Bayesian Approach for Distribution System State Estimation With Non-Gaussian Uncertainty Models

To deal with the increasing complexity of distribution networks that are experiencing important changes, due to the widespread installation of distributed generation and the expected penetration of new energy resources, modern control applications must rely on an accurate picture of the grid status, given by the distribution system state estimation (DSSE). The DSSE is required to integrate all the available information on loads and generators power exchanges (pseudomeasurements) with the real-time measurements available from the field. In most cases, the statistical behavior of the measured and pseudomeasured quantities cannot be approximated by a Gaussian distribution. For this reason, it is necessary to design estimators that are able to use measurements and forecast data on power flows that can show a non-Gaussian behavior. In this paper, a DSSE algorithm based on Bayes's rule, conceived to perfectly match the uncertainty description of the available input information, is presented. The method is able to correctly handle the measurement uncertainty of conventional and synchronized measurements and to include possible correlation existing between the pseudomeasurements. Its applicability to medium voltage distribution networks and its advantages, in terms of accuracy of both estimated quantities and uncertainty intervals, are demonstrated.