- What the performance ratio is counting
- Performance ratio beside degradation
- How to use the performance ratio without a fake target
- Frequently asked questions
- What is a good solar performance ratio?
- Can I calculate the performance ratio from my electric bill?
- Does a low performance ratio mean the panels are worn out?
- Sources and references
The performance ratio of a photovoltaic system is the final yield divided by the reference yield. In the definition used by Marion and colleagues at NREL, following IEC 61724 as it stood for their 2005 paper, the final yield is the system’s net energy divided by the array’s nameplate DC power. The reference yield is the in-plane irradiation divided by the reference irradiance, so if that reference is 1 kilowatt per square meter, the reference yield is the peak sun hours. Divide the first by the second and you have a dimensionless number. It is not the module’s efficiency. It is how much of the rated output survived the losses between the sun on the plane of the array and the energy you actually metered.
What the performance ratio is counting
Marion’s paper lists the losses that sit inside the ratio: inverter inefficiency, wiring, mismatch, and other losses on the way from DC to AC; module temperature; reflection from the front of the module; soiling or snow; downtime; and component failures. A low ratio can be any of those. The ratio does not name which one. A monitoring total that only shows kilowatt-hours cannot compute it, because the reference yield needs the in-plane irradiation. Without a sensor, or a defensible irradiance source for that plane, you have energy, not a performance ratio.
Temperature is why the number moves through the year even when nothing is broken. The same paper says performance ratios are normally higher in winter than in summer, because hot modules produce less power at the same irradiance, and that the values they were describing normally fell in a range of about 0.6 to 0.8. That range is that paper’s statement about the systems and the climate effects they were discussing in 2005. It is not a pass-fail grade for a system built this year, and it is not a reason to call a summer month a failure because it sits below the winter month. If you quote 0.6 to 0.8, quote the year and the reason.
The paper also says monthly or yearly values are the usual report, and that daily or weekly values can help spot a component failure. A single cloudy hour is a bad interval for a verdict. Decreasing yearly values may indicate a permanent loss. That is a trend to investigate, not an automatic warranty claim. Permanent loss and a dirty month look the same in one ratio until you wash the array and recalculate. Cleaning is in how to clean solar panels. Snow is a seasonal bite out of the same ratio: snow on solar panels.
Performance ratio beside degradation
It is not nameplate efficiency. Efficiency on a datasheet is power out of the module divided by irradiance on the module at standard test conditions, 1,000 watts per square meter and 25°C cell temperature. The performance ratio starts after that nameplate and asks how the system did in the weather. Mixing them is how a 20 percent module is described as a 20 percent performance ratio, which is a different and much worse claim. Field effects on the module itself are in what affects solar panel efficiency.
It is not a degradation rate. Jordan and Kurtz’s median near 0.5 percent per year is a literature result about how module power falls over time. A performance ratio can fall because of that, or because of dirt, a failed inverter, or a hotter year. One number does not prove the other. How long panels last is in how long solar panels last.
It is not proof the inverter is broken when the weather is the cause. A hot afternoon, a shaded string, and a cloud all change energy. Some of them change the ratio and some of them are already in the reference yield if the sensor saw the cloud. Shade that the sensor did not see will look like a loss. Sorting weather from faults is in solar inverter faults and the weather. Heat is in how heat affects a solar module.
How to use the performance ratio without a fake target
| If you have | You can say |
|---|---|
| Metered energy and nameplate DC power only | Final yield, in kilowatt-hours per kilowatt. Not a performance ratio. You have not divided by the sun. |
| In-plane irradiation as well | Reference yield, then the ratio. State the month. Expect a seasonal swing from temperature. |
| A ratio that fell and stayed down after a wash | Look for a failed string, an inverter limit, or a real loss of module power. The ratio did not name it. |
| A brochure that says “performance ratio 90 percent” with no sensor and no interval | Ask for the definition, the irradiance data, and the period. A bare percentage is not this metric. |
Marion’s group also warned that the nameplate you divide by changes the story. They used DC nameplate so that inverter and mounting losses would still show up. If someone computes a ratio against an already-derated AC rating, the number moves up and hides the conversion loss. Ask which power is in the denominator. A 2005 derate example in the same paper put a typical overall factor near 0.80 at standard test conditions and near 0.73 at a cell temperature they labeled NOCT, before soiling and shade were allowed to get worse. Those factors are that table’s estimates, not a performance ratio you should hit, and not a 2026 product specification.
Payback models need energy, not a borrowed ratio. PVWatts, which PNNL points to when monthly figures are missing, already folds weather and losses into an estimate. Replacing that estimate with a homemade performance ratio double-counts or contradicts it. The payback setup is in how a solar payback period is calculated.
Frequently asked questions
What is a good solar performance ratio?
Marion and colleagues wrote in 2005 that values normally fell between about 0.6 and 0.8, higher in winter than in summer because of temperature. That is a historical description, not a pass mark. A modern system can sit elsewhere. Judge a trend on your own array, with irradiation measured, not a single number from a different decade.
Can I calculate the performance ratio from my electric bill?
No. The bill can give energy. The ratio also needs the in-plane irradiation and the DC nameplate. Without the sun on the plane of the array, you have a final yield at best.
Does a low performance ratio mean the panels are worn out?
Not by itself. Dirt, heat, shade, downtime, and a failed inverter are inside the same number. Wash, check the strings, and look at a second period before you call it degradation. A falling yearly trend is the clue the 2005 paper flags. One bad month is not.
Sources and references
Final yield, reference yield, performance ratio, the loss list, the winter-versus-summer note, and the 0.6 to 0.8 range: Marion and colleagues, NREL/CP-520-37358, 2005. The range is quoted from that paper and is not used here as a current pass-fail. Standard test conditions for the nameplate the yield is based on: efficiency tables, version 68. Module degradation medians, which are not a performance ratio: Jordan and Kurtz, 2012.
