Neural network-based three-phase NPC rectifier for DC bus capacitor voltage balancing under perturbed mains supply conditions

Abstract

This paper presents an artificial neural network-based modified space vector pulse width modulation control approach for better performance of three-phase neutral-point clamped rectifier using optimised switching sequences. Use of optimised switching sequences even under ideal supply conditions, it is depicted that source side and load side parameters deviate from acceptable limits and a DC-bus capacitor voltage unbalance occurs. Under the influence of disturbed supply, source side and load side parameters deviate more beyond acceptable limits which causes a very large unbalance in DC bus capacitor voltages. This non-ideal performance of the converter is responsible for the deterioration of quality of source currents and a large stress on power semiconductor devices. The proposed control scheme employs a three-layer feed-forward neural network at different stages for capacitor voltage balancing of a three-phase three-level neutral-point clamped converter with improved power quality. According to the supply conditions, the neural network varies the speed of the reference vector and forms a required trajectory while passing through the most effective regions of SVPWM hexagon. The proposed controller scheme is modelled in MATLAB/Simulink software. Simulation results show that the proposed implementation of neural-networks controller in three-phase NPC converter displays better performance under ideal and disturbed mains conditions.