ANN model of a triple-junction cell in concentrating photovoltaic system

Abstract

Significant advancements in concentrating photovoltaic (CPV) systems have been achieved in recent years, also thanks to the definition of calculation methods of their energy performances in several operation conditions. Typically, the CPV systems electrical power is separately calculated or in terms of its temperature or concentration factor (C), but not simultaneously in terms of both variables. In this paper, an Artificial Neural Network model based on experimental data, linking electric power of CPV system with Direct Normal Irradiance and Triple-Junction cell temperature for different C values, is developed. Moreover, the model is also adopted to realize a feasibility analysis of point-focus CPV system used for different users: residential building and agricultural livestock farm. The optimal number of modules is determined to maximize the Net Present Value (NPV) of the investment. For the residential user, an optimal configuration of CPV system includes 16 modules, providing a peak power of 3.1 kW and covering an area of 130 m2. This configuration allows the maximization of NPV value, reaching 15.9 k€, with DPB of 9.8 years. As for the agricultural livestock, 36 modules, with peak power of 7.0 kW and covering an area of 292 m2, allow the maximization of NPV value equal to 16.3 k€, with DPB of 10.2 years.