Published on Show Me Mizzou Dec. 19, 2024
As artificial intelligence (AI) technology expands, so does the need for data centers to handle its growing computational demands. The problem? These centers consume vast amounts of energy, especially for cooling their servers. To address this, Mizzou researcher Chanwoo Park is pioneering an advanced cooling system that could significantly reduce energy use while preparing data centers for the next wave of AI computing.
Data centers house servers running websites, apps, and cloud data. Currently, they use fans or liquid-based systems to manage heat. In 2022, U.S. data centers consumed more than 4 percent of the nation’s electricity, with 40 percent of that energy spent on cooling alone. To reduce this burden, Park and his team, supported by a $1.65 million U.S. Department of Energy grant, are creating a two-phase cooling system that uses phase change, such as boiling liquid into vapor, to efficiently dissipate heat.
“The liquid goes in different directions and evaporates on a thin metal surface,” Park recently told Show Me Mizzou. “Using this boiling surface, we’re able to achieve very efficient heat transfer with low thermal resistance.” The system is designed to connect seamlessly within server racks, operate passively without energy in low-demand situations, and consume negligible power even in active mode.
Park’s system aligns with the mission of the Center for Energy Innovation, a forthcoming campus facility where interdisciplinary teams will tackle energy efficiency challenges. “The center will allow us to explore additional ideas and innovations around energy-efficient processes,” Park said. He envisions his cooling system being deployed in the next decade, ready to support the future of AI computing with sustainable solutions.
To read more articles like this, become a Mizzou Alumni Association member and receive MIZZOU magazine in your mailbox. Click here to join.
