Spectrally resolved internal heat generation and temperature prediction in multilayer photovoltaic modules under real operating conditions

For a photovoltaic module operating in the field, the dynamic temperature profile through the multilayer structure is an important issue. It governs the real-time conversion efficiency of the active layer, determines its reliability, and impacts PV module service life. Yet, most thermal models describe internal heat generation through empirical equations or uniform volumetric sources, and initialize transient simulations with a uniform temperature distribution. This work develops a spectrally resolved optical-thermal model predicting the transient temperature distribution across a multilayer PV module. Wavelength-dependent absorption and volumetric heat generation are computed in each active and passive layer by using the complex refractive index of the layer and Fresnel transmittance and reflectance at each interface for transverse-electric and transverse-magnetic polarization modes. The non-uniform heat source is incorporated into the one-dimensional transient heat conduction equation, discretized by an unconditionally stable implicit finite-difference scheme with convective and radiative boundary conditions. The transient regime is initialized from a physically consistent thermal state reconstructed from the module's own morning warm-up rather than an imposed uniform distribution. Crystalline silicon, Cadmium Telluride, and Copper Indium Selenide PV modules are simulated under nominal operating cell temperature conditions and outdoor conditions from the SIRTA-PV1 platform. The heat source is found to be concentrated within the first micrometers of the absorber, and the reconstructed initial state places the module below ambient temperature before sunrise. Predicted equilibrium temperatures agree with NOCT manufacturer values within 1 °C, and rear-surface temperatures reproduce measurements with R 2 above 0.99 under clear skies and 0.97 under cloudy skies. Because the model resolves the temperature of each layer rather than a single surface value, it applies to module thermal design, spectral-selective coating evaluation, and degradation and energy-yield studies.