Experimental And Computational Study Of An IT Pod And Its Cooling System

Abstract

Conventional data centers are extremely large buildings that have complex power distribution and cooling systems. These data centers employ relatively expensive cooling systems which are inefficient. These inefficiencies of traditional data centers can be overcome by partitioning the server load into modular sections which can be deployed, powered and cooled depending on availability and requirement. Modular data centers are increasingly being developed instead of using large data centers to save huge amount of capital investment. High energy efficiency is a priority and measures have been taken to increase it. Cooling techniques such as air-side economization and evaporative cooling are proving to be some of those very effective measures.An IT pod cooled by an indirect and direct evaporative cooling unit is the prime focus of my research. Computational fluid dynamics analysis is performed to study the airflow pattern, temperature distribution and the effect of solar loading inside the IT pod and the Aztec cooling unit. Both steady state and transient analysis are performed during the study by modeling the IT pod and the cooling system using a commercially available tool, FloTHERM. The temperature distribution inside the IT pod is validated by obtaining the temperature measurements through thermocouples and sensors at different locations. The study also reveals location of the servers which get cooled efficiently than the others.The study was done in collaboration with an industrial partner and as such most of the results of this study have been continuously or being implemented.