Validate Ecm Endpoint
Validate a fitted ECM on held-out data (synchronous, no fit, no pipeline).
Re-simulates the fitted model forward on the held-out current using the same
engine POST /ecm/fit runs internally, then returns model-vs-data traces
and RMSE/MAE/Max voltage error.
Body
Request body for validating a fitted ECM against a held-out trace.
Provide exactly one held-out data source (measurement_id or
example_id) and exactly one model source (fit_results or
parameterized_model_id). The model is re-simulated forward on the
held-out current — no fit and no pipeline — and compared to the held-out
voltage.
Held-out cell measurement to validate against.
Built-in example dataset id to validate against.
Optional inclusive start step (measurement only).
Optional inclusive end step (measurement only).
Known SOC (0–1) at the start of the held-out trace. When omitted it is recovered from the trace's first voltage via the fitted OCV(SOC) curve (assumes the trace starts near rest).
0 <= x <= 1Cell capacity [Ah] for the SOC integration. Defaults to the fit/model capacity (the SOC reference the curves were fitted against); it is never re-estimated from the held-out trace.
x > 0In-memory fit results to re-simulate.
Saved ECM parameterized model to re-simulate.
Response
Successful Response
Held-out validation result: error metrics + downsampled overlay traces.
Root-mean-square voltage error [mV].
Mean absolute voltage error [mV].
Maximum absolute voltage error [mV].
RC-pair count of the validated model.
Capacity used for SOC integration [Ah].
Initial SOC used for the held-out trace.
Where the validated model came from.
Downsampled time grid [s].
Held-out measured voltage [V].
Re-simulated model voltage [V].
Model − data voltage residual [mV].