Holistic management of mobile energy storage fleets for more resilient service restoration

• Main Author: Yao, Shuhan
• Other Authors: Wang Peng
• Format: Thesis-Doctor of Philosophy
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/140411
• Institution: Nanyang Technological University

The world economy continues to electrify. The power grid constitutes a vital cornerstone of critical infrastructures and serves as an essential foundation for the economy and society. Recent major blackouts caused by extreme weather events lead to catastrophic consequences. The impacts of extreme weather events pose unprecedented challenges to power grids and emphasize the importance of improving system resilience. It is indispensable to restore the electric service effectively in response to severe power outages, thus achieving more resilient distribution systems. When severe blackouts occur, a variety of local resources, e.g., microgrids and distributed energy resources, can be utilized to restore critical loads in distribution systems. Moreover, the emerging mobile energy storage systems (MESSs) can provide temporal-spatial mobility and coordinate with stationary local resources for an integrated distribution system restoration. To fully leverage the mobility and flexibility of MESS fleets in distribution system restoration, this thesis proposes a joint post-disaster restoration scheme for MESS fleets and generation scheduling in microgrids and network reconfiguration to minimize the total system cost, including customer interruption cost, generation cost, and MESS related costs. A temporal-spatial MESSs model that is related to both transportation networks and distribution systems is proposed to represent the difference between MESSs and energy storage systems (ESS)s in terms of flexibility and cost reduction of ESSs sharing among microgrids. The proposed restoration problem is formulated as a mixed-integer linear program (MILP) with considering various network and MESS constraints. Compared to conventional restoration strategy, the MESS can efficiently transfer energy among multiple micros within the distribution systems in appropriate times and locations, facilitating critical loads service restoration. Moreover, this thesis develops restoration strategies under uncertainties. A novel stochastic service restoration strategy is studied to coordinate the dynamic scheduling of MESS, resource dispatching of microgrids, and distribution network reconfiguration. It takes into account damage and repair to both the roads in transportation networks and the branches in distribution systems. The uncertainties in load consumption and the status of roads and branches are modeled as scenario trees using Monte Carlo simulation method. The operation strategy of MESSs is modeled by a stochastic multi-layer time-space network technique. In order to take advantages of multiple source data that improve situational awareness during the restoration process, a rolling optimization framework is adopted to dynamically update system damage, and the coordinated scheduling at each time interval over the prediction horizon is formulated as a two-stage stochastic mixed-integer linear program with temporal-spatial and operation constraints. The integrated restoration strategy is demonstrated to perform effectively under uncertainties in coupled transportation and distribution systems. Finally, this thesis investigates an on-line optimization framework inspired by deep reinforcement learning (DRL) for solving challenging large scale decision-making problems. The decision-making problem under uncertainties is formulated using Markov decision process (MDP) and solved iteratively by data-driven DRL algorithms. An MDP formulation for an integrated service restoration strategy is investigated to coordinate the scheduling of MESSs and resource dispatching of microgrids. The uncertainties in load consumption are taken into account. The deep reinforcement learning (DRL) algorithm is utilized to solve the MDP for optimal scheduling. The deep deterministic policy gradient (DDPG) and twin delayed DDPG (TD3) are applied to train the deep Q-network and policy network, then the well-trained policy can be deployed in on-line manner for solving restoration strategy efficiently. All the proposed methodologies and algorithms are verified by simulations and implemented in Python. Numerical results demonstrate the effectiveness and superiority of the proposed methods.