Photovoltaic module temperature prediction with Gaussian process

Abstract

The output power of solar photovoltaic (PV) arrays usually fluctuates with the changes of solar radiation and the module operating temperature. However, the module temperature is always substituted with the ambient one in the modelling or simulation of PV arrays since the later is easy to be measured. Unfortunately, the module temperature is quite different from the ambient one, which will result in great deviation in the ambient temperature-based simulation. To deal with this problem, the module temperature must be first obtained for improving the accuracy of the simulation of PV arrays. To this end, a Gaussian process-based prediction method of PV module temperature is proposed and validated. First, the relationship between the PV module temperature and the ambient one as well as the output power is analysed; then, the prediction model of PV module temperature based on Gaussian process is given; finally, some empirical experiments are conducted to validate the effectiveness of the proposed method.