Robust state estimation of electric power systems

Abstract

The exact fit points of the Least Median of Squares (LMS) and the Least Trimmed Squares (LTS) estimators in electric power systems are investigated. The expression of the maximum possible exact fit point /spl delta/*/sub max/ is derived, and the corresponding quantile index /spl nu/ of the ordered squared residual is determined. It is found that /spl delta/*/sub max/ as well as /spl nu/ hinge on the surplus of the network, defined as one less than the smallest number of measurements whose deletion from the data set decreases the rank of the Jacobian matrix. Based on the surplus concept, a system decomposition scheme is developed; it significantly increases the number of outliers that can be handled by the LMS and the LTS estimators. In addition, it dramatically reduces the computing time of these estimators, opening the door to their application in a real-time environment, even for large-scale systems. >