Performance Ratio (PR)

This page sets out the PR formula in solar, the performance ratio calculation that O&M engineers, asset managers and lenders use to judge whether a PV plant is delivering the energy its sunlight allows. It includes a worked solar plant example, the data boundaries that change the answer, and common errors.

Performance ratio at a glance

The International Electrotechnical Commission lists IEC 61724-1:2021 as edition 2.0, published on 21 July 2021. The standard covers photovoltaic monitoring terminology, equipment, and analysis methods. As checked on 26 September 2026, IEC also lists edition 3.0 as under development rather than published.

What does performance ratio measure?

Performance ratio measures how much energy a photovoltaic system delivered relative to the solar energy available on its collector plane. It compresses several system effects into one normalized result. Those effects can include temperature, shading, soiling, mismatch, wiring loss, inverter loss, transformer loss, clipping, curtailment, and downtime.

The word normalized needs care. PR normalizes output for plane-of-array irradiation and installed power. Standard PR does not automatically normalize cell temperature. Sandia National Laboratories’ PV Performance Modeling Collaborative notes that PR commonly falls during warmer periods and rises during colder periods because temperature still affects the result.

PR is therefore a diagnostic index, not a component efficiency. Specific yield reports energy per installed kilowatt-peak. Capacity factor compares energy with continuous nameplate output across clock time.

What is the PR formula in solar?

Calculate solar performance ratio by dividing final yield by reference yield. Final yield is AC energy divided by installed DC power. Reference yield is plane-of-array irradiation divided by reference irradiance. Every input must cover the same period and use the same measurement boundary.

Final yield:      Y_f = E_AC / P_0
Reference yield:  Y_r = H_POA / G_STC
Performance ratio: PR = Y_f / Y_r

The symbols mean:

The equivalent direct form is:

PR = E_AC / [P_0 × (H_POA / G_STC)]

The measurement period must match across every input. Do not divide monthly energy by annual irradiation. The power basis must also remain fixed. Mixing an AC rating in one calculation with a DC standard-test-condition rating in another changes the result before plant performance enters the discussion.

Worked performance ratio calculation

This example is illustrative. It shows the arithmetic and unit checks, not a contractual test procedure.

Assume a grid-connected photovoltaic plant has these monthly inputs:

First calculate final yield:

Y_f = 144,000 kWh / 1,000 kWp
Y_f = 144 kWh/kWp

Then calculate reference yield:

Y_r = 180 kWh/m² / 1 kW/m²
Y_r = 180 h

Now divide the two yields:

PR = 144 h / 180 h
PR = 0.800 = 80.0%

The units cancel, which is why PR has no unit. The result means the measured energy equals 80.0 percent of the reference energy for those stated inputs and that boundary. It does not prove that 80.0 percent is contract-compliant. That decision needs the agreed method, exclusions, uncertainty allowance, and comparison value.

Which data boundary should a PR calculation use?

A usable PR record identifies the energy boundary before anyone opens a spreadsheet. For a standard grid-connected system without storage, the numerator is commonly energy delivered to the grid. A behind-the-meter system may need grid export plus directly consumed solar energy. A storage system needs treatment for charge, discharge, conversion loss, and energy shifted between time intervals.

The PVsyst performance ratio documentation uses different effective-energy definitions for grid, self-consumption, storage, stand-alone, and pumping systems. It also warns that short time-step PR becomes less meaningful with storage because energy can be produced in one interval and consumed later.

Record these items beside every reported value:

1. Meter and energy channel used for the numerator.
2. DC power basis and treatment of replaced or unavailable modules.
3. Irradiance sensor, calibration status, position, and time alignment.
4. Calculation interval and missing-data rule.
5. Exclusions for curtailment, grid outage, maintenance, or force majeure.
6. Treatment of temperature, bifacial irradiance, shading, and storage.

Two PR values without this record may describe different questions. Comparing them as if they share a method creates false precision.

Why does performance ratio change by season?

