Multi-microgrid formation (MMGF) is a promising solution for enhancing power system resilience. This paper proposes a new deep reinforcement learning (RL) based model-free on-line dynamic MMGF scheme. The dynamic MMGF problem is formulated as a Markov decision process, and a complete deep RL framework is specially designed for the topology-transformable micro-grids. In order to reduce the large action space caused by flexible switch operations, a topology transformation method is proposed and an action-decoupling Q-value is applied. Then, a convolutional neural network (CNN) based multi-buffer double deep Q-network (CM-DDQN) is developed to further improve the learning ability of the original DQN method. The proposed deep RL method provides real-time computing to support the on-line dynamic MMGF scheme, and the scheme handles a long-term resilience enhancement problem using an adaptive on-line MMGF to defend changeable conditions. The effectiveness of the proposed method is validated using a 7-bus system and the IEEE 123-bus system. The results show strong learning ability, timely response for varying system conditions and convincing resilience enhancement.

This paper proposes a new decentralized restoration scheme in order to realize efficient coordinated restoration of the coupled transmission and distribution (TS-DS) system. First, considering the separated operation of the TS operator (TSO) and DS operators (DSOs) and to avoid the TSO data processing burden, the TS-DS system is divided into several subsystems according to their physical connection. Then, in order to implement the independent but coordinated decision-making of the subsystems, a decentralized decision-making framework is constructed by model decoupling and iterative interaction between TS and DS. Furthermore, for the sake of the convergence of the iterative process with non-convex models, a new decentralized optimization method is developed to obtain the optimal coordinated restoration strategy of the TS-DS system. The proposed decentralized restoration scheme achieves independent decision-making of the TSO and DSOs in the coupled TS-DS system, and solves the non-convergence problem of the decentralized optimization method. The effectiveness of the proposed method is validated using two IEEE standard test systems and a real-world power system.

To enhance the resilience of distribution systems (DSs), networked microgrids (MGs) can be employed to restore critical loads after a blackout. However, uncertainties of renewable energy sources (RESs) and loads bring challenges to load restoration. This paper proposes a distributed risk-limiting load restoration strategy for unbalanced DS with networked MGs. In the proposed strategy, the risk is represented by the conditional value-at-risk (CVaR), based on which a risk-limiting load restoration model is formulated considering renewable and load uncertainties. With the projection function-based alternating direction method of multipliers (P-ADMM) and sub-additivity of the CVaR, the proposed risk-limiting model is decoupled and solved in a distributed manner. This enables distributed risk limiting among MGs. Each MG can make load restoration decisions individually, while ensuring that the risk-limiting constraints of the overall system are satisfied. The modified IEEE 123-node system with networked MGs was used to conduct case studies. Simulation results show that the proposed strategy can restore loads efficiently with a smaller risk index and can provide distributed and flexible risk management. Moreover, the proposed distributed strategy can reduce the computation burden and is more robust against controller failures compared with the centralized strategy.

This paper proposes a new deep learning (DL) based model-free robust method for bulk system on-line load restoration with high penetration of wind power. Inspired by the iterative calculation of the two-stage robust load restoration model, the deep neural network (DNN) and deep convolutional neural network (CNN) are respectively designed to find the worst-case system condition of a load pickup decision and evaluate the corresponding security. In order to find the optimal result within a limited number of checks, a load pickup checklist generation (LPCG) algorithm is developed to ensure the optimality. Then, the fast robust load restoration strategy acquisition is achieved based on the designed one-line strategy generation (OSG) algorithm. The proposed method finds the optimal result in a model-free way, holds the robustness to handle uncertainties, and provides real-time computation. It can completely replace conventional robust optimization and supports on-line robust load restoration which better satisfies the changeable restoration process. The effectiveness of the proposed method is validated using the IEEE 30-bus system and the IEEE 118-bus system, showing high computational efficiency and considerable accuracy.