Logiciel Contact Us
Success Stories Tech News Contact Us
whitepaper

How an Energy Utility Built Grid-Trustable AI for Anomaly Detection.

An AI reliability playbook for VPs of Operations responsible for grid signal anomaly detection.

In depth

Your anomaly detection alerts too much.

01

Anomaly detection on the grid is an unusually hard machine learning problem.

02

Signal quality is the underrated half of the problem.

03

Alert calibration is the other half.

The detail

The 10-week program that gets you there.

Zone · 01

Weeks 1–3 - Signal quality before model quality

We spend the first 25 percent of every anomaly detection program on the data pipeline. Sensor health, missing data handling, time alignment across systems, label quality.

Zone · 02

Weeks 4–7 - Alert calibration with explicit precision-recall trade-off

We make the trade-off explicit. The operations team chooses where on the precision-recall curve they want to live.

Zone · 03

Weeks 8–10 - Operator-in-the-loop tuning

Operators tag every alert as true positive, false positive, or 'don't know yet.' The tagged data is the most valuable training signal in the system. We make tagging easy - one click in the existing operator console.

By the numbers

The figures that make it a board-level conversation.

78%
Daily alert volume - reduction
55 ppt
Precision (true positive rate) - +
6 hours
Average lead time on incipient failures
Inside the report

What you'll take away.

01

Signal quality before model quality

We spend the first 25 percent of every anomaly detection program on the data pipeline.

02

Alert calibration with explicit precision-recall trade-off

We make the trade-off explicit.

03

Operator-in-the-loop tuning

Operators tag every alert as true positive, false positive, or 'don't know yet.' The tagged data is the most valuable training signal in the system.

Questions

Frequently asked.

Why is alert volume the most important metric?
What about completely new failure modes?
How does this fit with NERC CIP and FERC?
Get the whitepaper

Have it emailed to you.

Drop your details and we'll send How an Energy Utility Built Grid-Trustable AI for Anomaly Detection straight to your inbox - no spam, unsubscribe anytime.

Download whitepaper
Next step

Alerts the operators act on, with a precision rate that reflects real anomalies.

Talk through how this applies to your roadmap with our engineering leads - a working session, not a sales pitch.

Download White Paper