You are in:Home/Publications/Effect of in-row cooling units numbers/locations on thermal and energy management of data centers servers

Ass. Lect. Abdullah Mohamed Gamal Abbas :: Publications:

Title:
Effect of in-row cooling units numbers/locations on thermal and energy management of data centers servers
Authors: Sameh A. Nada; Abdullah M. Abbas
Year: 2021
Keywords: data centers, economic study, experimental validated numerical simulation, in-row cooling, thermal performance
Journal: International Journal of Energy Research
Volume: Not Available
Issue: Not Available
Pages: Not Available
Publisher: Wiley
Local/International: International
Paper Link:
Full paper Not Available
Supplementary materials Not Available
Abstract:

The current research investigates the effects of numbers and arrangements of the cooling units on the thermal and energy management of in-row cooling architectures of data centers. Three different configurations with different numbers of cooling units; namely one rack between each two cooling units (Configuration A), two racks between each two cooling units (Configuration B), and three racks between each two cooling units (Configuration C), are numerically investigated. A scaled physical module of data center rack is experimentally constructed and tested for model validation. Two thermal evaluation indices (Supply Heat Index SHI and Index of Mixing IOM) and two energy efficiency metrics (Beta Index β and Energy Utilization Index ηr) are used to investigate, evaluate, and compare the thermal management and energy efficiency of the three configurations. The results show that the three configurations have good thermal homogeneity and the same good values of the thermal and energy efficiency metrics at the intermediate racks. However, problems of hot air circulations with hot spots and bad performances indices were noticed at terminal racks of the racks array. The energy utilization index is in the optimal range (about 2) for the intermediate rack and drop to 1.5 for the terminal racks. Economic study shows that configuration B has the minimum capital cost and configuration A has the minimum needed white area with a capital cost very close to that of B. Accordingly, configuration B is economically recommended, and configuration A is recommended for limited white area

Google ScholarAcdemia.eduResearch GateLinkedinFacebookTwitterGoogle PlusYoutubeWordpressInstagramMendeleyZoteroEvernoteORCIDScopus