Identifying a practical state estimation scheme incorporating substation measurement dependencies

Abstract

This paper evaluates existing methods to improve state estimation by incorporating substation measurement dependencies. Three main approaches are considered: a traditional weighted least squares scheme, an iterative algorithm using a 2-point estimate to approximate the covariance matrix, and a variant of the algorithm taking just the diagonal of the estimated covariance matrix. These schemes are tested under uniform conditions for systems ranging from 3 to 30 busses, and the results are analyzed to gain intuition about each method. Finally, this intuition is used to recommend a practical weighted least squares state estimation scheme which could potentially be used to improve online state estimation.