← Back to blog

Loss Attribution: Decomposing the Gap Between Expected and Actual Generation

"We generated less than expected" is the most expensive sentence in solar O&M. Loss attribution turns it into a contract-grade waterfall.

N
Lead AI Engineer · February 10, 2026 · 10 min read

The Operator's Real Question

Operators do not actually want a PR number. They want to answer: "Of the kWh I expected to generate today, where did the missing kWh go — and which ones can I get back?"

The Loss Waterfall

Enlyra's Loss Attribution engine decomposes the Expected vs Actual gap into deterministic buckets:

  • Curtailment: Grid-imposed clipping at the export meter, computed against allowed export.
  • Inverter downtime: Time-weighted unavailability per inverter from heartbeat and status registers.
  • Soiling: Modelled from short-term PR drift between cleanings, validated by reference cell where available.
  • MPPT loss: DC vs AC efficiency drift relative to inverter datasheet curves.
  • Weather (low POA): Below-expected irradiance days, isolated from controllable losses.
  • Inverter underperformance: Residual after the above, traced to specific inverters or strings.

Why Buckets Matter

A 4% PR shortfall in March looks identical on a dashboard whether it came from grid curtailment (recoverable via grid coordination), soiling (recoverable via cleaning) or inverter underperformance (recoverable via repair). Loss Attribution says exactly which one — and ties each kWh to a rupee or dollar at the contracted tariff.

From Number to Action

Every Enlyra Loss Attribution row is actionable: it links to the affected device, the time window, the analytic trace, and the recommended O&M ticket. Plant heads stop arguing about PR and start recovering kWh.

Ready to see Enlyra in action?

Start your free trial and turn every conversation into revenue.