Utilities

How do you improve SAIDI and SAIFI, and what data does it take?

Reliability indices are reported to regulators and argued in rate cases. Moving them takes data from systems that rarely talk to each other.

June 20264 min read

The indices, and what actually moves them

SAIFI and SAIDI, the frequency and duration of customer interruptions, are the numbers a utility answers for. They improve when outages are prevented, or restored faster, which depends on seeing the risk before it becomes an interruption.

Prediction needs the systems fused, not a single feed

Outage risk lives across many systems: the asset register, SCADA, the outage and distribution management systems, GIS, weather, and vegetation data. The value is in the correlation, not any one source.

Vegetation contact and asset condition are consistently among the top drivers, which is why trimming is shifting from a fixed cycle to risk-based, guided by imagery and AI.

Where the ERP closes the loop

A predicted feeder or transformer risk only helps when it becomes a prioritised work order with the right crew, parts, and switching plan. On Hudace the asset register and work management sit together, so risk scores drive the schedule and the evidence for NERC compliance is captured as the work is done.

Xenon AI prioritises the backlog by risk. The field crew and the regulatory obligation remain where they belong.

The numbers to watch

These are reported to regulators, so the baseline already exists.

SAIFI

SAIFI = total customer interruptions / total customers served. How often the average customer loses power.

SAIDI

SAIDI = total customer-minutes interrupted / total customers served. How long, in total, they are out.

CAIDI

CAIDI = SAIDI / SAIFI. Average restoration time per interruption.

Avoided outage minutes

The leading indicator that moves before the annual index prints.

See reliability on Hudace

Talk to our team about turning outage risk into prioritised, audit-ready work.

Request a demo