Fewer extension dayson a generating unit, priced at its replacement power
The problem
A municipal utility that owns generation pays twice when a planned outage goes wrong: once in crews standing by, and again in power bought to replace the unit. Outages get denied or extended over defects that were in the request all along, or because conditions changed after approval and nobody re-studied.
What we ran
A transmission-level model with an outage queue, planted card defects, and a load forecast that moves day to day between approval and the outage date.
What the agent did
Before submission it checks each outage request for switch statuses including closures, equipment IDs against the model, and overlaps. It applies the utility's own approval rules, re-studies any approved outage whose date, forecast or overlaps changed, and proposes only mitigations operators already use: a new window or a load transfer. Contingency analysis stays in the tools the utility already runs.
Results
| Measure | Today | With Evardi |
|---|---|---|
| Card defects caught before submission | Found by the reviewer, or after denial | Flagged at drafting |
| Approved outages re-studied when conditions change | When someone notices | Every one, automatically |
| Extension or deferral days on the unit per year | Baseline | 5 fewer |
What it is worth
- Utility cash: extension days avoided × unit size × 24 hours × gap between market price and the unit's own cost
* Study simulated with partners on public data.
Run it on your network
A year of outage requests with their outcomes, and the unit's capacity and typical replacement price. We return which past requests the agent would have flagged, and when.
Book a demoSee Outage requests on the platform
Full study
Request the full study.
The complete write-up of “Fewer extension days on a generating unit, priced at its replacement power”: method, data, results and assumptions, as a PDF. Leave your work email and we'll send it.
