Forecasting-Aided State Estimation in Distribution Networks Considering Behind-the-Meter Resources

Abstract

Distribution system state estimation (DSSE) is essential in active distribution networks (ADNs). However, new emerging load demands, including the profiles of behind-the-meter (BTM) resources and responsive loads, increase the uncertainty during pseudo-measurement generation. Most existing methods struggle to accurately generate pseudo measurements of load demands incorporating the BTM resources. To address this, a dynamic equivalent model (DEM) for each node within the distribution system is constructed using the Numerical algorithm for Subspace State Space System IDentification (N4SID) algorithm to mimic the dynamics of these nodes. These DEMs are then integrated into an unscented Kalman filter (UKF) to predict and refine the measurements. The output of the UKF stage is utilized to perform the state estimation process using forecasting-aided state estimation (FASE). The proposed approach was tested and validated using the modified IEEE-123 tests system. The numerical results demonstrated the effectiveness of the proposed DEMs, reducing the mean absolute error (MAE) from 5.90 kW, achieved by the state-of-the-art deep learning models, to 2.69 kW with UKF prediction and further to 1.64 kW with UKF correction.