- Home
- Documentation
- Time series
Time Series Simulation
pandapower
Time Series runs one AC load flow per hour (or time step). At each step, load, generation and battery power follow the profiles you enter, so you can see voltages and loadings over a day or a year. Time Series Playback then plays the whole run on the single-line diagram and exports it as a video.
Simulate → Time Series Simulation (pandapower).
When to use
- Daily or yearly voltage and loading with varying load and generation
- Checking whether a snapshot load flow hides an overloaded hour
- Showing how a battery shifts midday solar to the evening peak
How to run
- Build the network and confirm a single Load Flow converges.
- Open Time Series Simulation.
- Set the horizon (for example 24 hours) and a profile for each load, generator and battery.
- Run and review charts in the results dialog; export if you need a spreadsheet.
- Click Play on diagram to watch the run on the single-line diagram, and Export video to save it (see Time Series Playback).
How this relates to pandapower time series
Each step updates P/Q from your profiles and solves AC load flow. The idea matches the pandapower time series tutorial, but Electrisim drives the loop itself rather than the pandapower OutputWriter pipeline.
Dialog — simulation horizon
time_steps— number of sequential power-flow solves. Each step represents one hour by default (e.g. 24 = one day, 8760 = one year).
Dialog — load & generation profiles (per element)
Every Load, Generator, Static Generator and Storage (battery) on the diagram gets its own profile editor. Enter one numeric value per time step, separated by commas or spaces. You can also import a .csv or .txt file.
Batteries take absolute power in MW: positive charges, negative discharges (pandapower convention). Charge midday, discharge evening fills a solar-shifting day at the battery’s rated power (max_p_mw, otherwise the larger of sn_mva and p_mw): charging from 10:00 to 15:00, discharging from 17:00 to 21:00, with about 85 % of the charged energy coming back out. Idle (0 MW) switches a battery off for the whole run, which is handy for a with/without comparison. A battery without a profile keeps its diagram set point. State of charge and energy limits are not enforced in the run, so keep the profile within the battery’s energy.
| Profile mode | Meaning | When to use |
|---|---|---|
| Absolute P (MW) | Sets active power directly in MW at each hour (Q scales proportionally unless overridden). | Recommended for tutorial-style networks and when base-case P from the diagram should be replaced entirely. |
| Scale factor (× base P) | Multiplies the element’s base p_mw from the diagram (e.g. 0.8 → 80% of rated P). |
Use with built-in preset shapes (daily load, solar, wind) via Quick fill or Fill preset. |
Quick fill (optional toolbar above the element cards) lets you apply a preset shape to all loads or all generators at once — this replaces the older “Global preset” mode, which applied one profile to every element and often failed to converge on meshed or transformer-heavy networks. Per-element control with absolute MW values is the recommended approach.
Preset shapes available for quick fill:
- Load:
constant,daily(residential),industrial,variable - Generation:
constant,solar,wind,variable
The Random (tutorial) button fills absolute MW values between 0 and the element’s base P — matching the pandapower notebook example (for batteries: between full discharge and full charge).
Scale factors are used exactly as entered. Only values above 3 are rejected in Scale factor mode, because they are almost certainly MW; switch the element to Absolute P (MW) for those.
Dialog — advanced power flow settings
Collapsed under Advanced power flow settings in the dialog:
frequency— system frequency (Hz).algorithm—nr(Newton-Raphson),iwamoto_nr,fastdecoupled, ordc.calculate_voltage_angles—auto, yes, or no.init— initialization method. Auto is recommended; after the first converged step the backend warm-starts from previous results.
Convergence tips
- Verify the network converges in a single Load Flow before running time series.
- Use Absolute P (MW) when values look like notebook MW numbers — do not mix absolute MW values with scale-factor mode.
- Keep initialization on Auto unless you have a specific reason to change it.
- Large simultaneous swings in load and generation can stress weak networks; try smoother profiles first.
Results dialog
After a successful run, the results dialog shows:
- Summary — convergence status, duration, element counts, and non-converged step numbers if any.
- Charts — input profiles, bus voltages, line loading, load P, static generator P.
