All customer storiesProof of concept
Probabilistic grid load forecasting with an approval gate: proof of concept.
An energy operator dispatched against a single-point spreadsheet forecast, with no uncertainty range and no record of which forecast a decision rested on. A proof of concept demonstrated P10, P50 and P90 load forecasts, drift checks, a decision log and an approval gate in the tested workflow.
- Client
- An energy operator
- Sector
- Energy, grid operations
- Engagement
- Proof of concept
- Validated result
- P10 / P50 / P90probabilistic forecasts
P10 / P50 / P90probabilistic forecasts
The full account
01One number and no record
With more variable generation and demand on the grid, dispatch teams answer for both stability and cost. They worked from one spreadsheet forecast per interval, plus each operator's own buffer. There was no low or high case, no warning when the model drifted, and no record of which forecast a decision rested on.
02Forecast ranges, drift checks and an approval gate
Forecasts give P10, P50 and P90 load for each interval, so operators see the low, central and high cases. Realised load should stay under P90 in about nine intervals out of ten. Drift checks compare forecasts with realised load and flag the model for review when error leaves agreed bounds. An assistant beside the existing dispatch tools assembles each proposal but cannot dispatch: a named person approves it at a recorded gate.
03What the proof of concept showed
In the tested workflow, forecast ranges, drift checks, the decision log and the approval gate worked together. Each decision was logged with the forecast shown, the action and the approver, so any past day can be reviewed. The planning team now uses the forecast ranges in its weekly planning, with an engineer signing off each change. It is not connected to live dispatch systems; that step would connect live grid data, run alongside the current process first, and keep the gate and the log as fixed requirements.