Predictive maintenance on the distribution grid: from asset health to the capex plan

IEEE 1366 defines the distribution reliability indices most regulators use, including SAIFI (how often the average customer is interrupted) and SAIDI (for how long) ^[1]. Any maintenance programme should be able to say which interruptions it is expected to prevent and what that does to these numbers.

• Transformer condition: dissolved gas analysis of insulating oil detects overheating and arcing well before failure; IEEE C57.104 sets out how to interpret it ^[2].
• Loading: transformers that run above rating age faster; per-transformer load forecasts show which ones will cross it and when.
• Vegetation: growth into line corridors is a leading cause of outages, and aerial inspection can find it span by span.
• History: past interruptions, protection operations and work orders on the same asset or feeder.

A risk score is not a plan. Each candidate intervention (replace, augment, reconductor, add a sectionaliser, trim) has a cost and an expected effect on reliability and capacity. The decision is which set to fund this year within the budget, which is an optimisation problem, not a ranking.

Engineers will only trust the plan if the network analysis is checkable. EPRI's OpenDSS is an open-source distribution system simulator widely used as a reference ^[3]; cross-checking against it makes the numbers defensible to a regulator.

Found this useful?

Share the key finding on X.