A Numerical Study for Contained Cold Aisle Data Center Using CRAC and Server Calibrated Fan Curves

In the present work, we demonstrate the effect of Cold Aisle Containment Systems (CACS) on the airflow and temperature distribution inside a representative data center. Computational Fluid Dynamics (CFD) is used to conduct this analysis. This study includes calibrated fan curves in the Computer Room Air Conditioner (CRAC) and the servers in order to capture the impact of pressure changes on the flow field. The system characteristics curve for the open (uncontained) system and contained system including the leakage effect is established. Since the IT-equipment has a crucial effect on the performance of contained systems, the individual and combined effect of fully enclosing the cold aisle and IT-equipment are investigated. Partially contained systems including doors only and ceilings only configurations are also considered in this study. Steady state and dynamic scenarios are simulated to characterize different containment systems.It is found that the pressure inside the fully contained system is determined by the IT-equipment as well as the geometrical obstructions of fully containing the cold aisle. Increasing the pressure due to enclosing the cold aisle is reduced by introducing IT-equipment, load banks in this case, which leads to a reduction of the total static pressure and an increase of the flow rate. It is also found that the fully contained system presents the best configurations in achieving low and uniform temperature distribution; however, partially contained systems can be a good solution if a certain cold air provision is maintained during operation. The dynamic analysis shows a significant increase in the safe time at full CRAC failure scenario compared with uncontained system due to utilizing the trapped cold air inside the plenum by the load banks fans.© 2013 ASME