A Real-Time Robust Method for Post-Disaster Load Restoration of Coordinated Power-Transportation System With Vehicle-to-Grid Response
Houbo Xiong, Yan Xu, Zhao Yang Dong, Wei Gan, Chuangxin Guo, Mingyu Yan
Abstract
With the proliferation of electric vehicles (EVs), vehicle-to-grid (V2G) capability emerges as a potential resource for load restoration after a large disruption. This paper presents a real-time post-disaster load restoration method for the coordinated power distribution networks (PDN) and urban traffic networks (UTN) with V2G response. The multi-period restoration problem is modeled as a dynamic programming-based multi-stage robust optimization model, addressing uncertainties of renewable generation and traffic demands. It incorporates a dynamic traffic assignment scheme to characterize vehicle travels and V2G services within short time slots. Then, an improved robust dual dynamic programming algorithm is proposed to solve the multi-stage robust optimization problem. For online application, the solved value functions from each stage serve as per-period policies, leveraging knowledge of future uncertainties to quickly guide real-time load restoration through distributed resource dispatch, network reconfiguration, and V2G assignments. Numerical experiments with a 33-bus PDN and 20-road UTN, plus a real-world 91-bus PDN with 35-road UTN, validate the effectiveness of proposed restoration method.