State Estimation in Smart Distribution Systems with Deep Generative Adversary Networks

Abstract

The problem of distribution system state estimation using smart meters and limited SCADA (Supervisory Control and Data Acquisition) measurement units is considered. To overcome the lack of measurements, a Bayesian state estimator using deep learning is proposed. The proposed method consists of two steps. First, a deep generative adversary network is trained to learn the distribution of net power injections at the loads. Then, a deep regression network is trained using the samples generated from the generative network to obtain minimum mean-squared error (MMSE) estimate of the system state. Our simulation results show the accuracy and the online computation cost of the proposed method are superior to the conventional methods.