Thermal parameter identification of photovoltaic module using genetic algorithm

Operation temperatures of a photovoltaic (PV) module change constantly with ambient variables (e.g. temperature, irradiance, wind speed and direction) as well as electrical operation points. In order to establish an accurate thermal model to monitor PV module temperatures in a shorter time interval (e.g. a few minutes), which can be used in various operation conditions, a five-layer dynamic thermal model has been developed. Usually the parameters of such a model can be estimated from both manufacturer data and experimental tests. However, such an experimental approach does not provide satisfactory results, as these thermal parameters cannot be determined precisely enough due to the complexity of the phenomena and quantitative variations with the time evolution. In this paper, a genetic algorithm (GA) is employed as a optimisation method to identify the model parameters based on onsite measurements sampled from an on-line PV module. Comparisons have been made to validate the identified parameters of the five-layer PV thermal mode.