Comparative Analysis of Controllers for Dynamic Voltage Restorer in Power Quality Enhancement Applications

Abstract

The paper provides a comparative analysis of control systems for a dynamic voltage restorer (DVR) based on three types of regulators: a classic PID controller, a fuzzy controller and a NARMA-L2 neural network controller. A DVR simulation model has been developed in the MATLAB/Simulink environment with the implementation of control systems in a synchronously rotating coordinate system. The simulation was performed with a short circuit between phases b and c and with a nonlinear load connected. The results showed that the Fuzzy controller provides the best combination of low THD, minimal overshoot, and instantaneous adjustment time. The neural network regulator demonstrated the smallest integral error in the short-circuit interval, however, it has an increased THD and a longer adjustment time. It is concluded that the Fuzzy controller is the optimal compromise solution for most DVR applications, and the hybrid approach with switching between regulators depending on the type of disturbance is of the greatest interest for further research.