First study

One case. One version. Visible limits.

This walkthrough describes the supported route; it does not represent a completed customer installation or engineer-approved result.

Run the exact five-bus sample

The downloadable file is a sanitized derivative of the reviewed synthetic fixture. Only its personal user_name metadata was cleared; electrical values and every other archive member are byte-identical to the source fixture.

Download the native sample Read the public proof record

Verify the sample file SHA-256 before opening: 6741afaabf9ab0190cbec2d05d7cfa6521cfeaa9fe0e6f4d2530fe7e51a978d3. The archive contains five buses, seven lines, two generators and three loads. It is educational material, not an installer or qualified design.

Open the file in a compatible PowerSystem desktop. Open the downloaded .veragrid in the native project window. Keep the original sample unchanged and use a disposable project copy.
Inspect exact inputs in Plan. Ask the native study tools to run study_equipment for the summary and confirm the returned artifact SHA-256 matches the downloaded file. Review the buses, branches, loads and generators; do not edit or repair values.
Save the immutable case in Build. Use study_case_manage with operation="save", label Public synthetic base case, and request {"kind":"power-flow","expected_artifact":"<exact discovered SHA-256>"}. Use a fresh request key for submission.
Submit that exact case version. Use study_case_submit with the saved case_id and version; do not substitute a newer or guessed version. Poll study_status until a terminal status.
Read the retained result. Use study_runs and study_results for the returned receipt. Expect a balanced AC base-case run only; no N-1 is implied. Compare nothing unless a second compatible run exists.

The balanced AC base case converged. Bus voltage range: 1.000000–1.010000 pu. Active losses: 0.119856 MW. Independent line-pi nodal balance error: 1.86e-12 MVA (limit 0.0001 MVA). No N-1 study or engineering approval is implied. This result is reproducibility evidence for the synthetic case, not independent validation, an engineering recommendation, or human approval. If no compatible customer package/account is available, the sample cannot be run through an unavailable service.

Open a copy of your native project. Preserve stable IDs, diagrams, scenarios and original input values. Investigate data warnings rather than asking AI to repair them silently.
Inspect the exact equipment and units. Confirm load/generator P in MW, Q in MVAr, voltage setpoint in pu and continuous branch rating in MVA. Record source or synthetic assumptions beside the candidate.
Save a named case and version. Select one snapshot/scenario. Changed inputs produce a new immutable version; a prior receipt is not reused as proof for changed data.
Choose an admitted profile. Balanced AC power flow examines one exact saved snapshot. Select outages separately if screening is required; power flow never adds N-1 implicitly.
Submit and wait for the worker. Read status, warnings and the solver/configuration identities. A failed, islanded or unsupported study remains explicit instead of being converted to a zero result.
Review measured results. Inspect bus voltages, branch flows, losses and any limit violations in the native result view. Compare only compatible cases and independently verify answers within the admitted profile.
Retain evidence, then decide. Reopen the saved project and receipt. An AI response or completed job does not accept a candidate, approve engineering or Apply to Original.

Read study qualifications · When a run is unavailable