Numerical simulation on a thermal management system for a small data center

A system with optimum thermal control and noise control is proposed to allow placement of servers at workspace for low operation office cost. A computational fluid dynamics (CFD) model is developed to predict the air flow and temperature distribution of the system. To improve the accuracy of the simulation, the geometry of the axial-flow fan is reproduced in detail and the rotor region is treated with the multiple reference frame model. A porous model is applied to describe the hydraulic and thermal behavior of the evaporator. The integrated system simulation model is validated by measurements. Four layouts of fans are proposed to analyze the airflow organization and temperature distribution. The inhomogeneity of the temperature field is quantitatively evaluated via information entropy and variance. The results reveal that the airflow organization scheme avoids the occurrence of an evident stagnant zone or a hot spot. There are low-velocity zones with insufficient heat transfer in the evaporator. Increasing the number of fans can improve the heat transfer and result in a more uniform temperature distribution. Setting up ventilation fans to increase the ventilation rate of the server is an effective way to meet the challenges of higher heat density. (C) 2018 Elsevier Ltd. All rights reserved.