Power-system analysis in Julia: AC and DC power flow, quasi-static time series, WLS state estimation with tap estimation, IEC 60909 short circuit, scenarios and N-1 contingency; import of CGMES, MATPOWER, PowSyBl IIDM, DTF and SCF, export of CGMES, MATPOWER and SCF.
Nothing in Sparlectra is a black box: model construction, Jacobian assembly, PV/PQ switching and convergence are visible at runtime. Besides the rectangular Newton-Raphson solver there is a DC power flow and the analytic APSLF backend (AnalyticLoadFlow.jl). Runs are deterministic, configuration-driven and machine-readable. Suited for grid studies, solver development and teaching.
The full capability list is in the feature matrix.
The workshop notebooks run in Jupyter or VS Code. The Web UI installer has a
workshop mode (--workshops): one command installs Julia (if missing),
Sparlectra, IJulia and the notebooks into ~/sparlectra-workshops and
starts Jupyter.
Linux/macOS:
curl -fsSL https://raw.githubusercontent.com/Welthulk/Sparlectra.jl/main/tools/install_webui.sh | sh -s -- --workshopsWindows (PowerShell):
iwr -useb https://raw.githubusercontent.com/Welthulk/Sparlectra.jl/main/tools/install_webui.bat -OutFile install_webui.bat; .\install_webui.bat --workshopsOptions: --dir <folder> for another folder, --rev <branch> for
Sparlectra from a branch instead of the registered release, --no-start to
set up only. Without a Jupyter installation IJulia offers to install one. In
VS Code, open the folder and select the kernel
"Sparlectra workshops ( )": every workshop folder gets its
own kernel, named after the folder with a short hash of its path, and the
installer prints the name. Running the command again updates Sparlectra and
the notebooks; a notebook you changed is kept next to the new copy as
<name>.ipynb.<date>-<time>.bak.
The Web UI itself has no workshop button. If you already have the Web UI (a
copy from the one-line install or a git checkout), set up the workshops from
its folder; --path . takes Sparlectra from that folder, so the notebooks
match the version you run:
julia tools/install_workshops.jl --path .On Windows: julia tools\install_workshops.jl --path .. The options above
apply here as well.
| Notebook | File |
|---|---|
| Tour, basic: first network, model editing, Q-limits, slack types, OLTC and Q(U) control | workshop_tour.ipynb |
| Tour, advanced: remote voltage control, HVDC link, state estimation, FACTS incl. UPFC, N-1, threaded sweeps | workshop_tour_advanced.ipynb |
| Transformer control: parallel units, master/slave groups, CGMES RegulatingControl | workshop_tour_control.ipynb |
| Foreign formats: CGMES (anatomy of a delivery, bus-branch and node-breaker import, SV validation, export round trip), PowSyBl IIDM compared with OpenLoadFlow, a DTF deck with its outage records | workshop_tour_cgmes.ipynb |
| Slack types and short circuit: ideal, external grid, distributed slack, IEC 60909-0 currents | workshop_slack_short_circuit.ipynb |
| Distributed slack: participation weights, normalization, fallback | workshop_distributed_slack.ipynb |
| Transformer taps: OLTC, PST, X(α) characteristic, 3WT star equivalent | workshop_transformers.ipynb |
| FACTS flow control: TCSC to UPFC | workshop_series_compensation.ipynb |
| Scenarios and N-1 screening | workshop_scenarios.ipynb |
| QSTS: a day on sp_case14, snapshot run and sequential run with tap and bank controllers | workshop_qsts.ipynb |
| State estimation: measurement sets, observability, WLS, bad data | workshop_state_estimation.ipynb |
| SE bad data and parameter estimation: normalized residuals, sequential elimination, robust estimation, shunt estimation | workshop_se_diagnostics.ipynb |
| SE taps, phasors and topology: J and its degrees of freedom, tap estimation, PMU phasors, topology errors, islands | workshop_se_taps.ipynb |
| APSLF: analytic power-series solver on a Sparlectra model, convergence radius as a loadability margin, hybrid start | workshop_apslf.ipynb |
The notebooks are generated from docs/lit/ and are also part of the documentation.
