Experimental verification of the effectiveness of neural modified sliding mode technique in multi rotor wind turbine systems

Ripples in energy and current are two of the main issues with direct power control, as these fluctuations cause numerous drawbacks in the wind energy system. However, the conventional approach's many benefits make it one of the most popular approaches in the wind power industry due to its simplicity, ease of use, and ease of realization. In this paper, a neural-modified sliding mode approach has been proposed to control the power of a double-powered induction generator-based multi-rotor wind turbine system. The designed control is described as highly robust and has outstanding competence. MATLAB and experimental work were used to verify this performance compared to the usual approach. The outcomes demonstrated the efficiency of the designed approach in getting a better quality of generated power and supplied currents compared to the conventional approach. The suggested approach minimized the overshoot value by ratios estimated at 99.82% and 97.26% for both reactive and active power, respectively. Also, the value of current harmonic distortion was minimized by 48.80%, 46.35%, and 61.29% in all tests performed compared to the classical approach. The designed approach reduced the value of active power ripples by ratios of approximately 81.35%, 85.20%, and 84.04% in all tests. Empirical results using hardware-in-loop simulation based on dSPACE 1104 confirm the high competence and ability of the designed approach to significantly improve the quality of energy and current, allowing it to be relied upon in the future as a solution in the area of control.