Performance Guarantee Dispute After Irradiance Drift
Your firm operates a 1.62 MWDC / 1.25 MWAC community solar array under a five-year O&M agreement. The contract guarantees a weather-adjusted performance ratio (PR) of 80.0%, measured monthly using the site's secondary-standard pyranometer and a reference cell. For June, the monitoring portal reports PR of 74.6%; at the guarantee threshold the array would have needed 1,308 MWh, and it produced 1,219 MWh. Weather data from a nearby airport station suggests irradiance was about 3% lower than the on-site sensor recorded. The pyranometer's last calibration was 26 months ago, and its dome shows visible dust and a bird-dropping streak. Inverter 2 logged 14 hours of grid-fault downtime during a heat wave. The asset manager wants to file a warranty claim against the EPC by Friday. You have string-level current data, inverter logs, and the irradiance time series.
1. Before the asset manager files the claim, how do you determine whether the 5.4-point PR shortfall is a real production loss or a measurement artifact?
Read the worked answer
Start with the measurement chain, not the modules. A dirty dome or a drifting pyranometer biases the denominator of PR. Here the airport station reads about 3% lower irradiance than the on-site sensor, which suggests the on-site sensor may be over-reading; that alone would push a true 77.5% PR down to roughly 74.7% on paper, explaining most of the gap.
Pay attention to the direction of the bias: a soiled dome usually under-reads irradiance and would make PR look better, so dust alone cannot explain a low PR - calibration drift, leveling error, or a scaling error in the logger can. Request the sensor's calibration certificate and drift report, clean the dome, compare the pyranometer against the reference cell and against the airport data over the same daylight hours, and verify the irradiance scaling in the data logger.
In parallel, confirm array health from string currents and inverter DC/AC ratios so a real fault is not masked by a sensor argument. Only after the measurement chain is validated can you decide whether a claim is supportable.
2. If the sensor proves accurate, how should the 14-hour grid-fault outage be handled in the guarantee calculation?
Read the worked answer
Treat the outage as a separate, quantified adjustment rather than as an excuse. First establish cause from inverter logs and utility outage records: a grid fault is normally an excused event, while an inverter trip caused by maintenance negligence is not.
Estimate the energy that would have been produced during those 14 hours - either by applying the measured plane-of-array irradiance to the inverter's expected efficiency, or by using the sibling inverters' output as a reference. Then either add that estimate to actual production or exclude the corresponding irradiance from the PR calculation. Do not do both, or you double-count the adjustment.
Document the method, data sources, and result in the claim file. If the contract also has an availability guarantee, confirm that the same hours are not being counted under both the PR and availability provisions.
3. What operational and documentation changes should be put in place regardless of the claim outcome?
Read the worked answer
The evidentiary weakness is the calibration interval and the lack of redundancy. Move pyranometer calibration to an annual or at least two-year cycle consistent with IEC 61724-1 and the contract, and add a reference cell for cross-checking. Separate sensor cleaning from module cleaning so the sensor is not cleaned only when the array is washed.
Flag downtime, calibration events, and sensor maintenance in the monitoring database so future PR reports carry data-quality annotations. Keep calibration certificates with the monthly reports and review who is responsible for sensor upkeep under the O&M agreement. If the contract's data-quality clause requires calibrated instruments, an out-of-calibration sensor weakens the claim even when the array is healthy.