Data-Driven, Multi-Region Distributed State Estimation for Smart Grids

Abstract

Real-time wide-area monitoring of smart grids demands a low latency data processing of power system data. To enable the low latency requirements and to avoid the large overhead of communicating a large volume of time-sensitive data to central processing units, distributed and local processing of data is a promising approach that can improve system monitoring functions. Data-driven state estimation in power systems is an example of functions that can benefit from distributed processing of data and enhance the real-time monitoring of the system. In this paper, distributed state estimation is considered over multi-region, identified based on geographical distance and correlations among the state of the power system's components. Bayesian Multivariate Linear Regression (BMLR) combined with Auto-Regressive AR(p) process for distributed state estimation is considered over the multi-region power system. The performance of the distributed data-driven state estimation method and the role of regions are evaluated using the IEEE 118 test case under normal conditions as well as partially unobservable scenarios.