Design and optimisation of an advanced PID-based control strategy for microgrid power quality enhancement utilising GKAN-OHO algorithm

This paper presents an intelligent power quality enhancement technique for microgrid systems using a 3-Degree-of-Freedom PID Tilt-Integral (3DOF-PID-TI) controller. High penetration of renewable energy sources in microgrids causes voltage fluctuations, frequency instability, harmonic distortion, and inefficient power sharing. To address these issues, a hybrid Graph Kolmogorov-Arnold Neural Network and Opposition-based Hippopotamus Optimisation (GKAN-OHO) method is proposed. The GKAN model predicts load demand accurately, while the OHO algorithm optimally tunes the 3DOF-PID-TI controller parameters for inverter operation. The controller minimises Total Harmonic Distortion (THD), suppresses circulating currents, and improves voltage and frequency regulation under dynamic conditions. MATLAB simulation results demonstrate superior performance over Genetic Algorithm (GA), Elephant Herding Optimisation (EHO), and Modified Water Wave Optimisation (MWWO) methods. The proposed controller achieves a voltage THD of 0.8%, significantly lower than existing techniques, while also reducing settling time, voltage overshoot, and frequency overshoot, ensuring reliable and stable microgrid operation.