Because of the high penetration of renewable energies and the installation of new control devices, modern distribution networks are faced with voltage regulation challenges. Recently, the rapid development of artificial intelligence technology has introduced new solutions for optimal control problems with high dimensions and dynamics. In this paper, a deep reinforcement learning method is proposed to solve the two-timescale optimal voltage control problem. All control variables are assigned to different agents, and discrete variables are solved by a deep Q network (DQN) agent while the continuous variables are solved by a deep deterministic policy gradient (DDPG) agent. All agents are trained simultaneously with specially designed reward aiming at minimizing long-term average voltage deviation. Case study is executed on a modified IEEE-123 bus system, and the results demonstrate that the proposed algorithm has similar or even better performance than the model-based optimal control scheme and has high computational efficiency and competitive potential for online application.

With the development of integrated energy systems (IES), the traditional demand response technologies for single energy that do not take customer satisfaction into account have been unable to meet actual needs. Therefore, it is urgent to study the integrated demand response (IDR) technology for integrated energy, which considers consumers’ willingness to participate in IDR. This paper proposes an energy management optimization method for community IES based on user dominated demand side response (UDDSR). Firstly, the responsive power loads and thermal loads are modeled, and aggregated using UDDSR bidding optimization. Next, the community IES is modeled and an aggregated building thermal model is introduced to measure the temperature requirements of the entire community of users for heating. Then, a day-ahead scheduling model is proposed to realize the energy management optimization. Finally, a penalty mechanism is introduced to punish the participants causing imbalance response against the day-ahead IDR bids, and the conditional value-at-risk (CVaR) theory is introduced to enhance the robustness of the scheduling model under different prediction accuracies. The case study demonstrates that the proposed method can reduce the operating cost of the community under the premise of fully considering users’ willingness, and can complete the IDR request initiated by the power grid operator or the dispatching department.