Adaptive time granularity-based coordinated planning method for the electric-hydrogen coupled energy system with hybrid energy storage

• A novel adaptive time granularity-based planning method is proposed for electric-hydrogen coupled system. • The optimal time granularity combination is obtained to fit the characteristics of hybrid energy storage system. • The proposed method shows better performance in terms of computation time and accuracy. The electric-hydrogen coupled system (EHCS) is becoming an important way for renewable energy consumption and energy low-carbon transformation due to its ability of long-term energy transport and short-term power regulation. For the capacity configuration of EHCS, the traditional method with fixed time granularity is difficult to balance the contradiction between model complexity, computational cost and model accuracy. To this end, this paper proposes an adaptive time granularity-based coordinated planning method. Firstly, the seasonal-trend decomposition using losses (STL) algorithm is used to extract the characteristics of intra-day variation and seasonal fluctuation of net loads. On this basis, ward clustering algorithm is applied to realize the vertical typical day selection and horizontal time granulation. The optimal particle number of typical days and seasonal component is determined based on an improved PSO algorithm. Then, a planning model is constructed to determine the capacity of key devices of EHCS, whose results are used in reverse for updating the particle number of PSO algorithm, so that the optimal combination of time granularities can be realized according to the long-term and short-term operational characteristics of different devices. Finally, the effectiveness of the proposed method is verified based on numerous simulation analysis. Compared with the benchmark case based on 8760 h, the average planning error is only 4.52%, while the computation time is reduced by 65.13%, and the average planning error is improved by 10.23% compared with the fixed granularity method.