Numerical investigation of airflow and heat transfer in sun (orcale) hpc data centre with designs of exhaust fan installed on ceiling duct and overhead cooling system

Over the years, the power consumption and cooling demand in data centre have drastically increased. As a result, energy cost is increased, which requires the data centre to be properly designed to maintain the temperature within the allowable range in order to achieve high cooling efficiency. This p...

Full description

Bibliographic Details
Main Author: Shi, QingYun.
Other Authors: Li Hua
Format: Final Year Project (FYP)
Language:English
Published: 2012
Subjects:
Online Access:http://hdl.handle.net/10356/47624
Description
Summary:Over the years, the power consumption and cooling demand in data centre have drastically increased. As a result, energy cost is increased, which requires the data centre to be properly designed to maintain the temperature within the allowable range in order to achieve high cooling efficiency. This project is undertaken with the aim to optimize the cooling efficiency of data centre. The commercial CFD tools (Gambit and Fluent) are used to simulate the airflow and heat transfer within the data centre. The benchmark study is conducted with experimental validation, in order to evaluate the airflow and thermal distribution within SUN (Oracle) HPC data centre. The conditions used in the simulations are collected by previously experimental study of SUN HPC data centre. The important issues are identified, such as the recirculation of hot air, numbers of undesirable hot spots, insufficient cool air supplied in cold aisle and over-cooling at certain part of the data centre. By the simulation case studies, the better understanding of cooling system of the data centre is achieved. After that, the two designs are proposed. One of them focuses on the exhaust fan installed on the ceiling duct, the other on the overhead cooling system in the SUN (Oracle) HPC data centre, in order to improve the cooling efficiency and hence achieve higher energy efficiency.