Triple-junction cell temperature evaluation in a CPV system by means of a Random-Forest model

The temperature of a multi-junction (MJ) cell under high concentration is basic to evaluate the performances of a concentrating photovoltaic (CPV) system from electric and thermal point of view. The MJ cell temperature (Tc) in a CPV system depends on different parameters as concentration factor (C), cell material and environmental conditions; hence, its evaluation is a complex non-linear problem. In this paper, a Random-Forest (RF) model for the temperature analysis of two different triple-junction (TJ) solar cells mounted on an experimental CPV system, is studied. The two cells are respectively constituted by InGaP/GaAs/Ge and InGaP/InGaAs/Ge, and present areas equal respectively to 5.5x5.5 mm2 and 1.0x1.0 cm2. The temperature of two cells has been evaluated in different working conditions as function of C, environment temperature and solar radiation. The higher temperatures have been obtained with the larger cell (InGaP/InGaAs/Ge). Moreover, an Artificial Neural Networks (ANN) model and a Linear Regression Model (LRM) have been also studied to compare the RF model results. The RF model is resulted the best in comparison with the ANN and LR models both in terms of absolute error and fit capability. Tc has been evaluated for both cells by means of the RF model for C values equal respectively to 1x, 7.33x, 40x and 310x as function of environment temperature and global radiation in different hours included between 9:00 a.m. and 6:00 p.m. A percentage difference between the temperature values of two cells of more than 25% in favor of the larger cell has been noted. Hence, the RF model studied in this paper allows to predict the Tc values necessary to evaluate the thermal energy production of a concentrating photovoltaic and thermal system.