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.

With the expansion of power grids and the growth of renewable energy generation, more uncertainties of topology (e.g. components outages) and injection (e.g. renewable energy outputs variations) need to be analyzed. As a result, large numbers of system states have to be evaluated by optimal power flow (OPF) and this brings a significant challenge to the efficiency of reliability assessment. To address this, a Lagrange-multiplier-based reliability assessment method (LM-T&I) is proposed to accelerate the evaluation of system states. The Lagrange-multiplier-based function is constructed to obtain the optimal load shedding of topology changes and injection variations, avoiding time-consuming OPF computations. Moreover, combined with the impact-increment-based state enumeration method (IISE), the computational efficiency can be further improved. Case studies are conducted on the RTS-79 and IEEE 118-bus systems. The scalability of the LM-T&I method is verified on the practical Brazilian system. The results demonstrate that the LM-T&I method has a superior performance on the reliability assessment compared with traditional methods.

To deal with uncertainties of renewable energy, demand and price signals in real-time microgrid operation, this paper proposes a model predictive control strategy for microgrid economic dispatch, where hourly schedule is constantly optimized according to the current system state and latest forecast information. Moreover, implicit network topology of the microgrid and corresponding power flow constraints are considered, which leads to a mixed integer nonlinear optimal power flow problem. Given the non-convexity feature of the original problem, the technique of conic programming is applied to efficiently crack the nut. Simulation results from a reconstructed IEEE-33 bus system and comparisons with the routine day-ahead microgrid schedule sufficiently substantiate the effectiveness of the proposed MPC strategy and the conic programming method.

Hydrogen and renewable electricity-based microgrid is considered to be a promising way to reduce carbon emissions, promote the consumption of renewable energies and improve the sustainability of the energy system. In view of the fact that the existing day-ahead optimal operation model ignores the uncertainties and fluctuations of renewable energies and loads, a two-stage energy management model is proposed for the sustainable wind-PV-hydrogen-storage microgrid based on receding horizon optimization to eliminate the adverse effects of their uncertainties and fluctuations. In the first stage, the day-ahead optimization is performed based on the predicted outpower of WT and PV, the predicted demands of power and hydrogen loads. In the second stage, the intra-day optimization is performed based on the actual data to trace the day-ahead operation schemes. Since the intra-day optimization can update the operation scheme based on the latest data of renewable energies and loads, the proposed two-stage management model is effective in eliminating the uncertain factors and maintaining the stability of the whole system. Simulations show that the proposed two-stage energy management model is robust and effective in coordinating the operation of the wind-PV-hydrogen-storage microgrid and eliminating the uncertainties and fluctuations of WT, PV and loads. In addition, the battery storage can reduce the operation cost, alleviate the fluctuations of the exchanged power with the power grid and improve the performance of the energy management model.