Load-step response
How quickly does PCC power follow a change from 0.5 to 0.6 pu?
Run an equilibrated 21-state converter chain. Compare the requested load, PCC response and DC-link voltage.
Advanced power dynamics / five interactive studies
Follow a load change through the power system
Explore converter response, oscillation modes and interacting grid ports with the supplied Python study equations running in your browser.
Choose a question, predict, then run
Modes and frequency response
What changes when the voltage-loop bandwidth changes?
Inspect eigenvalues, damping, dominant state participation and a sampled small-signal frequency response.
Synthetic load spectrum
Which part of a repeating load swing reaches the PCC?
Apply a bounded sinusoidal load and compare its time response and spectrum. The input is explicitly synthetic.
Phase model versus QSS
What does the reduced formulation leave out?
Compare the supplied 41-state abc and 21-state quasi-steady formulations, including their hand-set startup and a shared step at 0.04 s in a bounded 0.08 s window.
Interacting grid ports
Does a solved power flow imply stable dynamics?
Explore a modified nine-bus network with synchronous-machine, grid-forming, grid-following and data-center ports.
Read the system at two timescales
The twelve power-continuity lessons account for energy during outages. These advanced studies investigate the dynamics of idealized converter and grid models. For an introductory electromagnetic transient, begin with the separate DC-link RLC voltage-sag lesson.
What a mode tells you
For a linearized mode with eigenvalue λ = σ + jω, σ below zero indicates decay around that operating point. A positive σ indicates growth in the linearized model. Its frequency is |ω|/(2π) Hz; damping ratio is −σ/√(σ² + ω²). A converged power flow and stable dynamic modes answer different questions.
Run one controlled comparison
- Select the load-step study and write a prediction before running.
- Keep the starting settings, run, and inspect the recorded solver diagnostics.
- Change only the DC capacitance from 2 to 3 pu. Run again and compare the transient shapes.
- Export each applied result. Draft input edits are not part of an earlier result.
Runtime and reproducibility
The scientific runtime downloads only when you choose Run. Computation stays in a cancellable browser worker; there is no shared simulation backend or usage telemetry. The simpler course and DC-link study retain their lightweight JavaScript first calculation.
python -m pip install -r requirements-studies.lock python -m datacenter_twin.research --study load-step --output outputs/load-step.json
Use a current checkout containing these studies. The historical v1.0.0 assets predate this addition. Results record the applied configuration, solver settings, assumptions, chart data and diagnostics. Per-unit values are normalized on the documented model bases, not automatically MW.
Provenance and interpretation
The owner supplied Python equivalents of MATLAB studies. The original MATLAB files and measured GPU trace were absent, so no MATLAB equivalence or measured-workload reproduction is claimed. The spectrum lesson uses a synthetic sinusoid. The phase-model comparison uses documented stiff integration instead of the original explicit Heun update. The nine-bus case is modified, not an unaltered standard benchmark. These are software teaching studies, not equipment or facility validation.