Modeling of a PV water supply system for different meteorological conditions

In this study, the behavior of a water supply system with PV technology was analyzed and modeled on different meteorological conditions. The photovoltaic water supply system, PVWSS, consists of a PV generator set, a frequency inverter, a motor pump and a storage system. This water supply system was thoroughly experimentally tested, on a test bench in real operating conditions (outdoor) in Recife, Brazil. The meteorological and geometric parameters of the system and its flow rates results were used to perform a modeling through ANN to estimate the pumped volume in function to meteorological conditions and the discharge heads. The ANN modeling predicted the flow rates with percentage deviation and normalized mean square error of − 0.6% and 6.85%, respectively. Using a residential water demand curve, the system was simulated for different meteorological conditions, resulting in a deficit of 0.2%, a unit cost of $0.29 per m3 pumped to a discharge head of 24 m with an optimum reservoir of 2 days of autonomy. This model can predict the flow rates for any location and weather conditions.