Powering the AI Era: Resilient Converter-interfaced Systems with Critical Digital Infrastructure
【讲座题目】Powering the AI Era: Resilient Converter-interfaced Systems with Critical Digital Infrastructure
【时 间】2026年5月29日 下午：15:00-17:00
【地 点】北京校区 主A322学术报告厅
【主讲人】侯云鹤教授，香港大学电机和计算机工程系终身教授、系常务副主任、IEEE Fellow、美国弗吉尼亚理工大学电气与计算机工程系终身教授。
【主讲人简介】
Dr. Yunhe Hou is a Tenured Professor and Deputy Head with the Department of Electrical and Electronic Engineering at the University of Hong Kong. He is an IEEE Fellow. He received his B.E. and Ph.D. degrees in Electrical Engineering from Huazhong University of Science and Technology, China, in 1999 and 2005, respectively. From 2006 to 2009, he was a postdoctoral researcher at Tsinghua University, Iowa State University, and University College Dublin. In 2010, he was a Visiting Scientist with the Laboratory for Information and Decision Systems at MIT. Dr. Hou has supervised more than 30 Ph.D. graduates and has authored more than 170 papers, 6 monographs. He has been recognized as a Top 1% Scholar by Clarivate Analytics and as a career-long World's Top 2% Scientist in the Stanford/Elsevier ranking. He chairs an IEEE PES task force on power system restoration with renewable energy sources. He served as an Associate Editor of the IEEE Transactions on Smart Grid and the IEEE Transactions on Power Systems and several other leading journals. Dr. Hou was elevated to the IEEE Fellows for contributions to theory and decision support tools for power system restoration. Dr. Hou will join the Bradley Department of Electrical and Computer Engineering at Virginia Tech as a Tenured Professor in 2026 August.
【报告内容简介】
Abstract: Power systems are entering a new regime. Both supply and demand are becoming more converter-interfaced, fast-acting, and digitally coordinated. The challenge is not simply larger load growth. Some of the assumptions behind traditional planning, operation, and restoration are starting to weaken.
This talk focuses on AI data centers as a new class of critical digital infrastructure. Their power demand is shaped by computation, communication, power-electronic conversion, and control. A profile that appears acceptable under conventional system criteria may still create planning, operation, and restoration problems. In this talk, the gap between grid-feasible and service-feasible operation is examined. It also discusses how structured load behavior interacts with vulnerable system modes. A restoration method is further developed for large digital infrastructure with staged return and control-mediated behavior, rather than conventional passive load pickup. The talk closes with the broader view that critical digital infrastructure should be treated not simply as load to be served, but as part of the new challenges facing reliable and resilient power system operation.
