Enabling Resiliency Through Outage Management and Data-Driven Real Time Aggregated DERs

Distribution power grids are rapidly changing from a centralized and unidirectional distribution network to a decentralized, bidirectional network that supports the penetration of large amounts of observable/unobservable distributed energy (DERs). The additional variability, uncertainty and scale of control variables generated by DER integration greatly increase the complexity of power system operation. This paper presents a method to improve the resiliency of distribution systems using increasing distributed energy resources (DERs). DERs and loads are aggregated into a switch-level equivalent model while retaining controllability information for real-time power outage management analysis after a power outage. Pre-fault smart meter measurements serve as input to a machine learning integrated cold load pickup (CLPU) model to obtain short-term load predictions, while optimal recovery analysis is carried out to determine restored network topology and utility control loads and DERs set-points.

I. Introduction

The increasing of distributed energy resources brings new challenges to the safe operation of the power system, especially the observability and controllability during an outage. However, with the digitization of the grid and deployment of intelligent electronic devices (IEDs), more reliable data is becoming available. Therefore, data-driven technology in this area has become a viable solution and offers many new opportunities.