Predictive control of an integrated PV-diesel water and power supply system using an artificial neural network

This paper discusses the development of a predictive artificial neural network (ANN)-based prototype controller for the optimum operation of an integrated hybrid renewable energy-based water and power supply system (IRWPSS). The integrated system, which has been assembled, consists of photovoltaic modules, diesel generator, battery bank for energy storage and a reverse osmosis desalination unit. The electrical load consists of typical households and the desalination plant. The proposed Artificial Neural Networking controller is designed to be implemented to take decision on diesel generators ON/OFF status and maintain a minimum loading level on the generator under light load and high solar radiation levels and maintain high efficiency of the generators and switch off diesel generator when not required based on predictive information. The key objectives are to reduce fuel dependency, engine wear and tear due to incomplete combustion and cut down on greenhouse gas emissions. The statistical analysis of the results indicates that the R2R2 value for the testing set of 186 cases tested was 0.979. This indicates that ANN-based model developed in this work can predict the power usage and generator status at any point of time with high accuracy.