Future energy systems aiming to achieve high shares of renewable energies need energy storage systems to mitigate the highly volatile electrical energy generation from renewable sources like wind farms and photovoltaic parks. Furthermore, seasonal energy storage is needed to decarbonize the electrical grid fully. Thus, energy storage solutions for short-term fluctuations to seasonal variations in energy availability are needed. Decentralized hydrogen energy storage systems, in combination with battery storage systems, are promising solutions for energy storage, as renewable energy sources are distributed, hydrogen storage systems are scalable, and battery storage systems are highly efficient. The control of hybrid battery-hydrogen energy storage systems is challenging due to the high system complexity and the wide range of system dynamics up to seasonal variations. Mathematical optimization problems can be utilized in a model predictive control framework to derive the optimal operational strategy for such an energy storage system. However, current model predictive control frameworks do not sufficiently consider seasonal variations in energy availability due to the computational complexity that arises while solving optimization problems with long prediction horizons. This work presents a two-level, hierarchical, model-predictive control approach to integrate a long-term operating strategy into the scheduling of an energy system. Therefore, two optimization problems are developed to manage the challenges arising from the various time scales that need to be considered. For each optimization problem, methods to reduce the computational complexity are devised to enable the utilization of detailed mathematical models, accounting for component degradation, minimal operational times, part-load constraints, and non-linear operational behaviors while utilizing long-prediction horizons. The control approach investigated in this work is developed for and applied to a novel hybrid battery-hydrogen energy storage system considering liquid organic hydrogen carriers for long-term hydrogen storage.The hierarchical model predictive control framework is compared to an adapted model predictive controller and a classical rule-based controller. The performance of all three control approaches is evaluated by applying the derived operational strategy to a detailed simulation model of the energy storage system. The results show that the hierarchical model predictive controller outperforms both control approaches with and without perfect foresight for deriving a long-term strategy. Furthermore, an investigation of the component sizing of the energy system showed that the hierarchical model predictive control framework improves the performance of the energy storage system for a wide range of configurations compared to a rule-based controller.
Alexander Holtwerth (Wed,) studied this question.