Speaker

Caisheng Wang, Professor, Electrical and Computer Engineering, Wayne State University

Abstract

The rapid growth of artificial intelligence (AI) data centers is fundamentally transforming power system planning and operations. Global electricity consumption from AI data centers has already reached hundreds of terawatt-hours and is projected to grow substantially over the coming decade. Recent analyses suggest that data centers could account for as much as 17% of total U.S. electricity demand by 2030, while contributing on the order of 30% to 50% of net demand growth over the same period. This expansion poses new challenges for energy development and power generation, and the reliability, operation, and planning of the grid. Modern AI data centers also exhibit highly dynamic and complex load behavior, making both their internal power management and their coordination with the grid pressing problems to address.

This talk presents an integrated view of these challenges, by beginning with an overview of energy development and power generation, placing the electricity demand of AI data centers in the context of legacy and emerging generation resources. The discussion then turns to power supply for data centers, including the power delivery architectures and conversion stages that feed racks of graphics processing units (GPUs), where high current density, fast transients, and stringent efficiency requirements shape design choices. The talk also examines data center load profiles and characteristics, highlighting the need of data center load flexibikity quantification. It then addresses power and energy management at two levels: power management within the data center, covering the management of possible on-site sources; and coordination with the power grid, including demand flexibility, frequency and voltage support, and dispatch. The talk will be concluded by an introduction to the Electric, Connected and Autonomous Technologies for Mobility (eCAT) at Wayne, funded by the NSF IUCRC (Industry–University Cooperative Research Centers) program.

Bio

Dr. Caisheng Wang received the BSEE and MSEE degrees from Chongqing University, China, in 1994 and 1997, respectively, and the Ph.D. degree from Montana State University (MSU), Bozeman, in 2006, all in electrical engineering. In August 2006, he joined Wayne State University (WSU), Detroit, MI, where he is now a Professor in the Electrical and Computer Engineering Department. Supported by NSF, DOE, Great Lake Protection Foundation (GLPF), WSU, industry sponsors, and other sources with a total of over $19 million, he has been teaching and conducting research in renewable and alternative energy systems, power systems, power electronics, distributed generation, microgrids and smart grids, battery storage, electric vehicles, and machine learning applications in energy and transportation systems. He is the center director of the Electric, Connected and Autonomous Technologies for Mobility (eCAT) Center, https://www.ecat.center/, sponsored by the NSF IUCRC (Industry–University Cooperative Research Centers) program. He has authored/co-authored over 260 journal and conference papers, two books, and several book chapters.

He has served as an associate editor for several international journals and magazines, including IEEE Electrification Magazine, IEEE Transactions on Smart Grid, IEEE Power Engineering Letters, Electric Power Components and Systems, SAE Journal of Electronic & Electrical Systems for Passenger Cars, special issue guest editor for International Journal of Powertrains and IEEE Transactions on Transportation Electrification. He has won several awards, including the WSU President’s Award for Excellence in Teaching, Engineering Research Excellence Award, IEEE PES EDPG Prize Paper Award, an MSU Foundation Graduation Achievement Nomination Award, and an Honorary Citizenship, City of Bozeman, MT. Dr. Wang is a Fellow of IEEE and a member of WSU’s Academy of Scholars.

Virtual

If you are unable to join us in person, please join us on Zoom.