High-Performance ANN-SMC MPPT Control with Adaptive Super-Twisting Observer for Enhanced Stability in Grid-Connected PV Systems

Grid-connected photovoltaic (PV) systems face significant stability challenges due to rapid irradiance fluctuations and severe grid faults. Conventional maximum power point tracking (MPPT) algorithms often struggle with slow convergence and oscillations under these conditions. This study aims to improve system stability, dynamic performance, and power quality through a novel adaptive hybrid control strategy. The proposed controller integrates an artificial neural network (ANN) for precise reference voltage generation with sliding mode control (SMC), enhanced by an adaptive super-twisting observer (ASTO). This combination effectively estimates unmeasured states and mitigates chattering phenomena. The system is validated using MATLAB/Simulink under rapidly changing solar irradiance and both symmetrical and asymmetrical fault scenarios. Simulation results show that the proposed method significantly reduces DC-link voltage overshoot and settling time compared to conventional methods, ensuring fast and stable tracking. The method achieves a superior total harmonic distortion (THD) of 1.34% and maintains robust operation during faults. The novelty of this approach lies in the ASTO-based ANN-SMC framework, which ensures robust tracking under severe nonlinear disturbances. This solution provides a resilient method to enhance energy conversion efficiency in modern grid-integrated renewable systems.