Bi-Level Planning of Microgrid Considering Seasonal Hydrogen Storage and Efficiency Degradation of Electrolyzer

Abstract

Microgrids that contain a high percentage of renewable energy face the challenge of having insufficient resources for long-term regulation of the energy balance. Seasonal hydrogen storage emerges as a promising option. To analyze the feasibility and economic viability of seasonal hydrogen storage in microgrids, this paper proposes a bi-level planning approach. First, considering the dynamic characteristics of electrolyzer power-efficiency-degradation, a linearized method for calculating the dynamic efficiency is proposed. Taking into account the randomness of the data, a data-driven Markov Chain Monte Carlo method is used to generate typical daily time series. The upper level of the bi-level model focuses on investment decisions for electrolyzer capacity and seasonal hydrogen storage, while the lower level optimizes system operational revenue and electrolyzer usage costs. Then, addressing the nonlinear term issue in optimization calculations, a piecewise McCormick envelope method considering the power characteristics of the electrolyzer is proposed to convexify the optimization problem. The results of the case study show that the proposed planning method can increase the annual revenue by 11% and 595% compared to only considering constant efficiency and fixed electrolyzer lifespan. Additionally, the convex relaxation method enhances convergence speed while maintaining solution accuracy.