4 hoursper plant, down from about 120
The problem
NERC's cold-weather standard EOP-012 asks every generator owner to calculate an extreme cold weather temperature for each plant from years of hourly weather data. Today an engineer assembles it by hand over several weeks, filling gaps in station data, and the result is hard to reproduce for an auditor.
What we ran
No network simulation is needed. The test is a plant's own documents: the agent must reproduce the number the plant engineer calculated by hand, from public weather station data.
What the agent did
It pulls hourly temperatures from nearby public weather stations, normalizes timestamps, fills gaps with a documented method, computes the required percentile of winter hours, and assembles the evidence package. The compliance engineer reviews and signs.
Results
| Measure | Today | With Evardi |
|---|---|---|
| Engineer hours per plant per calculation | About 120 | About 4 |
| Same inputs give the same answer | Depends on who ran it | Every run |
| Gaps and fills documented for the auditor | Partly | Every one |
What it is worth
- Per plant: engineer hours saved × loaded engineer hour
- Fleet: the same, × plants per update cycle
- Penalty exposure counted only at expected value, never the maximum
* Study simulated with partners on public data.
Run it on your network
Plant locations and the documents behind the last hand calculation. We return the reproduced temperature with its full evidence trail.
Full study
Request the full study.
The complete write-up of “An auditable extreme-cold temperature in hours instead of weeks”: method, data, results and assumptions, as a PDF. Leave your work email and we'll send it.
