Validation of energy valley optimization for adaptive fuzzy logic controller of DFIG-based wind turbines

Abstract

This study presents a novel optimization algorithm known as the Energy Valley Optimizer Approach (EVOA) designed to effectively develop six optimal adaptive fuzzy logic controllers (AFLCs) comprising 30 parameters for a grid-tied doubly fed induction generator (DFIG) utilized in wind power plants (WPP). The primary objective of implementing EVOA-based AFLCs is to maximize power extraction from the DFIG in wind energy applications while simultaneously improving dynamic response and minimizing errors during operation. The performance of the EVOA-based AFLCs is thoroughly investigated and benchmarked against alternative optimization techniques, specifically chaotic billiards optimization (C-BO), genetic algorithms (GA), and marine predator algorithm (MPA)-based optimal proportional-integral (PI) controllers. This comparative analysis is crucial in establishing the efficacy of the proposed method. To validate the proposed approach, experimental assessments are conducted using the DSpace DS1104 control board, allowing for real-time application of the control strategies. The results indicate that the EVOA-AFLCs outperform the C-BO-based AFLCs, GA-based AFLCs, and MPA-based optimal PIs in several key performance metrics. Notably, the EVOA-AFLCs exhibit rapid temporal response, a high rate of convergence, reduced peak overshoot, diminished undershoot, and significantly lower steady-state error. The EVOA-AFLC outperforms the C-BO-AFLC and GA-AFLC in terms of efficiency, transient responses, and oscillations. In comparison to the MPA-PI, it improves speed tracking by 86.3%, the GA-AFLC by 56.36%, and the C-BO by 39.3%. Moreover, integral absolute error (IAE) for each controller has been calculated to validate the system wind turbine performance. The EVOA-AFLC outperforms other approaches significantly, achieving a 71.2% reduction in average integral absolute errors compared to the GA-AFLC, 24.4% compared to the C-BO-AFLC, and an impressive 84% compared to the MPA-PI. These findings underscore the potential of the EVOA as a robust and effective optimization tool for enhancing the performance of adaptive fuzzy logic controllers in DFIG-based wind power systems.