PV Temperature Derate

Photovoltaic modules lose power as they warm, and the correction is linear over the range that matters: $P = P_{stc}[1 + \gamma(T_c - 25\,^\circ\mathrm{C})]$. The coefficient $\gamma$ is negative — typically $-0.30$ to $-0.45$ %/°C for silicon — and it is the single largest item inside a real system's performance ratio.

The physics is almost entirely in the voltage. Heat narrows the semiconductor band gap and raises recombination, so the open-circuit voltage falls roughly 0.3–0.4% per degree while the short-circuit current creeps up by only about 0.05%. The product falls, and it falls at a rate that is a property of the cell chemistry rather than of the workmanship. Thin-film CdTe sits near $-0.25$ %/°C, which is a genuine advantage in hot climates and partly offsets its lower efficiency.

Cell temperature is not air temperature, and assuming otherwise is the usual error. A module in full sun runs 25 to 35 C° above ambient, which is why datasheets quote a nominal operating cell temperature (NOCT) near 45 °C at only 20 °C of air. Mounting matters as much as climate: a rack with free air behind it runs 10 to 15 C° cooler than a panel laid flat against a roof deck, which is worth several percent of annual yield for nothing but standoff height. Building-integrated PV, sealed against the structure, pays the largest penalty of all.

The effect runs both ways, and the upward direction has a design consequence. A clear, cold, bright morning in March can push an array above its nameplate rating, and the voltage coefficient — larger than the power coefficient, and also on the datasheet — is what governs cold-weather string sizing. A string designed for summer conditions can exceed the inverter's maximum input voltage on the coldest sunny morning of the year, which is why the sizing calculation is done at record low temperature rather than at anything typical.