Improving the cooling performance of a data center: impact of unit arrangement and CFD setups | ||||
Engineering Research Journal | ||||
Article 2, Volume 168, Issue 0, December 2020, Page 117-145 PDF (1.87 MB) | ||||
Document Type: Original Article | ||||
DOI: 10.21608/erj.2020.140126 | ||||
View on SCiNiTO | ||||
Authors | ||||
S. Nassar* ; A. Abdelsamie; A. Mohamed; M. Sedrak | ||||
Department of Mechanical Power Engineering, Faculty of Engineering at El-Mattaria, Helwan University, Masaken El-Helmia P.O., Cairo 11718, Egypt. | ||||
Abstract | ||||
Data center can be considered as the heart of any large institution; it uses for a lot of purposes such as data transfer, data storage, data analysis, etc. The performed operations release huge amount of heat, which should be cooled down. Therefore, a huge amount of energy is required for this cooling process. This work aims to improve the cooling efficiency and consequently, decrease the energy consumption, which is required for a specific data center. The improvements have been achieved by introducing two modifications: modifying the locations of heat sources (hereinafter will refer as racks units) and cooling machines (hereinafter will refer to CRACS units), and the airflow distribution path; where the suggested arrangement is different than the standard one, and changing the location of the server fan compared to the standard one. As a first step, the quality of the numerical simulation of the suggested modifications should be evaluated. This assessment was done by testing the impact of the mesh size on the air velocity and temperature as the main parameters in the current analysis. CFD Numerical investigation is done using ANSYS 16. The geometry is created using AUTOCAD 18, then it is exported to ANSYS. The meshing is created using ANSYS workbench with taken into account several mesh refinement. It was found that the suggested modifications increase the cooling performance and reduces energy consumption | ||||
Keywords | ||||
Data centers; Energy saving; Cooling performance | ||||
Statistics Article View: 263 PDF Download: 409 |
||||