Convert from pandapower¶

If you already have an electrical network defined in pandapower, monee can import it directly. The helper function serialises the pandapower network to a temporary MATPOWER .mat file and reads it back into monee for you.

Warning

This feature is experimental. Complex pandapower networks with special elements (three-winding transformers, DC lines, switches) may not convert correctly. Always verify results against the original network.

Prerequisites¶

pandapower (and simbench, for the section below) are optional dependencies. Install them via the simbench extra:

pip install monee[simbench]

or install pandapower alone if you only need the converter:

pip install pandapower

Converting a network¶

import pandapower as pp
import pandapower.networks as pn
from monee.io.from_pandapower import from_pandapower_net

# Load a built-in IEEE test case from pandapower
pp_net = pn.case9()

# Convert to a monee Network
net = from_pandapower_net(pp_net)

The returned Network is a standard monee electricity network. Run an energy flow immediately:

from monee import run_energy_flow

result = run_energy_flow(net)

What is preserved¶

pandapower element	monee equivalent
`bus`	`Bus` node (id, name, geodata)
`line`	`GenericPowerBranch` (r, x, b parameters, current limit)
`ext_grid`	`ExtPowerGrid` child
`gen` / `sgen`	`PowerGenerator` child
`load`	`PowerLoad` child (aggregated per bus, see below)

Node ids, names and positions¶

The MATPOWER format numbers buses from 1, so the monee node id is the pandapower bus index plus 1. Bus names and coordinates are copied onto the nodes when present. After conversion, node.name holds the pandapower bus name and node.position holds the (x, y) coordinates. Coordinates come from net.bus.geo (the GeoJSON Point column used by pandapower 3.x) or from the legacy net.bus_geodata frame on pandapower 2.x.

Note

pandapower 3.x models a transformer’s winding vector group as a branch phase shift (a multiple of 30°, for example Dyn5 gives 150°) and exports it through to_mpc; pandapower 2.x dropped it. For monee’s single-slack PQ power flow that shift is a pure reference rotation: it offsets all downstream bus angles by a constant and leaves every magnitude (voltages, branch flows, currents) unchanged. The importer zeroes these transformer phase shifts so the flat-start AC solve stays well-conditioned. Reported magnitudes still match pandapower’s; only the absolute downstream angle reference differs.

Line current limits¶

The intermediate MATPOWER round-trip drops current limits, so from_pandapower_net() recovers them afterwards from the original pandapower tables. Each line’s limit is rebuilt as max_i_ka * parallel * df and written back to the matching monee branch.

Note

Transformers deliberately keep a placeholder limit. monee’s single max_i_ka per branch cannot represent the different rated currents of the HV and LV sides at once, so any tight bound would make one side infeasible. Check transformer limits manually before running loading-constrained studies on a converted network.

Load aggregation and naming¶

All pandapower loads connected to the same bus are aggregated into a single PowerLoad. The aggregated child is named deterministically by joining the original pandapower load names with + (see aggregated_pp_load_name()), for example "LV1.101 Load 1+LV1.101 Load 2". This deterministic naming lets simbench timeseries match back to the aggregated loads by name.

Tip

After converting, you can save the result to the native OMEF format using monee.io.native.write_omef_network() so you do not need pandapower installed in future sessions. See Import MATPOWER case files for details.

SimBench¶

SimBench provides standardised benchmark grids with full-year load profiles. monee.io.from_simbench wraps the pandapower converter and turns the simbench profiles into a TimeseriesData ready for run_timeseries():

Function	Returns
`obtain_simbench_net(sb_code)`	`Network`
`obtain_simbench_profile(sb_code)`	`TimeseriesData`
`obtain_simbench_net_with_td(sb_code)`	`(Network, TimeseriesData)` from a single simbench fetch
`obtain_simbench_profile_by_pp_net(pp_net)`	`TimeseriesData` from an already loaded simbench pandapower net

Load profiles are scaled per load (base_p_mw * profile) and summed per bus, matching the per-bus PowerLoad aggregation described above. Each summed series is registered under the aggregated load name via add_child_series_by_name(), for both the p_mw and q_mvar attributes. Profiles of other element categories pass through under their raw simbench column names as p_mw series.

Feeding the result into a timeseries simulation takes four lines:

from monee import run_timeseries
from monee.io.from_simbench import obtain_simbench_net_with_td

net, td = obtain_simbench_net_with_td("1-LV-rural1--0-no_sw")
ts_result = run_timeseries(net, td, steps=96)

Note

SimBench profiles cover a full year at 15-minute resolution. Pass steps to simulate only the first part of the horizon, as above.

See Timeseries simulation for everything you can do with the resulting TimeseriesResult.