Artificial Neural Network-Based Control of Vienna Rectifier for Power Factor Correction and Capacitor Voltage Balancing

Abstract

This paper uses a modified control approach for a three-phase Vienna rectifier based on hysteresis current control and artificial neural network (ANN) for dynamic power factor correction and improved input current quality. The Vienna rectifier is a high-efficiency device and an ideal solution for medium-power applications that require unidirectional power flow and must adhere to strict power quality regulations. Nevertheless, conventional hysteresis current control has a variable switching frequency, which leads to increased electromagnetic interference (EMI) and a more complicated filter design. Therefore, an adaptive control method based on ANN is proposed, which can adaptively update the hysteresis band and current reference profiles according to real-time operating conditions. The ANN is employed to predict the intended current that will maximize the efficiency of the Vienna rectifier and to balance voltages across capacitors on the DC side under load disturbances. The simulation results confirm that the proposed control scheme achieves a near-unity power factor, low THD of the input currents, and stable DC–link voltage regulation in various dynamic operating conditions. The proposed strategy based on ANN achieves the objective of a future smart-grid interfaced rectifier system with resilience, intelligence, and hardware efficiency.