Adaptive Artificial Neural Network-Based Control of Active AC-DC Boost Converter for Power Factor Correction

Abstract

The control strategy of a three-phase diode bridge rectifier and a boost-type DC-DC converter system by artificial neural networks (ANNs) for efficient power factor correction (PFC) and output voltage regulation is proposed in this paper. Three-phase AC input is rectified by a six-pulse diode bridge, generating a pulsating DC voltage, which is then fed to a boost-converter stage. In the outer voltage loop, amplifying the reference current's amplitude, the first ANN controller is also introduced, and the second one is incorporated in the inner current loop to establish the inductor current with an appropriate phase in line with the input voltage. This phase-locked loop (PLL) locks the current reference to the AC source and allows the system to operate at a unity power factor. Compared with conventional proportional-integral (PI)-based approaches, ANN controllers are trained to mitigate the effect of nonlinearities, parameter uncertainties, and dynamic disturbances. The simulation results confirm the proposed method and show enhanced dynamic performance, lower THD and compliance with the power quality standards.