A novel design of WACS based multi-output support vector machine (M-SVM) for oscillation damping on power system

Abstract

This paper proposes a novel design of wide area control system (WACS). WACS is utilized to damp the oscillation on power system. WACS consists of wide area monitor (WAM) and wide area control (WAC). WAM is used to monitor the dynamic behavior of power system, while WAC is used as the additional controller on power system. The proposed method is called WACS based multi-output support vector machine (M-SVM). M-SVM used in this paper is M-SVM for regression. The input signal which used by WAM is the input signal to the AVR ΔVwi, the mechanical power ΔPmi and the electrical power ΔPei of each generator. The output of WAM is utilized to predict the speed deviation of the generator. WAC is using the input signal of mechanical power ΔPmi and electrical power ΔPei from generator, while the output of WAC is the signal control which injected to the AVR ΔVwi. A two-area-four-generator is utilized as a tested system to evaluate the performance of the proposed method. From the simulation results that has been conducted, the proposed method can reduce the overshoot and compress the settling time better than other methods that presented in this paper.