Trust the numbers, not the pitch

Methodology

PsFlow is benchmarked on open synthetic networks — the Texas A&M ACTIVSg cases, including the 2,000-bus Texas-footprint and larger Eastern-Interconnection-scale systems — so that every result can be independently reproduced without access to confidential (CEII) data.

For each benchmark case we publish:

• Agreement tables — post-contingency flows, violation sets, and transfer limits compared element-by-element against established industry reference tools, with error distributions rather than cherry-picked examples.
• Binding-constraint match rates — whether PsFlow identifies the same limiting elements a reference study identifies, which is the result screening decisions actually turn on.
• Runtime scaling — wall-clock performance from a single core to a full multi-core batch, on documented hardware.

Why public cases

Real network models are restricted data, and vendors who benchmark only on private cases are asking you to take accuracy on faith. Synthetic grids remove that excuse: the inputs are downloadable, the scripts are shareable, and any engineer can rerun our numbers. When you evaluate PsFlow on your own models — under your own data agreements, in your own environment — you'll already know what the engine does on cases you can inspect.

Evaluate it on your own study

The strongest validation is your last study, rerun. Pilot evaluations start by reproducing a screening study your team has already completed, so the comparison baseline is work you trust. Contact us to set one up.