Improving Data Center Energy Efficiency Through End-To-End

Cooling Modeling and Optimization

Sponsored by U.S. Department of Energy (DOE) (2016 – 2020)

Project Description

Led by Dr. Zuo, this is a joint research project with Lawrence Berkeley National Laboratory and Schneider Electric. We have developed an open-source, free software package (see the following links) which provides practical, end-to-end (from the IT equipment to heat rejection to the ambient) modeling and optimization for data center cooling. It can be used as a stand-alone tool by data center designers, service consultants and facility managers, or be integrated into existing data center management software for autonomous optimal operation. It is the first practical tool that couples the modeling of airflow-management and cooling systems to enable a global data center cooling optimization. Its self-learning regression model enabled by an in situ adaptive tabulation algorithm and a fast fluid dynamics model can predict the critical airflow information under various operational conditions within a few milliseconds. The equation-based modeling language allows a fast and flexible modeling of various cooling system configurations. We have demonstrated the usage of our tool and identified 53% energy saving potentials in a Florida data center and 74% in the Massachusetts data center.

Open Source Models for Data Center Cooling

Technical Advisory Group

Name Institution
Amistadi, Henry R. MITRE Corporation
Cleaver, Donald Keystone Critical Systems & Advisors
Doppelhammer, F. James University of Miami
Geraghty, Edward P. CEC Group, LLC
Groenewold, John JPMorgan Chase & Co.
Herrlin, Magnus Lawrence Berkeley National Laboratory
Kaiser, Raymond Amzur Technologies
Meneghan, Brian W. Carrier Corporation
Plamondon, David University of Massachusetts Medical School
Sartor, Dale Lawrence Berkeley National Laboratory
Sorell, Vali Syska Hennessy Group, Inc.

Collaborators

Press Release

Project Report

Ph.D. Theses

Journal Articles

Conference Papers

Presentation