Solar panel degradation rates

Solar panel degradation is the permanent loss of power capability as a PV module ages. A figure near 0.5% a year is often used as a planning assumption, but it is not a promise about every panel and it is not a reliable diagnosis for one household system. Published results vary by product, climate, age, measurement method and whether the study is measuring modules or complete systems.

The best-supported broad finding is a range, not a universal rate. An influential review assembled nearly 2,000 published degradation rates for modules and systems exposed outdoors over four decades and found a median of 0.5% a year. A later compendium covered more than 11,000 rates from almost 200 studies in 40 countries. For crystalline silicon it found medians around 0.5% to 0.6% a year and means around 0.8% to 0.9%.

Those results describe mixed populations. They should not be rewritten as “a modern panel loses exactly 0.5% each year”.

What newer module studies show

A study published in 2023 tested 834 fielded modules representing 13 module types from seven manufacturers in three climates. Over the study period, the mean measured rate was 0.62% a year and the median was 0.58%. The researchers also found non-linear behaviour, seasonal variation and meaningful differences between modules. Six of the 23 systems studied were on trends that could exceed their warranty limits, while 13 showed the potential to operate for more than 30 years if their measured trends had stabilised.

NREL’s 2024 PV Lifetime report gives a useful warning against simple technology rankings. Its individual product cohorts ranged from better than 0.25% a year for some tested module groups to around 2% a year for one heterojunction product group. Several cohorts showed much of their measured loss in the first year. These are results for named samples under a particular test programme, not market-wide rates for each cell architecture.

The evidence therefore supports two statements that can sit together: degradation around half a per cent a year is a reasonable central reference for many crystalline-silicon populations, and the actual result for a particular product can be materially better or worse.

Early change is not always a straight-line trend

Performance can change more quickly after first exposure and then settle. Light-induced degradation and light and elevated temperature-induced degradation are two mechanisms associated with early-life changes in some crystalline-silicon products. The 2023 study found that degradation-rate estimates stabilised after roughly three to four years in its tested population, with some modules showing seasonal recovery as well as loss.

This matters when reading a warranty graph. A straight line printed from year two to year 25 is a contractual threshold, not evidence that the product will follow a perfectly straight physical curve. It also matters when estimating a rate from a short monitoring history. A few months or a single year cannot reliably separate module ageing from weather and other system effects.

Module degradation and system performance are not the same

A module degradation test tries to compare the panel’s power capability under controlled or weather-corrected conditions. The energy total in an inverter portal includes far more:

• the sunlight reaching the array and module temperature
• temporary shading, dirt, snow or leaves
• inverter availability and conversion losses
• clipping, curtailment and export limitation
• faults, failed strings, connectors and cabling
• changes to the monitoring equipment or its data record

NREL’s US fleet work found median system performance loss between roughly 0.5% and 0.75% a year depending on method and treatment of soiling. That fleet is largely commercial and utility-scale, so it is not a direct forecast for a UK house. It does demonstrate why system-level and module-level figures must be labelled separately.

A 2026 study of 1.25 million German PV systems reported a weather-normalised decline in system performance ratio of about 0.59% a year. That result is useful European system evidence, but it still does not isolate the panels from the rest of each installation.

Why annual generation cannot prove degradation by itself

UK solar output varies substantially with the weather. Comparing one calendar year’s generation in kWh with the next can therefore produce a large apparent gain or loss even when the equipment has not changed. Tree growth, new buildings, soiling and periods when the inverter was offline can add further differences.

Long-term monitoring can identify a trend if irradiance and temperature are accounted for and poor-quality periods are excluded. A module-level assessment is more demanding. Research programmes use calibrated current-voltage measurements under controlled conditions or carefully corrected field data. A homeowner portal normally cannot reproduce that test.

Where output appears to have fallen, first check for a discrete fault or changed site condition. A sudden step down, repeated inverter alarms or one string behaving differently is not the normal pattern of gradual ageing. An installer or independent solar specialist can compare strings, inspect the array and review weather-normalised performance before attributing the loss to module degradation.

Reading a performance warranty

A module performance warranty states the minimum warranted power at specified points in time, subject to its terms and measurement rules. It is not the manufacturer’s prediction of the exact output you will receive. It also applies to module power, not directly to annual household generation.

Before comparing warranties, check:

• the initial allowance before the annual decline begins
• whether later limits are annual steps or a final-year threshold
• the reference power and measurement tolerance
• the claim procedure and who pays for testing, removal or transport
• exclusions for installation, environment and system voltage
• whether the product warranty and performance warranty have different periods

Two warranties that both advertise 25 or 30 years can provide different minimum power and very different practical remedies. The current warranty document for the exact model is the relevant source, not a generic rate for its cell technology.

A defensible expectation

For early financial estimates, a degradation assumption around 0.5% a year is within the central range of major crystalline-silicon reviews. Sensitivity testing at a higher rate is more honest than presenting that central figure as certain. A model-specific warranty can supply a contractual floor, while independent field evidence for that exact product may be limited or unavailable.

There is not enough evidence to assign every P-type, TOPCon, heterojunction or back-contact panel a dependable generic rate. Materials, manufacturing and the complete bill of materials can differ within the same broad cell category. Product-level evidence should take priority over a technology table.

Related reading

• Solar panel ratings and degradation
• PV module warranty types
• MCS PV performance estimates
• Monitoring portal data
• PV fault codes