Research on collaborative scheduling optimization of multienergy systems in parks based on deep reinforcement learning

With the rapid development of energy Internet technology, the problem of coordinated dispatch optimization of multi- energy systems in parks has attracted increasing attention. This paper proposes a method for coordinated dispatch optimization of multi-energy systems in parks based on deep reinforcement learning. Firstly, a model of multi-energy systems in parks, which includes multiple energy forms such as electricity, heat, cooling and gas, is constructed, and the coupling relationship and energy conversion characteristics between subsystems are considered. Secondly, are inforcement learning algorithm based on deep Q-network (DQN) is designed to solve the high-dimensional complex decision-making problem of multi-energy system dispatch optimization. The algorithm adopts techniques such as experience replay and target network to effectively improve the learning efficiency and convergence performance. Then, a multi-objective reward function considering system operation constraints and user needs is proposed, and the comprehensive optimization of economy, environmental protection and reliability is achieved. Finally, the effectiveness of the proposed method is verified by simulation cases. The results show that compared with traditional methods, the dispatch optimization method based on deep reinforcement learning proposed in this paper can effectively reduce the system operation cost, improve the utilization rate of renewable energy, and ensure that the energy demand of users is met. This method provides a new solution for the intelligent coordinated dispatch of multi-energy systems in parks, which is of great significance to promote the practical application of energy Internet technology.