数据中心
网格
电
计算机科学
可再生能源
电力系统
发电
需求响应
智能电网
多样性(控制论)
经济调度
数据建模
环境经济学
功率(物理)
工程类
电气工程
数据库
物理
人工智能
经济
操作系统
量子力学
数学
几何学
作者
Yasmine Abdennadher,Julia Lindberg,Bernard C. Lesieutre,Line Roald
标识
DOI:10.1109/naps56150.2022.10012198
摘要
The demand for computing is projected to grow rapidly over the next decade. At the same time, technology companies have pledged to reduce carbon emissions associated with the electricity generation needed to meet this computing demand. To achieve this goal, data centers need to be sited at locations with access to low carbon electricity while continuing to meet transmission limits within the electric grid. In this paper, we present a model that determines the optimal placement of future data centers that accounts for the underlying power system dynamics. This model utilizes a bilevel program where the lowerlevel problem calculates the electricity dispatch of the entire electric power network by solving a DC OPF and the upperlevel problem identifies data center locations that minimize the carbon emissions of this dispatch. We test this model under a variety of operating conditions and show that to minimize carbon emissions from new data center loads, it is not enough to place new data centers near renewable generation sources. Instead one needs to consider grid operations and limits.
科研通智能强力驱动
Strongly Powered by AbleSci AI