Situation-Aware Load Restoration Considering Uncertainty and Correlation

In order to minimize power supply interruption to customers, power system restoration needs to be performed rapidly after a blackout. With the integration of intermittent renewable energy and flexible loads, power system load patterns are becoming more uncertain. The goal of this paper is to tackle the challenging task of developing a secure yet speedy load restoration planning strategy considering load uncertainty. The key is to leverage the real-time measurements of the restored load blocks to refine the statistical knowledge about the unrestored load blocks derived from historical data, such that the load restoration plans can be less conservative without harming security. Specifically, as the restoration process rolls forward and more loads are restored, we use a Kalman filter to update the probability distribution of the unrestored loads using the newly available measurements from the restored loads. With the updated distribution, a rolling-horizon load restoration problem for load-pickup decision making is formulated with given specified confidence levels under frequency, voltage, and power flow constraints. The effectiveness of the proposed method is evaluated through case studies on the IEEE 39-bus test system and the NPCC 140-bus system. Results demonstrate that the proposed method can restore loads more quickly and securely than conventional methods under load uncertainty.