Operation optimization for integrated energy system with energy storage

Dear editor, The integrated energy system has attracted widespread attention owning to its high energy efficiency, low emission, and reliability [1]. It is a typical cyber-physical integrated energy system (CPIES) that combines information and energy technologies, in which the cyber information platform collects information from the energy plant and controls most of the energy devices [2]. Many optimized operation strategies have been presented for the information platform to guarantee that the energy plant works stably and efficiently. Generally, the operational optimization of energy systems needs to solve a combinatorial optimization problem. Several methods have been employed to calculate the optimal set-point of each energy device, including the genetic algorithm (GA), particle swarm optimization (PSO), non-linear programming (NLP), and mixed-integer non-linear programming (MINLP). However, the integrated energy system usually uses an energy storage unit to address fluctuations in load demands [3]. For this complex energy system, the above methods need to transform the optimization of the dynamic storage process into a high-dimensional non-linear optimization problem. However, it is hard to solve this optimization problem with multiple constraints. Some studies combined fixation strategies with optimization methods to achieve a satisfying solution [4–6]. Deng et al. [4] developed a fixation storage strategy that stores thermal energy at night and supplies energy in the daytime, and employed MINLP to determine the optimal set-points of other devices. Bao et al. [5] defined that the energy storage unit continuously runs at valley time and presented an improved PSO method that can be used to correct the particle’s position to satisfy the non-linear constraints. Zheng et al. [6] determined the set-points of the storage unit and generator by calculating the shortest distance between the load points and the output curves of the energy equipment. Nevertheless, these methods cannot guarantee that the obtained operational strategy is optimal for an integrated energy system with energy storage. In this study, a hybrid method based on dynamic programming (DP) and GA is proposed in order to find the optimal operational strategy for an integrated energy system. Compared with the existing methods, the proposed method does not require the conversion of the optimization of the energy storage to a high-dimensional static optimization problem. Instead, it uses DP as the main optimization framework to calculate the optimal energy storage state for the entire optimization period. The GA is then embedded into the DP to optimize the corresponding set-points of the device in each energy storage state for every period. DP can be used to find the global optimal solution to the energy storage problem. The determination of optimal device set-points is guaranteed by the GA because there are only one or two optimiza-