Sparlectra is two packages in one repository: the library Sparlectra
(registered) and the application SparlectraApp under app/ (service API,
Web UI, sysimage build).
Library:
using Pkg
Pkg.add("Sparlectra")Application (Web UI), from a checkout:
git clone https://github.com/Welthulk/Sparlectra.jl.git Sparlectra
cd Sparlectra
julia --project=. start_webui.jlThe first start installs and compiles and reports the time. Scripts that use
the service API run with julia --project=app. Other REPL setups, sysimage
use and embedding: integration guide.
Installs Julia if missing, downloads the latest tagged release into Sparlectra/ in the current directory, and starts the Web UI.
Linux/macOS:
curl -fsSL https://raw.githubusercontent.com/Welthulk/Sparlectra.jl/main/tools/install_webui.sh | shWindows (PowerShell):
iwr -useb https://raw.githubusercontent.com/Welthulk/Sparlectra.jl/main/tools/install_webui.bat -OutFile install_webui.bat; .\install_webui.batImportant
This downloads and runs a script. If you prefer to read it first: install_webui.sh, install_webui.bat.
The installer offers a desktop shortcut; re-running it updates the copy.
Options and unattended installs: Web UI.
From a checkout, ./start_webui.sh or start_webui.bat starts the Web UI directly.
Julia compiles on first use; a sysimage removes that wait.
Every release ships an SPDX SBOM and SHA256SUMS; Pkg.add verifies the
registry tree hash. Details and security contact: SECURITY.md.
run_sparlectra runs the whole chain: import, configuration, control loops, solve, report.
using Sparlectra
# ensure_casefile downloads case14.m if it is not present locally
case_path = ensure_casefile("case14.m")
result = run_sparlectra(
casefile = basename(case_path),
path = dirname(case_path),
)
println(result.outcome)
println(result.iterations)
println(result.final_mismatch)A CGMES model: diagnose, import, solve, compare with the delivered SV profile. The example uses the CGMES demo files shipped with the package:
using Sparlectra
demo = joinpath(pkgdir(Sparlectra), "data", "cgmes_demo", "sp_case14")
summary = summarizeCGMES(path = demo)
result = importCGMES(path = demo)
runpf!(result.net, 30, 1e-8, 0)
cmp = compareWithSV(result)
@show cmp.max_dvmFor an ENTSO-E delivery of your own, path takes a folder, ZIP files or XML
files, for example path = ["grid.zip", "boundary.zip"].
Embedding Sparlectra in your own tooling (scripts, sysimage, API): integration guide.
Case management, run history and results in the browser. Cases can be taken from the local cache, uploaded (MATPOWER, DTF, CGMES ZIP, PowSyBl IIDM .xiidm) or fetched by name; three PowSyBl networks ship under data/powsybl. IIDM files are read in Julia, no Python is needed.
Every run file opens in the browser: CSV as a sortable table, Markdown rendered.
A QSTS run shows its controllers, switching counts and a time-series chart per element.
Full documentation: https://welthulk.github.io/Sparlectra.jl/
- Feature matrix
- Integration guide
- Web UI
- MATPOWER cases
- Solver guide
- State estimation
- Short circuit
- N-1 contingency analysis
- Function reference
- Changelog
Bug reports, test networks and import edge cases (CGMES, MATPOWER, DTF, PowSyBl IIDM) are welcome. Questions: Discussions. Please read CONTRIBUTING.md and CODE_OF_CONDUCT.md first.
The development of Sparlectra.jl is supported by SOPTIM AG, Aachen, Germany.
@software{sparlectra_jl,
title = {Sparlectra.jl: A Power-Flow and State-Estimation Framework in Julia},
author = {Schmitz, Udo},
year = {2026},
url = {https://github.com/Welthulk/Sparlectra.jl}
}Apache License 2.0, see LICENSE.




