Enhancing power quality in renewable energy systems through MAO-RERNN integrated harmonic mitigation

Integrating renewable energy sources (RES) into power systems introduces critical power quality (PQ) issues, like voltage sags (VS), voltage swells (VSW), interruptions, and harmonic distortions, which compromise system stability and performance. This study presents a novel control strategy that combines an enhanced Mexican Axolotl Optimization (MAO) technique with a Recalling-Enhanced Recurrent Neural Network (RERNN), termed MAO-RERNN. The MAO is developed utilizing adaptive crossover and mutation strategies, enhancing global search efficiency and avoiding premature convergence. The proposed method employs a Shunt Active Power Filter (SAPF) to mitigate harmonic distortions, with MAO optimizing the parameters of a Fractional Order Proportional Integral Derivative (FOPID) controller. The RERNN predicts and generates optimal control signals that regulate the SAPF's switching pulses, effectively eliminating harmonic distortions in RES systems. Implemented in MATLAB and validated through simulations, the proposed technique attains a Total Harmonic Distortion (THD) of 1.56 %, outperforming existing techniques such as Spider Monkey Optimization (SMO), Wild Horse Optimizer (WHO), and Side-Blotched Lizard Algorithm (SBLA), which result in THD values of 2.1 %, 3.5 %, and 4.45 %, respectively. These results confirm that the MAO-RERNN technique offers a superior solution for developing PQ in RES systems, with significant implications for the design and optimization of future RES integration technologies.