A plant reports a performance ratio of 76% in May. The O&M contractor calls it normal. The owner calls it a breach. Both may be right, because a single PR number means little without the season, the sensor, and the definition behind it.
Quick answer. A solar performance ratio benchmark for a modern, well-run grid-connected plant is usually above 80% a year. IEA PVPS Task 13 (2014) puts the best near 90%, while new plants range from 70% to 90%. In hot Indian summers, monthly PR often falls into the low or mid 70s because modules run hot.
TL;DR
- PR is final yield divided by reference yield. It removes the effect of how sunny the site is.
- Annual PR above 80% is typical for modern plants. In Indian field studies, summer PR ran 10 or more points below the winter peak.
- A monthly PR above about 95% usually means a sensor problem, not a great plant.
- Use a temperature-corrected PR for guarantees, or agree a seasonal target curve.
- Check the irradiance sensor, availability, and exclusions before blaming the modules.
This article is for owners, asset managers, and EPCs who set or test PR guarantees. For the monitoring standard behind the metric, read our guide to IEC 61724 performance monitoring.
What is the performance ratio?
The performance ratio (PR) compares the energy a plant delivered with the energy a lossless plant would deliver under the same sunlight. IEA PVPS Task 13 calls it one of the most useful measures of PV system efficiency.
The formula is simple:
- Final yield (Yf) = AC energy delivered (kWh) / DC rated power at STC (kWp).
- Reference yield (Yr) = plane-of-array insolation (kWh/m2) / 1 kW/m2.
- PR = Yf / Yr.
Because both yields are in hours, PR has no units. It does not depend on how sunny the site is. That is why it differs from specific yield and capacity factor, which both rise with better sun.
A worked example
This example uses illustrative numbers, not data from a real plant.
| Item | Value |
|---|---|
| DC capacity | 1,000 kWp |
| Plane-of-array insolation in May | 180 kWh/m2 |
| Reference yield Yr | 180 h |
| Energy delivered in May | 135,000 kWh |
| Final yield Yf | 135 h |
| PR | 135 / 180 = 75% |
What is a good solar performance ratio benchmark?
IEA PVPS Task 13 tracked how PR improved over decades. Its 2014 report on long-term PV performance gives these bands:
| Period | Typical PR | Source |
|---|---|---|
| Late 1980s | 50% to 75% | IEA-PVPS T13-05:2014 |
| 1990s | 70% to 80% | IEA-PVPS T13-05:2014 |
| Modern plants | Usually above 80%, some near 90% | IEA-PVPS T13-05:2014 |
| Newly installed plants (spread) | 70% to 90% | IEA-PVPS T13-05:2014 |
Hot climates post a lower PR for the same hardware, because of temperature losses. That matters for Indian sites.
What Indian field studies report
Published Indian measurements sit in a similar band, with a clear seasonal swing:
| Plant | Annual PR | Monthly range | Study |
|---|---|---|---|
| 10 MW, Telangana | 86.12% | 73.88% (April) to 97.5% (December) | Kumar and Sudhakar, Energy Reports, 2015 |
| Rooftop, Northern India | See study | 72.67% (May) to 82.50% (January) | Energy for Sustainable Development, 2018 |
The 10 MW study is open access in Energy Reports. Its authors attribute the low months to incorrect operation and inverter faults, not heat alone.
The December figure of 97.5% deserves suspicion. A PR that high leaves almost no room for temperature, wiring, inverter, and soiling losses. It more often points to a sensor or data problem.
Why does PR change by season?
Module temperature is the main driver. Crystalline modules lose output as cells heat above 25°C, at a rate set by the temperature coefficient on the datasheet.
NREL’s Weather-Corrected Performance Ratio report (Dierauf et al., 2013) makes the point directly. A PR measured in summer and one measured in winter can differ substantially. That makes plain PR a weak basis for a precise guarantee.
How much can heat move PR?
Here is the same illustrative plant, corrected for temperature:
| Step | Value |
|---|---|
| Measured May PR | 75% |
| Irradiance-weighted module temperature | 55°C (assumed) |
| Temperature coefficient of power | -0.35%/°C (assumed datasheet value) |
| Temperature loss vs 25°C | 30 × 0.35% = 10.5% |
| PR corrected to 25°C | 75% / (1 - 0.105) = 83.8% |
So a May PR of 75% can describe a healthy plant. Check your own module datasheet, because coefficients differ by technology.
Which corrected PR should a contract use?
The NREL method corrects to the plant’s annual average module temperature, not 25°C. That keeps the result comparable with the annual PR in the energy model. PVsyst documents both the plain and temperature-corrected PR.
Pick one definition and write it into the contract. State the reference temperature, the data interval, and the exclusions.
Is PR the same as CUF?
No. Capacity utilisation factor (CUF) is energy delivered divided by energy at full rated output for every hour of the period. CUF rises with better sun. PR does not. Indian tenders usually set CUF obligations, while O&M contracts often guarantee PR.
How to check a PR number before you trust it
Run these five checks in order. Most bad PR readings fail at step 1 or 2.
- Irradiance sensor. Is it in the plane of the array? Is it clean and calibrated? A dirty or tilted sensor shifts PR directly.
- Data gaps. Were hours with missing irradiance or energy data removed from both yields, not just one?
- Availability and exclusions. Were grid outages and curtailment excluded as the contract says?
- Temperature. Compare the temperature-corrected PR with the model, not the raw PR.
- Losses. Only after steps 1 to 4, look at soiling, string faults, and inverter clipping.
What do the readings usually mean?
| Monthly PR reading | Likely meaning | First check |
|---|---|---|
| Above 95% | Sensor or data error | Irradiance sensor and data gaps |
| 80% to 90% | Healthy in cool months | Compare with model |
| 72% to 80% | Normal for hot months, or a real loss | Temperature-corrected PR |
| Below 70% | Fault, outage, heavy soiling, or curtailment | Availability log, string currents |
These bands are a screening aid based on the sources above, not contractual thresholds. Compare against the month-by-month PR in your own energy model.
Where should a PR target come from?
From the energy model, month by month. A PVsyst report gives monthly PR, and its loss diagram shows which loss drives each month. A guarantee set as one flat annual number ignores the seasonal swing. A monthly target curve avoids most disputes.
Good acceptance data helps too. Our commissioning checklist sets out the full test sequence. An IEC 62446 Category 2 test gives every string an I-V baseline. Later PR shortfalls can then be traced to specific strings.
Soiling is often the largest controllable loss. Our soiling guide for India covers measurement and cleaning intervals.
FAQ
What PR should I expect from a new plant in India?
Expect the annual figure from the plant’s own energy model, usually in the high 70s to low 80s for hot sites. Monthly values will run lower in summer and higher in winter.
Does PR fall over time?
Yes, mainly through module degradation. IEA PVPS Task 13 (2021) found a median performance loss rate of about 0.6% to 0.7% per year across 120 systems, depending on method. See the Task 13 loss rate report.
Is a higher PR always better?
For the same site and design, yes. Across sites, no. A cold site posts a higher PR than a hot site with better hardware.
Do bifacial plants break the PR formula?
They can push PR up, because rear-side gain adds energy not counted in front-side irradiance. Agree the PR definition for bifacial plants before signing. Our bifacial module design guide covers how rear-side gain is modelled.
Set a PR target you can defend
A defensible target comes from a site-specific energy model with monthly PR and a clear loss breakdown. Heaven Designs prepares PVsyst yield reports and bid-stage 3D pre-design with monthly PR tables.
To scope a yield study or review a PR guarantee, ask for a project quote. We reply within 1 business day.