- Statistics — min / max / average voltage and line loading across all hours.
- Detailed data — scrollable table of bus results for every hour.
Enable Download Excel report automatically in the setup dialog, or click Export to Excel in the results dialog. Sheets include Summary, res_bus, res_line, res_load, res_sgen, and profiles. If the Excel library cannot load, a CSV fallback is offered.
Time Series Playback
Click Play on diagram in the results dialog. The dialog closes and the run plays on the single-line diagram, hour by hour:
- Lines and transformers take the loading colour (green to red at 100 %, the same scale as the load-flow arrows). Lines at 90 % or more and every transformer show their loading.
- Buses take the voltage colour (green at 1.0 pu to red at ±0.1 pu).
- Flow dots run along each line in the direction of active power and move faster as the flow grows, so a reversal (for example a feeder exporting at midday) is easy to see.
- Labels show generator MW, battery MW with charging or discharging, and grid import or export.
The control bar under the diagram has play/pause, step back/forward, a time slider, speed (0.5× to 4×), Loop, and the Flow and Labels switches. Its chart shows total generation, load, battery power and grid import over the run; click or drag on it to jump to a time. Next to it: the highest loading, the voltage range, and the totals at the current step. Results reopens the results dialog; × closes the playback. Between hours the colours and values are interpolated, so the animation runs smoothly.
Export video
Export video records the playback with a title, the clock, key numbers and the totals chart, in one of three formats: 16:9 (1920 × 1080), 4:5 for the LinkedIn or Instagram feed (1080 × 1350), or 1:1 square (1080 × 1080), at about 12, 20 or 30 seconds. Chrome and Edge save MP4; browsers without MP4 recording save WebM. The video records in real time, so keep the tab open until it finishes. Runs longer than 72 steps export the day you are viewing.
Limits: Playback needs a run made after November 2026: older results have no transformer, battery or grid rows, so only buses and lines are coloured. Very large networks (more than 800 lines) start with flow dots off; switch them on in the bar if the browser keeps up.
API payload
Posted as TimeSeriesSimulationPandaPower Parameters together with the diagram JSON:
time_steps,profile_mode(alwayscustomfrom the UI),element_profiles— per-element{ element_type, mode, values, display_name }.element_typeisload,sgen,genorstorage. An explicitmode(absoluteorscale) is used as sent; without one, the backend guesses from the values.load_profile,generation_profile— preset names used only for quick-fill reference.frequency,algorithm,calculate_voltage_angles,init.
Outputs (JSON)
| Key | Meaning |
|---|---|
timeseries_converged |
true only if every time step converged. |
time_steps |
Requested horizon (hours). |
profiles_used |
Per-element profiles actually applied, including display_name for chart labels. |
busbars |
{ name, id, time_step, vm_pu, va_degree, p_mw, q_mvar } for each bus and hour. |
lines |
{ name, id, time_step, loading_percent, p_from_mw, p_to_mw, from_bus, to_bus }. from_bus / to_bus are the bus ids, so a positive p_from_mw flows from from_bus to to_bus. |
loads, sgens, gens |
Simulated P/Q per element and hour (gens also vm_pu). |
transformers, transformers3w |
{ name, id, time_step, loading_percent, p_hv_mw } per transformer and hour. |
storages |
{ name, id, time_step, p_mw, q_mvar }; positive p_mw is charging. |
externalgrids |
{ name, id, time_step, p_mw, q_mvar }; positive p_mw is import from the grid. |
time_step_converged |
One true/false per step. |
voltage_statistics, loading_statistics |
Per-bus / per-line min, max, average across steps. |
time_stamps |
Synthetic timestamps, hourly from 2024-01-01 00:00 (runs longer than 24 steps continue into the next days). |
Extraction tip: Filter busbars / lines by time_step to plot trajectories; use the statistics blocks for KPI cards.
pandapower alignment
Official reference: Time Series Simulation. Electrisim approximates the profile-driven loop using repeated runpp calls, which is appropriate when you need bus/line summaries and Excel export rather than registered OutputWriters.
Last updated: November 2026