Using Cold Underground Thermal Energy Storage to Manage Data Center Cooling Loads

The global expansion of data centers, driven by the increasing demand for data-intensive services such as artificial intelligence, leads to a surge in electricity consumption and facility cooling requirements. The electricity consumption for data center cooling systems is second only to the electricity consumption of IT equipment, placing significant electrical loads on the power grid and increasing grid connection requirements. Cold underground thermal energy storage (Cold-UTES) is one solution for managing data center cooling loads. These systems leverage the natural thermal capacity of the earth (e.g., reservoirs, borehole thermal energy storage) by chilling the subsurface during off-peak hours and seasons and using the resource to provide cooling during peak periods. Particularly, Cold-UTES systems can flexibly meet the steady cooling loads of data centers as needed, while reducing dependence on conventional, energy-or water-intensive cooling technologies. Cold-UTES has opportunities in scenarios where water usage must be minimized: Cold-UTES reduces water usage compared to evaporatively-cooled systems and reduces electricity consumption compared to dry air-cooled mechanical chillers. In this article, we describe data center cooling systems that include Cold-UTES. We develop technical models of these systems to illustrate the potential to manage cooling loads - and particularly to minimize energy consumption at times of peak grid stress (which often coincide with higher electricity prices). We describe a price-taker model that enables the optimal dispatch of the Cold-UTES cooling resource to minimize energy costs and quantify the potential cost reductions.