Project Title: “Data Center Load Characterization for Grid Analysis and Planning” 

 

Project Summary:

The rapid expansion of data centers, fueled by the growth of artificial intelligence and global data demands, has created significant challenges for power grids. These centers are high-power-consuming, fluctuating loads, especially GPU-intensive ones, which pose risks to voltage and frequency regulation in steady-state operation. During transient conditions, such as faults, data centers may experience sudden load shedding due to their interface converters detecting abnormal grid conditions, which can lead to imbalances and transient stability issues. To address these challenges, this project aims to develop comprehensive models of data center loads, including power electronics, cooling systems, GPU-intensive behaviors, etc. These models will be validated through lab testing and industry collaboration. Moreover, this project will also explore the ride-through margin of data center loads, investigating strategies to optimize their performance during grid disturbances and helping utilities define ride-through requirements. The findings will inform power flow calculations and transient stability assessments for grids with high data center penetration.

University Team Members:
Lead PI: Dr. Dingrui Li – Clemson University

Co-PI Members: 
Dr. Ramtin Hadidi – Clemson University
Dr. Badrul Chowdhury – UNC Charlotte