Grid Impact of Reservoir Thermal Energy Storage for Data Center Cooling

Abstract

Data centers are projected to account for up to 12% of total U.S. electricity demand by 2028 according to Lawrence Berkeley National Laboratory, up from 4.4% in 2023. Expanding electricity system infrastructure to reliably serve this rapid data center demand growth is one of the largest near-term challenges that the U.S. electricity system faces. In addition to increasing power demand, rapid data center growth could drastically increase water demand for cooling if evaporative cooling is used. If dry cooling and air-cooled chillers are used to reduce data center water consumption, cooling would drive peak data center demand, which in turn drives the electricity system investments needed to serve that demand. Cold reservoir thermal energy storage (RTES), a form of cold underground thermal energy storage (UTES), can be used to shift data center cooling loads to off-peak hours, reducing electricity system costs while also reducing data center water consumption. Using advanced, open-source power system modeling tools GenX and ReEDS, this paper investigates the impact of RTES for data center cooling on the bulk power system. We estimate the electricity system investment and operational cost savings that RTES could provide and evaluate the impact of RTES system duration on these savings. We find that seasonal RTES could reduce the grid cost of adding data center capacity by 3% to 6% in Virginia. These savings are primarily achieved by reducing cooling demand during net peak load hours, which decreases the firm generation capacity required to meet the planning reserve margin with additional data center load. This firm capacity reduction could be achieved with durations as short as 12 hours in Virginia. These findings suggest that RTES for data center cooling could appreciably lower the power system costs of serving rapidly growing data center load.