Role of redox flow battery and AI-based controller in frequency regulation of weak microgrids

The intermittent behavior of renewable technologies, which is driven by changes in the weather, causes the mismatch between power supply and load demand on Microgrids (MGs). The stability of such systems is more threatened when they are integrated with bio-renewable generating units. Under such circumstances, the deterministic load frequency stabilization methodologies can no longer guarantee the MG's stability. To address this issue, this work focuses on the design of an AI-based adaptive controller for frequency stabilization of an isolated bio-renewable MG (BRMG) which is formed from renewable energy sources (RES), bio-renewable generating sources (BRGS), and storage units. In particular, a second-order sliding mode control (SOSMC) has been adaptively designed by the multi-agent fuzzy Q-learning (MAFQL) algorithm to dampen the frequency fluctuations in the integrated power system. Finally, hardware-in-the-loop real-time (HiL) real-time emulations are conducted to demonstrate the applicability of the SOSMC controller-based MAFQL in a real-time framework. The comparative values of various performance indexes revealed a significant improvement of the proposed scheme to other prevalent controllers where the value of the integral of time-weighted absolute error (ITAE) has been further minimized than whale optimization algorithm (WOA) based on the SOSMC controller (66 % improvement), a classical SOSMC controller (97 % improvement), and a PID controller (98 % improvement), respectively. •A novel control strategy is proposed by combining the MAFQL algorithm and SOSMC controller for the LFC in a BRMG.•The utilization of RFB technology has been implemented to enhance the stabilization of current power grids.•The MAFQL is employed to dynamically regulate the gains of the SOSMC.•Real-time evaluations utilizing OPAL-RT are employed to assess the viability of the proposed novel approach.