Neural Network Control Techniques of Hybrid Active Power Filter

A multi-object optimization approach was developed for the design of hybrid active power filters (HAPF) to give better mitigation of the harmonics and better reactive power compensation. The neural network technique was used with optimization theory to improve the algorithm precision and stability. The optimization is more effective since the performance goals and optimization parameters were optimized together. Secondly, this paper presents the design of a hierarchical fuzzy current control scheme for a shunt active power filter compared with a single fuzzy controller scheme. It provides superior current tracking capability and switch frequency signal is limit in the permit range. Finally, many simulations and experimental result demonstrate the validity of the theory.