Performance Ratio, and What a Good Number Actually Looks Like

Performance & Operations   6 min read

Performance ratio is the number owners quote at each other and the number contractors defend. It is genuinely useful, and it is also easy to compute in a way that flatters whoever is doing the computing.

The definition

PR is measured energy divided by theoretically expected energy, where expected energy comes from the irradiance that actually landed in the plane of the array, multiplied by installed capacity, referenced to standard conditions. IEC 61724-1 sets out the method.

Because it normalises against measured irradiance, PR strips out weather. A cloudy month and a clear month should produce similar PR from the same plant. That is what makes it comparable across time and across sites.

What a realistic figure is

ContextTypical annual PR
Well-built rooftop, temperate climate (UK, Germany, northern US)80 - 85%
Well-built rooftop, hot climate (Spain, Texas, Gulf states)75 - 80%
Ground mount, good design, temperate82 - 87%
Ageing plant, ten years, average maintenance72 - 78%
Plant with a real problembelow 70%

The single biggest driver of the climate difference is temperature. Modules lose roughly 0.3-0.4% of output per degree above 25°C, and a rooftop array in Seville spends much of the summer at 60°C+. That is not a fault; it is physics, and a good plant in a hot climate legitimately posts a lower PR than an average plant in a cool one.

Which is why comparing PR between sites in different climates tells you very little. Comparing a site against itself over time tells you a great deal.

Where the losses sit

A typical breakdown between nameplate and delivered energy:

  • Temperature: 4-10% depending on climate. The largest single item in most plants.
  • Inverter conversion: 2-3%. Modern units are 98%+ efficient at good load, less at very low load.
  • Soiling: 1-5%. Enormously site-dependent.
  • DC wiring and mismatch: 1-3%.
  • Shading: 0-8%. Should be near zero if the design was done properly.
  • Availability: 1-3%. Downtime for faults and maintenance.
  • Reflection at low sun angles: 2-3%.

How PR gets misreported

Temperature-corrected PR. There is a variant that normalises out module temperature as well as irradiance. It is legitimate and useful for isolating design and maintenance quality – but it produces a number several points higher than standard PR. If a report quotes 88% for a hot-climate rooftop, check which definition is in use.

Short measurement windows. PR measured across a mild, clear spring month will beat the annual figure comfortably. Guarantees tested over short favourable windows are testing the season, not the plant.

Uncalibrated or dirty sensors. If the pyranometer under-reads, PR rises, because the plant appears to be producing more from less light. A soiled reference sensor is the most common cause of a suspiciously good PR figure.

Excluding downtime from the calculation. Some reports compute PR only across hours when the plant was operating. That measures conversion efficiency while running and hides availability problems entirely. Both numbers have uses; presenting the first as overall PR is misleading.

Using it well

Track PR monthly and plot it against the same month in prior years. Seasonal variation is normal – PR typically peaks in cool clear months and dips in hot ones – so year-over-year comparison of the same month is the meaningful signal.

A gradual decline of roughly half a percent a year is expected module degradation. A step change points to a fault. A steady decline steeper than about 1% a year means something is wrong: soiling accumulation, a failing string, connector degradation, or growing shading from something new.

The plant tells you what is happening. PR is just the language it uses.