Unit Output Optimization of Integrated Energy System Based on Digital Twin and Reinforcement Learning

The integrated energy system is different from the power system in that it integrates a variety of energy forms and coupling devices. The structure and scale are more complex, and the collection of system parameters is brutal. In this case, how to optimize the unit output to reduce the operation cost and maintain the safe and stable operation of the system has become an important research direction. This paper presents a method to build a virtual environment of an integrated energy system based on digital twin technology and find the optimal output strategy of the unit through a deep reinforcement learning decision model. Firstly, the real-time operation data of the integrated energy system are collected by multi-dimensional sensors to build a digital twin virtual environment. Secondly, the real-time output of the unit is simulated through the digital twin, which provides the relevant operation parameter basis for the follow-up deep reinforcement learning decision-making. Finally, the deep reinforcement learning decision model is used to find the optimal strategy for unit output, and the information is fed back to the terminal controller in real-time through the digital twin to guide the output of the integrated energy system unit. The simulation results show that the proposed method can effectively improve the unit output optimization accuracy of the integrated energy system and reduce the system operation cost.