Standard PR can change even when the plant has no new fault. Higher cell temperature reduces module power, so a hot month can report lower PR than a cool month at similar equipment condition. Seasonal shading, soiling, snow, tracker geometry, clipping, and availability also alter the ratio.

Weather-corrected PR applies a temperature adjustment to reduce this seasonal bias. It is a separate method, not a hidden improvement to standard PR. PVsyst explains that the correction mainly addresses array temperature and does not remove every weather-linked effect. The selected method should therefore appear in the report title, calculation sheet, and acceptance procedure.

Use three comparisons instead of one borrowed benchmark:

• Compare measured PR with the approved simulation under matched assumptions.
• Compare the same plant with its own prior matched periods.
• Compare loss channels and data quality before blaming a component.

A PR decline is a prompt to investigate. It does not identify the failed asset by itself.

How do standard, weather-corrected, and bifacial PR differ?

IEC 61724-1:2021 introduced monitoring provisions for bifacial systems. PVsyst explains that a conventional front-side denominator can make rear-side production appear as a gain and can push standard PR above 100 percent for some configurations. A bifacial PR method adds a defined rear contribution to the denominator.

Name the method every time. A report labelled only PR leaves the reviewer guessing about temperature correction and rear irradiance.

How should PVsyst PR be compared with measured PR?

PVsyst reports a simulated PR using its own energy and irradiation variables. A fair field comparison recreates those boundaries. The sensor plane, horizon treatment, shading treatment, export point, self-consumption, and missing intervals must agree with the simulation as closely as practical.

The PVsyst guidance on PR guarantees says that the on-site method must be specified when PR supports a performance guarantee. It discusses sensor placement, calibration, transposition, horizon effects, tracker shading, and the choice between a direct measured calculation and a resimulation with measured weather.

For an EPC handover or lender review, keep one calculation sheet with the raw meter total, irradiance total, installed-power basis, excluded intervals, formulas, and sensor records. The PR result should be reproducible without copying values from a screenshot.

Performance ratio versus specific yield and capacity factor

Specific yield retains the effect of the site’s solar resource. PR places measured or simulated output against measured or simulated collector-plane irradiation. Capacity factor places output against clock time and nameplate power. None can replace the other two.

For design review, use specific yield to compare energy productivity, PR to audit the modeled loss chain, and capacity factor for asset-level utilization. If the three numbers tell conflicting stories, check their power and energy boundaries before revising the design.

Common PR calculation errors

1. Using global horizontal irradiation. A tilted or tracking array needs plane-of-array irradiation, or a documented transposition from horizontal data.
2. Mixing timestamps. Meter and irradiance intervals must share time zone, clock convention, and aggregation boundaries.
3. Ignoring sensor quality. Calibration drift, dirt, shading, or a poorly chosen sensor position biases the denominator.
4. Changing the nameplate basis. AC power, DC standard-test-condition power, and available power are different inputs.
5. Comparing standard PR with corrected PR. The methods can move in different seasonal patterns.
6. Treating PR as fault isolation. The ratio shows a performance gap, but the loss diagram and operating data locate the cause.
7. Applying one benchmark everywhere. Climate, system boundary, bifacial method, storage, exclusions, and contract wording all matter.

What should an auditable PR worksheet contain?

An auditable worksheet lets another engineer reproduce the result. Include the reporting period, source file names, meter identifier, and irradiance sensor identifier. Add calibration dates, the power basis, excluded records, unit conversions, equations, intermediate yields, final PR, and reviewer name.

Keep the source data unchanged. Perform cleaning and exclusions in a separate table with reason codes. This makes a later warranty or operations review much easier because every removed interval remains visible.

For pre-construction work, a PVsyst loss diagram provides the modeled reference. Plane-of-array irradiance and temperature coefficient inputs deserve separate review because they shape the denominator and seasonal response. After commissioning, soiling loss and inverter clipping are two loss channels to test when measured PR diverges from the matched model.

Heaven Designs prepares solar engineering reports and design-stage energy models. The project owner, EPC, independent engineer, and contract administrator should agree the field test boundary and acceptance terms. For a project-specific model or report scope, review the solar engineering reports service or contact the engineering team.