This work presents various models developed and implemented within the SOPHIA European project in order to thermally characterize PV modules in a rooftop BIPV configuration. Different approaches have been considered, including a linear model, lumped elements models and models that make use of commercial software solvers. The validation of the models performed by comparing the results of simulations with experimental data recorded on a test bench over an entire year is presented and discussed on a seasonal basis. The results have shown that all the models implemented allow achieving a good prediction of the PV modules back surface temperature, with the minimum value of the coefficient of determination R2 around 95% on a yearly basis. Moreover, the influence of season weather conditions and of the incident solar irradiance magnitude on the accuracy of the considered thermal models is highlighted. The major result of the present study is represented by the fact that it has been possible to perform a better thermal characterization of the BIPV module by tuning some of the heat transfer coefficients, such as those relative to the effects of the wind velocity, and to the evaluation of sky temperature. (C) 2017 Elsevier Ltd. All rights reserved.