(4, 1)

Scaling up: replicated districts¶

create_large_urban_mes_net() replicates the urban-district pattern n_districts times under a shared HV slack bus and a shared external gas feeder. Each replicated district adds its own normally-open internal ties (on_off=0, named tie_*) across all three carriers plus a second heat consumer. Adjacent districts are linked by normally-open MV power ties and live gas trunks, so electricity and gas each remain a single connected component. The heat sector is intentionally local per district (one supply-return consumer pair each) and therefore decomposes into one connected component per district.

from monee.network import create_large_urban_mes_net

net = create_large_urban_mes_net(n_districts=6)   # ~96 nodes
big = create_large_urban_mes_net(n_districts=20)  # ~320 nodes

Each district contributes 5 power buses, 4 gas junctions, 5 water junctions, 2 heat consumers, 3 coupling points (CHP, P2G, G2P), and 3 internal ties.

Seeded benchmark nets with backup lines¶

Two further fixed MES grids live in monee.network.mes. Both are seeded internally, so they are reproducible despite using the randomized overlay generators under the hood:

create_monee_benchmark_net(): a 7-bus 120 kV power grid with full gas and heat overlay, a CHP, a P2G, two G2P units, and a GridFormingGenerator (useful together with islanding). Two extra power lines are created as backup assets with backup=True and on_off=0: open in normal operation, available to a restoration algorithm.
create_mv_multi_cigre(): the pandapower CIGRE MV benchmark (with PV and wind DER) converted to monee and overlaid with gas and heat grids, a P2G, a CHP, and one open backup line. Requires pandapower.

The restoration benchmark¶

create_restoration_benchmark() (also re-exported from monee.network) is a curated multi-energy grid built for restoration-sequence planning, resilience studies, and multi-period optimisation:

def create_restoration_benchmark(*, linepack=False, ltc=False, misocp=True)

Electricity: two 110 kV feeders (each with its own external grid) joined by an HV tie, four 20 kV substations behind 63 MVA transformers, and three five-bus load chains (industrial, commercial, residential) with distributed solar and wind generation.
Gas: two high-pressure feeders with a trunk pipe, three compressors feeding three eight-junction medium-pressure chains plus two spur junctions, and fourteen gas sinks.
Heat: a long district-heating supply chain (120 junctions) fed by a 358 K heat plant, with a common return junction and consumer heat exchangers along the chain.
Coupling points: two CHP-HG units (gas → power + heat, node-based heat injection), one P2G electrolyser, and one G2P peaker.

The keyword-only flags select optional features:

linepack=True attaches the GasLinepack extension (inherent gas storage in pipes),
ltc=True attaches LumpedThermalCapacitance (thermal inertia at heat junctions),
misocp=True (the default) applies EL_MISOCP_FORMULATION to the power grid, replacing the nonlinear AC formulation.

Both extensions only take effect in multi-period runs, see Temporal extensions.

Tip

With the default misocp=True, pair the benchmark with a MIQCP-capable solver such as Gurobi. For the nonlinear variant (misocp=False), use the PyomoSolver with ipopt and relaxed tolerances.

import monee
from monee.network import create_restoration_benchmark

# Default: MISOCP formulation already applied - solve with Gurobi.
net = create_restoration_benchmark()
result = monee.run_energy_flow(net, solver="gurobi")

# Nonlinear variant for interior-point solvers.
from monee.solver.pyo import DEFAULT_SOLVER_OPTIONS

DEFAULT_SOLVER_OPTIONS["max_iter"] = 5000
DEFAULT_SOLVER_OPTIONS["tol"] = 1e-4

net_nl = create_restoration_benchmark(misocp=False)
result = monee.run_energy_flow(net_nl, solver="ipopt", backend="pyomo")

Load shedding on a reference network¶

The urban district network is a good starting point for load-shedding experiments: it couples power, gas, and heat through a CHP, a P2G, and a G2P.

import monee
from monee.network import create_urban_district_net

net = create_urban_district_net()
net.apply_formulation(monee.EL_MISOCP_FORMULATION)

problem = monee.create_min_load_shedding_problem(
    bounds_vm=(0.9, 1.1),
    bounds_t=(0.9, 1.1),
    bounds_pressure=(0.9, 1.1),
    include_ext_grids=True,
)

result = monee.run_energy_flow_optimization(
    net,
    problem,
    solver="gurobi",
    exclude_unconnected_nodes=True,
)

print(result.dataframes["PowerLoad"][["regulation"]])

Import / export capacity limits¶

Each external grid connection accepts optional capacity bounds that cap the power or mass-flow that can be exchanged with the upstream network. These are set at network construction time:

import monee.express as mx

net_cap = mx.create_multi_energy_network()
bus_cap = mx.create_bus(net_cap)

# Limit the electrical connection to 4 MW import, 1 MW export.
mx.create_ext_power_grid(net_cap, bus_cap, max_import_mw=4.0, max_export_mw=1.0)

# Gas connection: import capped at 0.1 kg/s, export unbounded.
junc_cap = mx.create_gas_junction(net_cap)
mx.create_ext_hydr_grid(net_cap, junc_cap, max_import_kgs=0.1)

The reference networks use these limits to reflect realistic import capacities for each grid type.

Reference networks¶

Available networks¶

Quick start¶

Scaling up: replicated districts¶

Seeded benchmark nets with backup lines¶

The restoration benchmark¶

Load shedding on a reference network¶

Import / export capacity limits¶

See also¶