A new algorithm for improving the numerical stability of power system state estimation

State estimation (SE) is based on an iterative process for solving weighted least squares (WLS) via the so-called normal equations (NE). This process is prone to numerical instability and may lead to an ill-conditioned state estimator in many cases. Although several methods have been proposed in the past to deal with the ill-conditioning problem in high-voltage transmission systems, the SE stability of low-voltage distribution systems remains a challenge due to fewer measurements and high R/X ratios. This paper highlights the main reasons for the ill-conditioning problem and focuses on the impact of the R/X ratio of the distribution systems. NE-based SE uses a comparatively simple algorithm with much lower storage size. Hence, this paper proposes a regularized version of the WLS state estimator to solve the problem of ill-conditioning using an adjustable regularization parameter. Simulation results on U.K. 18-bus and radial 33-bus distribution systems show improved performance in terms of reducing the computational complexity, measurements redundancy, the impact of high R/X ratios and improving the accuracy and numerical stability of the SE solution. (c) 2018 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.