Handling of long-term storage in multi-horizon stochastic programs

This paper introduces a method for incorporating long-term storage into the multi-horizon modelling paradigm, thereby expanding the scope of problems that this approach can address. The implementation presented here is based on the HyOpt optimization model, but the underlying concepts are designed to be adaptable to other models that utilize the multi-horizon approach. We demonstrate the effects of several formulations on a case study that explores the electrification of an offshore installation using wind turbines and a hydrogen-based energy storage system. The findings suggest that the formulations offer a realistic modelling of storage capacity, without compromising the advantages of the multi-horizon approach.