Role of Fuzzy Sets in Power System State Estimation

Abstract

This paper presents an enhancement to the least median of squares (LMS) estimation in the power system state estimation. A weighted least square estimation may be used in conjunction with LMS, yet LMS could face difficulties for eliminating bad data in specific cases. The proposed method, denoted as fuzzy LMS+LAV estimation, employs the applications of fuzzy sets to improve the accuracy of LMS. The proposed method applies the LMS method to determine the bounds of residuals and then applies fuzzy sets to the least absolute value (LAV) estimator to eliminate bad data. The paper shows that the proposed method produces more accurate estimate of power system states than that of LMS. The complementary advantages of fuzzy LAV and LMS estimators without lengthy computational efforts, which are targeted by this paper include: * less computational effort than the fuzzy LAV. * better accuracy than LMS. * same if not better robustness than both estimators. A qualitative comparison illustrates the superior robustness of the proposed method in the presence of leverage bad data.