Development of modular air containment system: Thermal performance optimization of row-based cooling for high-density data centers

Modular air containment (MAC) prototype was developed for optimal row-based cooling system for high-density data centers. The main purpose of this study is to evaluate the thermal performance , compare the cooling provisioning of different configurations of row-based cooling system to find the optimum placement of in-row CRAC units. Techno-optimal estimations were performed considering different in-row CRAC placements and aisle layouts. A matrix combination was analyzed based on total 2592 cases that can be implemented. By the whole optimization process of in-row cooler placement, we can achieve the best solution with the statically balanced cooling provisioning with a margin of error of 0.2%. The cooled air from in-row CRAC units was supplied with almost no heat loss in the air distribution paths to each server. Numerical simulations confirmed that the cold-hot-cold (C-H-C) aisle layout can improve the minimum 9% of heat balance for the provisioned CRACs compared to the hot-cold-hot (H-C -H) aisle layout. Temperature distributions, air streamlines and net cooling usages showed that the C-H -C aisle layout has better thermal performance and cooling provisioning especially for the row-based cooling. The C-H-C aisle layout is more proper than the H-C-H aisle layout for achieving the desired cooling efficiency.& nbsp; (c) 2021 Elsevier Ltd. All rights reserved.