Optimal PMU placement approach for power systems considering non-Gaussian measurement noise statistics. (March 2021)

• Record Type: Journal Article
• Authors: Chen, Tengpeng; Cao, Yuhao; Chen, Xuebing; Sun, Lu; Zhang, Jingrui; Amaratunga, Gehan A.J.
• Is Part Of: International journal of electrical power & energy systems. Volume 126(2021)Part A
• Publication Date: 2021-03
• Subjects: Optimal PMU placement -- Robust estimator -- Non-Gaussian noise -- Influence function -- State estimation; Electrical engineering -- Periodicals; Electric power systems -- Periodicals; Électrotechnique -- Périodiques; Réseaux électriques (Énergie) -- Périodiques; Electric power systems; Electrical engineering; Periodicals; 621.3
• DOI: 10.1016/j.ijepes.2020.106577
• Languages: English
• ISSNs: 0142-0615

Abstract

Highlights: A new Gain matrix is derived from the influence function when the probability density function of measurement noise is known. The inverse of the new Gain matrix can be utilized to approximate the covariances of estimated states of robust estimators even when the distribution of measurement noise is non-Gaussian. A new optimal PMU placement approach based on non-Gaussian noise and robust estimators is proposed to deal with the monitoring system upgrade in reality. Abstract: This paper investigates how to add a limited number of Phasor Measurement Units (PMUs) to the existing monitoring system so as to improve the estimation accuracy further. The existing methods are usually based on Gaussian noise assumption and the weighted least squares (WLS) estimator is taken into account. However, the Gaussian noise assumption is not always true in reality and the WLS is non-robust in this case. This paper proposes a new optimal PMU placement approach where the distribution of measurement noise can be non-Gaussian or Gaussian and many robust estimators such as the maximum likelihood estimator, Multiple-Segment, Quadratic-Linear, Square-Root and Schweppe-Huber Generalized-M estimator are considered. Based on the new Gain matrix obtained from the influence function approximation, the D-optimal and E-optimal experiment criterions are exploited in the optimal PMU placement problem. A convex relaxation in conjunction with an optimization improvement method based on the Fedorov exchange algorithm is utilized to solve the optimizing problem. Simulations on the IEEE 57-bus system and the Polish 2383-bus system are carried out to evaluate the effective performance of the proposed approach.