What Is Solar PV Performance Ratio (PR)? A Guide to Calculation, Interpretation, and Troubleshooting

Energy production alone does not tell you how well a solar plant performed. A plant that exported 1,440 MWh in one month may have performed exceptionally or poorly; the answer depends on its nameplate capacity, available irradiance, temperature, availability, and the selected measurement boundary. Performance Ratio (PR) normalizes actual production against the available solar resource and plant size to support a more meaningful comparison.

PR is not an absolute health score. It summarizes the combined effect of losses and operating conditions, and it is trustworthy only when the data, time interval, measurement boundary, and exclusion rules are clearly defined.

What exactly does PR measure?

In the commonly used IEC 61724 definition, PR is the ratio of final system yield to reference yield. Final yield expresses measured AC energy relative to the plant's rated DC power. Reference yield converts in-plane irradiation into equivalent hours at reference irradiance.^[1]

Y_f = E_AC ÷ P_STC

Y_r = H_POA ÷ G_STC

PR = Y_f ÷ Y_r = E_AC ÷ (P_STC × H_POA ÷ G_STC)

• E_AC: AC energy measured at the selected boundary, commonly the revenue meter or inverter outputs.
• P_STC: the array's rated DC power under standard test conditions.
• H_POA: accumulated plane-of-array irradiation, not simply horizontal irradiation.
• G_STC: the reference irradiance of 1,000 W/m².

Because Y_f and Y_r both have units of time, PR is dimensionless and is usually shown as a percentage. The normalization reduces the direct effects of plant size and solar resource, but it does not remove every environmental or operational influence.

Worked example for a 10 MWp plant

Assume a 10 MWp plant records the following values during one month:

Therefore:

PR = 144 ÷ 180 = 0.80 = 80%

The result means that measured AC output was equivalent to 80% of the theoretical energy based on nameplate power and measured irradiation. The remaining 20% is not automatically a fault. Temperature loss, inverter and transformer efficiency, cabling, mismatch, soiling, power limitation, downtime, and measurement uncertainty may all contribute.

Why there is no universal “good PR” number

A general threshold such as “anything above 80% is always good” can lead to the wrong decision. DC/AC design ratio, module technology, climate, cell temperature, mounting geometry, bifacial configuration, transformer losses, meter location, and curtailment rules vary across projects. Conventional PR is also temperature-sensitive and commonly appears lower during warm periods and higher during cold periods.^[2]

The professional baseline is the project-specific design model, the contractual guaranteed PR, the approved calculation method, and the plant's historical trend. A sudden deviation from a seasonally comparable baseline is often more useful than comparison with a generic number found online.

Define the measurement boundary before calculating

If one team reads energy from the sum of inverter outputs while another uses the point-of-interconnection revenue meter, their PR values will differ. The latter includes transformer and internal network losses. The IEA PVPS O&M guidelines emphasize that the testing boundary and factors inside or outside the responsible party's control must be explicit.^[3]

Seven common causes of low PR

1. High module temperature: higher cell temperature reduces available power, so uncorrected summer-to-winter comparisons can be misleading.
2. Soiling, shading, and snow: irradiance reaches the sensor plane, but parts of the array cannot convert it into usable power.
3. String loss or mismatch: an open fuse, connector damage, PID, or uneven string current may remain hidden under total inverter power. Our PV string monitoring guide explores this diagnostic path in detail.
4. Inverter outage or derating: faults, internal temperature, grid limits, or unsuitable firmware settings reduce output.
5. Electrical losses: DC/AC cable, transformer, and substation equipment consume part of the available energy.
6. Unavailability and curtailment: grid absence and commanded power reduction must be recorded separately so the source of loss remains visible.
7. Measurement and timing errors: poor pyranometer calibration, sensor tilt, a dirty sensor, or timestamp misalignment between meters and weather data can produce an incorrect PR.

Raw, adjusted, and weather-corrected PR

Raw PR is useful for routine monitoring and trend detection, but it retains temperature effects. For performance-guarantee testing or comparisons across different weather periods, an adjusted or weather-corrected PR may be more appropriate. NREL describes a method intended to reduce weather-related bias and produce a more stable result for contractual testing.^[4]

The KPI name is not enough. The temperature model, weather source, aggregation interval, irradiance threshold, exclusions, and method version should be stored with the result.

Daily, monthly, or annual PR?

• Instantaneous or hourly: helpful for fast diagnostics, but sensitive to clouds, data alignment, and meter latency.
• Daily: suitable for trend alarms and comparisons between similar days; low-irradiation days require caution.
• Monthly: more balanced for owner and O&M reporting and less affected by short-term noise.
• Annual: useful for asset assessment and design-model comparison, but capable of hiding short outages.

A useful dashboard does not show only one large number. It presents PR at several time scales alongside irradiation, module temperature, availability, curtailment, and lost energy.

Implementing defensible PR in SCADA

For a calculation that can support a technical or financial report, SCADA should preserve the complete data lineage:

1. Read energy from the reference meter and retain quality and timestamp information.
2. Collect POA irradiation from a calibrated sensor and apply data-quality checks.
3. Synchronize devices through NTP/PTP and aggregate all inputs over identical intervals.
4. Store the as-built DC capacity rather than relying only on the initial design value.
5. Label grid outages, curtailment, maintenance, and substituted data.
6. Retain raw PR, adjusted PR, and every calculation component for audit.
7. Trigger alarms from baseline deviation and sustained underperformance, not only a fixed threshold.

Harmony SCADA can align energy, POA irradiation, temperature, availability, and equipment status in the historian, then present PR over multiple intervals with calculation detail, trends, and PDF/Excel reports. Combining the KPI with hidden-loss monitoring and string analysis changes the question from “why is the number low?” to “which asset and interval caused the loss?”

Checklist for reviewing a PR report

1. Is the energy boundary and meter location explicit?
2. Does DC capacity come from the as-built record and use the correct units?
3. Which POA sensor supplied irradiation, and when was it calibrated?
4. Are timezone, sampling interval, and timestamps aligned?
5. Are gap filling, low-irradiance filtering, and invalid-data rules documented?
6. How are grid outages, curtailment, and maintenance treated?
7. Is the metric raw or corrected for temperature/weather?
8. Is it compared with the design baseline and a similar historical period?

Conclusion

PR is one of the most useful solar-plant KPIs because it normalizes production for capacity and available solar resource. Its value, however, extends beyond the simple formula: sensor quality, calculation boundary, exclusion rules, and temperature context determine whether the final number is credible. Treat PR as the starting point for diagnosis, not the final verdict on plant health.

References

1. IEC — IEC 61724-1:2021, Photovoltaic system performance — Part 1: Monitoring.
2. Sandia PV Performance Modeling Collaborative — Performance Ratio.
3. IEA PVPS Task 13 — Guidelines for Operation and Maintenance of Photovoltaic Power Plants in Different Climates.
4. NREL — Weather-Corrected Performance Ratio.