Hybrid energy storage sizing in energy hubs: A continuous spectrum splitting approach

Renewable powered energy hub is a promising way to realize efficient use of renewable resources through multi-energy integration. To cope with the fluctuation of renewable power at different timescales, long-term and short-term energy storage devices are essential. This paper proposes a frequency-domain approach to determine the appropriate capacities of hydrogen and battery energy storage units in an electricity-hydrogen-heat integrated energy hub. The net demand is mapped to the frequency domain via discrete Fourier transformation (DFT). A continuous spectrum splitting method is developed to allocate the frequency components among generator, hydrogen storage and battery storage. Compared with the time-domain method, it makes better use of the spectral characteristics of the year-round data, leading to robust sizing results without complex modeling of uncertainty. In contrast to the existing cut-off frequency method which assigns high-frequency components to battery storage and low-frequency components to generators according to a cut-off frequency, the proposed framework allows a single spectrum component to be allocated to multiple system devices, which not only makes the optimization model convex but also improves optimality. In addition, the spectrum clustering process further reduces the number of core variables and the computational efficiency is high. The rationality of the sizing results is verified through online operation, where the proposed method achieves an average of 1.93% total load shedding rate compared with 4.42% of the time-domain method, verifying the operation reliability under uncertainty. While fully utilizing the spectrum characteristics, the investment cost is about 22% less than the cut-off frequency result.