27 March 2023 in IEEE Transactions on Smart Grid
The networked Multi-Energy Microgrids (MEMGs) can significantly improve the economic efficiency of operation through energy cooperation and complementarity. However, the complex energy trading mechanism and multi-energy coupling decision process present various challenges for the optimal operation of MEMGs, such as the large decision space and the difficulty of converging to the global optimal solution. This study presents a novel energy management method for MEMG based on an improved Deep Q-network (IDQN) to solve these problems. Firstly, the equivalent model of an external interactive environment is constructed for each MEMG by using a novel Kriging surrogate enhanced Gate Recurrent Unit-Temporal Convolutional Network (GRU-TCN), which requires only the accessible energy exchange data and the public and independent information as input. The computational complexity of the reinforcement learning reward function is reduced accordingly by using this equivalent model. Subsequently, an improved k-crossover sampling strategy with reduced access frequency of the low reward actions is proposed to replace the traditional.-greedy strategy to further improve the original DQN. This effectively addresses the problem of inefficient exploration in a large-scale action space. Lastly, the proposed method is verified using a test system with three MEMGs. The result demonstrates that the proposed IDQN method presents better convergence and stability than the traditional DQN, and presents better energy management and operational efficiency.
20 April 2023 in Energy Reports
26 June 2024 in IEEE Transactions on Smart Grid
With the evolution of the power market, the active involvement of prosumers in both consuming renewable energy and maximizing financial gains has emerged as a pivotal and compelling trend in the prospective landscape of energy market development. However, the optimal energy trading strategy for prosumers under a complicated peer-to-peer (P2P) market environment is still a great challenge, due to the incomplete information and large decision space. To solve this problem, this paper proposes an improved deep deterministic policy gradient (IDDPG) algorithm to optimize the P2P energy trading and operation strategy for prosumers. The proposed IDDPG algorithm uses a wave-attention network to model the external interactive environment of each prosumer to simplify the complexity of multi-agent trading environment. Moreover, a priority sampling mechanism is proposed to learn valuable experiences purposefully and adopts a cyclic learning rate to avoid local optimality. Finally, the feasibility of the method is verified by the case study of P2P trading with different scales of prosumers. The numerical results show that the proposed IDDPG algorithm outperforms the traditional reinforcement learning algorithms in both calculating efficiency and convergence, and provides a reference for the P2P transaction strategy of prosumers.
30 October 2025 in International Journal of Electrical Power & Energy Systems
2 June 2026 in Energy and AI
26 February 2026 in SSRN Electronic Journal
Transient stability assessment (TSA) plays a crucial role in the analysis of power system. Owing to the large-scale integration of distributed energy resources (DERs), the resulting proliferation of inverters further exacerbates the nonlinearity and high dimensionality of power system. However, existing physics-driven methods often yield conservative results and involve high computational complexity, while learning based data-driven approaches suffer from limited interpretability. To address these challenges, this paper proposes a novel physics-informed deep learning framework to evaluate the transient stability of power systerm. First, an adjoint-based physics-dynamics learning (APDL) method is introduced to capture the transient behavior under arbitrary initial conditions. Next, leveraging these transient results, a Transformer-based multi-class classification model is developed to further identify the transient stability under various disturbances. Finally, the effectiveness and feasibility of the proposed method are validated through case studies on single machine infinite bus system, IEEE-39 bus system and WECC-240 bus system. The results suggest that, compared with the conventional PINN and data-driven approaches, the proposed method improves accuracy by 9.18\% and 5.94\%, respectively.