The waste heat generated by high-performance computing (HPC) represents an opportunity for advancing the decarbonization of energy systems. Seasonal storage is necessary to regulate the balance between waste heat production and demand. High-temperature aquifer thermal energy storage (HT-ATES) is a particularly well-suited technology for this purpose due to its large storage capacity. However, integrating HT-ATES into energy systems for district heating is complex, affecting existing components. Therefore, this study applies a bi-objective mixed-integer quadratically constrained programming (MIQCP) approach to optimize the energy system at Forschungszentrum Jülich (FZJ) regarding total annualized costs (TAC) and global warming impact (GWI). The exascale computer Jupiter, which is hosted at FZJ, generates a substantial amount of renewable waste heat that is suitable for integration into district heating networks and seasonal storage. Case studies show that HT-ATES integration into the investigated system can reduce GWI by 20% and increase TAC by 1% compared to the reference case. Despite increased TAC from investments and heat pump (HP) operation, summer charging of the HT-ATES remains flexible and cost-effective. An idealized future scenario indicates that HT-ATES with a storage capacity of 16,990 MWh and HPs could cover most of the heating demand, reducing GWI by up to 91% while TAC increases by 6% relative to the reference system.
Hampel et al. (Thu,) studied this question.