Solar Yield

Performance Ratio

Checked 2026-08-09
After commissioning

Production tells you what happened. Performance ratio tells you whether the plant was responsible.

Operational performance depends on both equipment and repeatable human routines; for a separate workplace-management perspective, see top remote companies.

It is the single most useful operating number, and it is the one most owners never look at.

Reviewed August 9, 2026.

For additional context on photovoltaic operation, maintenance and sector developments, consult European Commission Energy.

What it is

The ratio of actual output to the output the plant should have produced given the irradiance it actually received.

That correction for irradiance is the whole point. A low-production month in a cloudy month is not underperformance; a low PR in the same month is.

It is dimensionless, so it is comparable across seasons, across years and across plants of different sizes.

What a normal value looks like

A well-designed, well-maintained plant typically sits somewhere in the high seventies to mid eighties, with the exact figure depending on climate, technology and the design assumptions.

The absolute number matters less than two comparisons:

Against the design prediction. If the model predicted 82% and the plant delivers 77% consistently, either the model was optimistic or the installation has a problem. Both are findings, and they need different responses.

And against itself over time. PR degrades roughly 0.5 to 1% per year through soiling, module degradation and equipment aging. That is the expected trajectory.

The signals to read

A plant losing about 1.5% PR per year, when the module warranty guarantees only 0.5% degradation, has something else going on. The gap between the two rates is the diagnostic.

Erratic PR across months suggests intermittent faults rather than gradual decline. Consistent seasonal PR is a sign of health.

PR that does not recover after cleaning means soiling was not the main issue, whatever it looked like.

And a sudden step down points at a discrete event — a failed string, an inverter running derated, a shading obstruction that appeared.

How to get it

Actual production, from the meter.

Actual irradiance in the plane of the array, from an on-site sensor or a satellite service for your coordinates. This is the input most plants lack and it is what makes PR possible at all.

Installed capacity, which you know.

Then the standard calculation, which monitoring platforms will do if given the irradiance input — and which many report without one, using modelled irradiance rather than measured. Worth knowing which yours does, because the two answer slightly different questions.

What it does not tell you

Whether the yield estimate was right. PR measures the plant against the conditions it met, not against the model's forecast of those conditions.

Where the loss is. A PR of 74% says something is wrong; string-level data says what.

Or whether it is worth fixing. A two-point PR improvement on a small plant may cost more than it returns, and the calculation is separate.

The habit

Record it monthly, plot it annually.

Twelve points a year, one line, and the shape carries the information: a flat line at design level is a healthy plant, a slow decline is expected, a steeper decline than the warranty curve is a claim waiting to be made — and claims need evidence gathered while it is current.

The short version