Implementation of neural-network-controlled three-leg VSC and a transformer as three-phase four-wire DSTATCOM

Abstract

In this paper, a neural-network (NN)-controlled distribution static compensator (DSTATCOM) using a dSPACE processor is implemented for power quality improvement in a three-phase four-wire distribution system. A three-leg voltage-source-converter (VSC)-based DSTATCOM with a zig-zag transformer is used for the compensation of reactive power for voltage regulation or for power factor correction along with load balancing, elimination of harmonic currents, and neutral current compensation at the point of common coupling. The Adaline (adaptive linear element)-based NN is used to implement the control scheme of the VSC. This technique gives similar performance as that of other control techniques, but it is simple to implement and has a fast response and gives nearly zero phase shift. The zig-zag transformer is used for providing a path to the zero-sequence current in a three-phase four-wire distribution system. This reduces the complexity and also the cost of the DSTATCOM system. The performance of the proposed DSTATCOM system is validated through simulations using MATLAB software with its Simulink and Power System Blockset toolboxes and hardware implementation.