Significance Level of Factors for Different Airflow Management Configurations of Data Centers

Abstract

Data centers are among the highest energy consuming facilities and are projected to continue to increase in their power consumption for the foreseeable future. Due to the increase of computing power and the decrease in available floor space, maintaining the reliability of the electronic equipment in data centers is a big thermal challenge and, cannot be achieved solely by increasing the cooling capacity of the room. The overall thermal performance of data centers is highly dependent upon the thermal architecture of the facility. This paper presents numerical results of a parametric study, carried out for seven, fairly common, candidate configurations available for the air ducting design for data centers. Among the many factors associated with the data center thermal performance, three main factors at different levels have been selected to characterize their effect. The factors studied are ceiling height, tile flow rate and the location of the return vents. The numerical modeling is performed using a commercially available computational fluid dynamics (CFD) code based on the finite volume approach. This study also includes a summary of the statistical analysis carried out on the data obtained from the numerical parametric analysis, to determine the significance level of each of the individual factors and their interactions, on the thermal performance of the data center. The approach used here is to take an Analysis of Variance (ANOVA) approach, as a tool for determining the significance level of the different variables that affect the overall data center thermal performance. The tile flow rate is found to have significant effect on the thermal performance of all data center configurations studied.