Interactive Decision Tree based Fuzzy-Proportional Integral Derivative Controller for Reactive Power Compensation using STATCOM for Non-linear Loads

Abstract

Abstract The STATCOM is extensively used in the power system to address the voltage sag and other power quality issues that arises due to the inductive and nonlinear loads. The STATCOM uses a power electronics system with a voltage and current controller to actively compensate the reactive power requirements of the load. The STATCOM’s performance depends on the underlying controller operation. Consequently, controllers that can handle sudden load changes and disturbances optimally with faster response is desired. For this purpose, this investigation presents a controller design that handles nonlinear load and ensures optimal operation. The proposed controller consists of a fuzzy-PID controller that is capable of handling non-linear dynamics and sudden changes in the load. In order to tune parameters, the bio-geography based optimization method used to determine the fuzzy-PID controller parameters that minimizes the integral absolute error of the controller. Furthermore, a single objective function is formulated on top of the fuzzy-PID controller. It consists of importance weighted performance metrics such as integral squared error and integral absolute errors of the voltage and current controllers, respectively. The significance weights are determined by interactive decision tree technique to detect weaker metrics and improve their influence in the control scenarios. The algorithm terminates once all the metrics are tuned to satisfaction with operator’s consent. The proposed IDT based fuzzy PID controller is tested on a power system containing nonlinear load, and its performance is compared with existing controller. The proposed controller improves the reactive power compensation offered by the STATCOM significantly.