Lawrence Berkeley National Laboratory

Electronic equipment in data centers generates heat during operation, which should be dissipated through a cooling system to prevent overheating and maintain optimal performance. Electricity consumption for the data center cooling system becomes significant as the demand for data-intensive services increases. Although various technologies have been developed and integrated into the data center cooling system, there are limited high-efficiency alternatives for data center cooling. In this study, we designed a reservoir thermal energy storage (RTES) system that stores cooling energy during winters and produces it during summers for data center cooling. We then demonstrated the techno-economic performance of the RTES incorporated with dry coolers and heat recovery for a year-round 5 MW cooling load. The RTES cooling production was reliable during the 20-year lifetime. We estimated the levelized cost of cooling as $5/MWh, significantly lower than $15/MWh for the base scenario where chillers and dry coolers supply the same cooling load without the RTES. We also estimated that the RTES-based cooling system annually avoids CO2 emissions up to 1488 tCO2e compared to the base case. These results highlight techno-economic feasibility and environmental benefits of the RTES and its potential to be deployed for various applications at large scales as well as for data center cooling.