Tighten the gap between assumed yield and actual.
Combine on-site pyranometer and reference-cell data with mesoscale and satellite-derived irradiance to verify yield assumptions, validate developer forecasts, and tighten dispatch.
The challenge
Most PV plants start operating on a P50 yield assumption that hasn't been re-verified since financial close. The on-site pyranometers and reference cells that could improve the picture publish to a separate system. The trading desk submits day-ahead nominations off a third-party forecast that nobody benchmarks. Imbalance penalties are the result.
What success looks like
POA
Forecasts and resource estimates at the plane of array, not just GHI
Bias-corrected
Mesoscale forecast continuously corrected against on-site measurement history
Audited
Daily forecast-vs-actual scorecard so the value is provable to the trading desk
How SolarTwin delivers it
- Mesoscale + satellite ingest — ECMWF, CAMS, and Meteosat/GOES nowcasts, with POA interpolation per tracker geometry.
- Pyranometer & reference-cell fusion — on-site measurements weighted into the short-term horizon where they outperform the mesoscale.
- Long-term resource statistics — rolling 12-month POA distribution, clear-sky vs. measured ratio, capacity-factor by hour and month, against the P50 you bought the plant on.
- Power-output forecasts derived from your plant model plus live availability.
- Daily forecast-vs-actual scorecard — MAE, RMSE, and bias by horizon and inverter, so the trading desk can audit the platform.
Where this lands first
- Imbalance-exposed portfolios — even small forecast accuracy gains pay back quickly.
- Refinancing decisions — a credible long-term resource view changes lending terms.
- Curtailment-heavy regions — knowing when curtailment is likely lets you schedule maintenance into it.
What you need to start
- At least one on-site pyranometer or reference-cell stream we can ingest.
- A history of past nominations vs. actual production for the bias backtest.
- Plant model parameters (DC capacity, AC capacity, tracker geometry, module temp coefficients).
See it backtested on your own data
Send us a month of measured POA and a month of past forecasts. We'll show you what SolarTwin would have done with them.
Book a walkthrough