Enhancing infrastructure sustainability: reliability and sensitivity analysis of localized integrated renewable energy systems using feed forward backpropagation neural network

Abstract

Establishing robust infrastructure and securing a sustainable power supply can be costly and time-consuming. Localized power generation from natural resources through integrated renewable energy systems (IRESs) offers a solution. This study explores the reliability and key statistics of an IRES incorporating solar photovoltaic modules, wind turbines, and battery banks. The failure and repair rates follow an exponential distribution due to the increased risk in integrated structures. Neural networks (NN), particularly the feed forward back propagation neural network (FFBPNN), enhance the consistency and precision of reliability parameters. The learning process of FFBPNN adjusts neural weights, improving parameter values. Utilizing the MATLAB algorithm, this study iterated until achieving accuracy close to 0.0001. The proposed system's real-time operations can be effectively managed by analyzing operational costs and system sensitivity to different parameters.