Catch underperforming strings before the monthly PR report does.
Real-time PR monitoring, irradiance-conditioned baselines, soiling-rate tracking, and rapid root-cause triage combine to recover kWh that would otherwise stay lost.
The challenge
A 1–3% PR deviation across a 100 MWp plant is millions of euros of lost revenue a year — and it rarely shows up in a standard availability report, because the inverter is technically "available". By the time it surfaces in a monthly performance review, the deviation has been bleeding for weeks. Soiling losses, slow connector degradation, and string-level drift all hide in the noise.
The data needed to catch them lives in the same SunSpec register map your CMMS imports. The reason most operators don't catch them earlier is that nobody is conditioning the comparison on irradiance, module temperature, and clipping state continuously.
What success looks like
How SolarTwin delivers it
- String-level monitoring — DC voltage and current per string, AC power per inverter, POA irradiance, and module temperature. The deviation you care about is computed on every reading, not on monthly aggregates.
- Irradiance-conditioned PR — compared to the plant model with corrections for module temperature, POA versus GHI, and spectral mismatch.
- Soiling-rate estimation — derived from peer-string and reference-cell comparison; quantifies the marginal value of the next cleaning.
- Clipping-aware — inverters at the DC/AC inverter cap are excluded from the comparison instead of dragging the score down.
- Multi-condition alarms — "PR deviation > 3% conditioned on irradiance 600–900 W/m² for > 6 hours" instead of single-threshold rules.
- Tickets & mobile capture — the deviation turns into a dispatched job; the crew records what they found; the closing report attaches to the asset.
A typical workflow
- A string drifts below its peer group by more than 5% in the 600–900 W/m² irradiance bin over the rolling 6 hours.
- SolarTwin checks for shading, clipping, module-temperature anomalies; the alarm carries that context.
- The performance engineer opens the twin, reviews recent IV-curve history, and decides whether it's soiling, a diode, a connector, or a module-string issue.
- A work order goes out. The crew's closing report attaches to the string and the campaign.
Where this lands first
- Soiling-prone sites — most measurable PR recovery; lowest-effort win.
- Post-warranty plants — proof of underperformance has direct commercial value at end-of-warranty.
- Mid-life plants — most likely to have started drifting from their commissioning baseline.
What you need to start
- SCADA history at 15-minute (or higher) resolution — string current/voltage, AC power, POA irradiance, module temp, clipping state.
- The plant's commissioning PR baseline.
- Curtailment / dispatch log so we can exclude those periods cleanly.
Bring us a month of SCADA data
Send us 15-minute SCADA for one plant and the commissioning PR. We'll show what SolarTwin would have flagged across the period.
Book a walkthrough