Modeling of a Solar Pumping System in Agrivoltaic Environments Using Artificial Neural Networks

Abstract

The aim of the study was to employ mathematical modeling in the analysis of a photovoltaic (PV) water pumping system designed for agriculture irrigation in an Agrivoltaic environment that combines crops in the ground with PV panels, using as a case study the region of Khemis Meliana in Algeria in northwest Africa. System performance was optimized by considering the variation in solar irradiance level, ambient temperature, generated PV power, pump flow rate and head, and water demand for irrigation. The maximum power point tracking (MPPT) of the PV module was estimated from the I-V curve model. The performance equation that expresses the flow rate of the pump as a function of input power and the total head was modeled based on the manufacturer’s performance curves using a feed-forward backpropagation network. Results showed that an increase in irradiance intensity increased the pump flow rate. A higher head gave a lower flow rate, regardless of the change in solar irradiance intensity. The power excess was predicted monthly. This power can be further
utilized for secondary applications. The PV water pumping arrangement model enables advance estimation of system outputs and allows for optimization and improvement in the performance and reliability, accounting for variation in solar irradiance level, different pump operating parameters, and the water demand for irrigation.
Future studies are suggested on excess power analysis including economics and considering the targeted
plants to better optimize the system in terms of water need and frequency for irrigation.