Voltage management in the presence of Distributed Energy Resources - Field implementation of a robust Distribution State Estimator with errors in sensor data

Abstract

The deployment of Distributed Energy Resources (DER) complicates voltage management on existing MV networks and requires dynamic centralized setting of On Load Tap Changer (OLTC) set point to optimize the network DER hosting capacity. In this frame, Enedis (the French DSO) has been experimenting a Weighted Least Square (WLS) Distribution State Estimator (DSE) in VENTEEA and Smart Grid Vendée (SGV) demonstrators. This algorithm uses sensor-based measurements and pseudo-measurements to determine the state of a network at any time and to allow the OLTC dynamic management. The VENTEEA experimentation showed that the DSE behaviour can be strongly affected by sensor outliers. These bad data could be due to many reasons such as sensor misuse or failure. As the occurrences of this bad data are not negligible, the accuracy of the DSE becomes inacceptable to Enedis and a way to improve the DSE algorithm had to be found. This paper deals with improving the robustness of the DSE against sensor errors and failures.EDF R&D and Enedis studied, then implemented an M-estimator running alongside the operational WLS estimator on a substation MV networks of the Smart Grid Vendée demonstrator. This paper introduces the M-estimator principle, the way it deals with measurements errors, how to set it up, and the outcomes achieved by both simulation and field tests. It also compares the results of the M-estimator against the WLS state estimator.