Improved shunt APF based on using adaptive RBF neural network and modified hysteresis current control

In this paper, a new combination is proposed to control shunt active power filters (APF). The recommended system has better specifications in comparison with other control methods. In the proposed combination, an RBF neural network is employed to extract compensation reference currents for a variable non-linear load. In order to make the employed model much simpler and tighter, an adaptive learning algorithm for RBF network is proposed. In addition, a modified hysteresis current control technique based on defining a variable hysteresis band is employed to avoid any power system resonance. In this method the hysteresis band is expressed as a function of source voltage, rate of reference current variations and voltage of DC link capacitor in such a way that the switching frequency of the inverter switches remains almost constant. In summary, extraction of compensation reference current is done with lower amount of computations. Beside, the threat of resonance occurrence is cancelled. The simulation results which are done by MATLAB/Simulink illustrate the validity and effectiveness of the proposed combination.