Abstract
Hourly water flow rate (HWFR) forecasting is very important to photovoltaic water pumping system (PVWPS) planning, operation, and control. In this paper, a nonlinear autoregressive with exogenous input-recurrent neural network (NARX-RNN) is investigated for the prediction of water flow rate (WFR) using experimental data collected from a PVWPS installed at Madinah site (Saudi Arabia). Results showed that the developed NARX-based model is able to reach acceptable accuracy for 1-12 hrs (next-day) ahead predictions. The developed methodology provides valuable information to PVWPS operators for controlling the production, storage, and delivery of water.