Robust Hybrid Intelligent Control Using Probabilistic Feature for Active Power Filter

The escalating harmonic issues in the power grid pose a significant threat to the stability and security of the power system. While active power filter (APF) is a dependable solution for harmonic suppression, they grapple with challenges related to uncertainties in control design. Addressing this concern, an enhanced terminal sliding mode control (TSMC) scheme is proposed for APF, leveraging a probabilistic wavelet fuzzy neural network (PWFNN). Initially, the dynamic model of APF, accounting for external disturbances and component parameter perturbations, is derived based on its internal structure. Subsequently, TSMC is applied to APF, and its stability and finite-time convergence are demonstrated through Lyapunov theory. Additionally, a PWFNN, integrating probabilistic features to augment wavelet neural networks, is constructed to approximate uncertain functions, with adaptive laws governing network parameters. Simulation results validate that PWFNN outperforms in harmonic suppression.