AI-Based Hybrid Control for Renewable Energy Systems: A Review

Abstract

Efficient control strategies are needed for renewable energy systems due to the nonlinear nature of their dynamics, their intermittent characteristics, parameter uncertainties, as well as grid connection limitations and the need for real-time implementation. Recently, hybrid control structures where traditional real-time control principles combine with artificial intelligence adaptation, tuning or decision support are being used more often to address these issues. This paper presents a systematic review of the recent literature concerning the use of hybrid controls in renewable energy systems. In particular, it will focus on solar energy, wind energy, microgrids, as well as hybrid renewable configurations, providing both a theoretical framework and practical examples. The methodology applied to review scientific articles of both AI-tuned PID controllers, which will include literature about the following groups of controllers: artificial intelligent neural networks, fuzzy and neuro-fuzzy approaches, adaptive PID and predictive-control schemes that highlight their workflow, performance enhancement, robustness considerations, computational complexity, practicality of application/controller family and usage modes, as well as other issues. The analysis of scientific works presented so far shows three main tendencies in the research: researchers mostly rely on computer simulation results to present their findings; comparisons of controller types fail to provide methodological consistency; and their failure to address practical problems of complexity, tuning, and embedded implementation. In conclusion, this systematic review will indicate what methodological and practical aspects are missing in the literature and provide some suggestions on how to improve future research on intelligent hybrid control schemes to achieve more manageable solutions.