A Simulation-Based Approach to Data Center Thermal Efficiency Optimization

Abstract

The energy consumption for cooling electronic equipment in data centers using central systems is significant and will continue to rise. The motivation of the present research study is based on the need to determine optimization strategies to improve and optimize the thermal efficiency of data centers using a simulation-based approach. Here, simulation is used to model and optimize a proposed research data center for use as an environment to test equipment and investigate best practices and strategies such as containment and hybrid cooling. The optimization technique used in this study finds the optimal operating conditions and containment strategies of the data center while meeting specific thermal conformance criteria. More specifically, optimum supply airflow rate and temperature setpoint of cooling units are sought under different containment configurations, including both hot aisle and cold aisle containment strategies in both full and partial setups. The results of the computational fluid dynamics (CFD) simulations indicated a lower probability of hot spots with full hot aisle containment strategy in a data center operating at lower supply airflow rate and higher supply temperature setpoint. The optimization approach helped to determine a more efficient cooling system without the risk of under-provisioning. The study considered steady-state conditions with static heat load and fixed equipment layout. However, the generalized optimization process developed in the present study should add to the repertoire of tools presently used for the optimization of new air-cooled data centers.