Use of PMUs in WLS and LAV based state estimation

Abstract

This presentation is motivated by the recent increase in the number of phasor measurement units (PMU) installed in substations which suggests that it is not unrealistic to assume that most transmission systems will soon be fully observable by only PMU measurements. When there are sufficient number of PMU measurements to make the system observable, then state estimation problem can be formulated in a simpler manner due to the linearity of the measurement equations. The linear characteristics of PMU measurements also facilitate easy avoidance of leverage measurements which have been a major obstacle in maintaining robustness of estimators against bad data. In this presentation we will focus on two different estimation approaches, namely the weighted least squares (WLS) and least absolute value (LAV) estimators, both using only PMU measurements. We will show that there are some computational simplifications that implicitly exist in WLS formulation when only phasor measurements are considered. These simplifications allow computational savings to be realized in implementing this type of estimators. In the case of LAV estimators, it will be shown that their computational performance becomes comparable and in some cases even better than their WLS counterparts and therefore such estimators can be viable alternatives particularly for systems where communication errors are more common. Simulation results of implementing these estimators for typical power systems will be presented to illustrate the claimed properties of these estimators.