Involvement is not the same as causation
A fire involving PV may start in the system, or reach a system affected by a fire from another cause. Reports of “fires involving solar panels” do not always make this distinction. A list of press reports is therefore not a sound statistical record of electrical PV fire causes.
Assessment needs, where possible, the origin, affected components, building type and investigation status. Where these details are missing, an individual incident should not be used to make a general claim about the failure of a material family.
The QBE example: from 107 to 171 incidents
In 2025, QBE reported UK fire service information on fires involving solar installations. It cited 107 incidents for 2022, 128 for 2023 and 171 for 2024. The increase from 107 to 171 is approximately 60%. [1]
The dataset comprised responses from 37 of the 49 fire service organisations contacted. It is neither a complete German statistic nor a statement that PV caused every incident. The substantial growth in the installed base also matters to the interpretation.
The Netherlands: TNO separates presence from origin
In 2024, TNO, NIPV and NEN reported on approximately 10,000 building fires from 2022–2023. A solar energy system was present on the building in 152 cases. Of 70 cases examined in more detail, 30 originated in the solar energy system, 29 had another origin and the cause remained unknown in 11. [3]
The researchers stress the limited dataset and the need for longer-term monitoring. These figures do not represent the probability of fire per installed PV system. They demonstrate why the presence of solar panels must be distinguished from a fire originating in the PV installation.
| Established origin | Cases |
|---|---|
| In the solar energy system | 30 |
| Outside the solar energy system | 29 |
| Unknown | 11 |
Why absolute numbers do not show individual risk
As the number of installations increases, more incidents can occur even if the risk per installation remains unchanged. Reporting practices, data coverage, system age and building use also change the comparison. Without a suitable denominator, incident counts cannot establish a reliable risk trend.
A private home, a commercial flat roof and a solar farm are different scenarios. For the protection of an industrial roof, its construction, PV geometry, use and potential consequential losses are often more useful than a single general percentage.
What can still be learned for practice
Fire reports can highlight recurring inspection points: electrical connections, cable routes, inverters, spread beneath modules and accessibility. Specific measures combine these observations with professional design and suitable guidance. [2]
The resulting strategy is straightforward: reduce possible ignition sources and account for the consequences of an incident in the roof concept. Mineral layers are one element alongside installation, maintenance, layout and operational arrangements.
After an incident: document before drawing conclusions
Following a roof fire, electrical hazards, structural capacity, waterproofing and possible residues require professional assessment. Photographs alone do not reliably show which component failed first. Damage assessment and cause investigation belong with qualified specialists.
Accurate records of the previous assembly, installed materials and earlier changes help with reinstatement. Our documentation guide explains which information is worth retaining before any loss occurs.
Compare fire counts only on a consistent time period, reporting coverage and installation basis. Keep PV presence separate from fires originating in PV.