ANN-Based Control and Evaluation of Grid-Integrated Solar Desalination Systems: An Architecture-Oriented Study

Abstract

As the demand for renewable energy and clean water increases, photovoltaic (PV)-powered systems of desalination have arisen as a promising and distinct solution for sustainable production of water. This paper proposes the use of artificial neural networks (ANNs) for performance improvement and intelligent control for PV systems based on grid connection with desalination system. Several architectures of ANN, including Scaled Conjugate Gradient (SCG), Levenberg-Marquardt (LM), and Bayesian Regularization (BR) are evaluated and explored in terms of rate of water production, accuracy of tracking, adaptability to dynamic conditions of environmental, and control efficiency. The suggested models of ANN are utilized for optimizing the parameters operational of system, such as utilization of solar energy and output of desalination, guaranteeing reliable operation whether grid support or PV power is dominant. Comparative results of simulation highlight the distinct behaviour of ANN-based methodologies over traditional Incremental Conductance (IC) algorithm, proving enhanced energy efficiency, decreased losses of system, and improved stability water production. This study emphasizes the prospective of ANN as a powerful and efficient tool for sustainable, resilient, and smart desalination systems.