Fast CFD for Data Center Cooling Design
. by Jim VanGilder, P.E., Director of Cooling Prediction R&D for APC by Schneider,
The power and flexibility of traditional CFD tools are not always justified for data center modeling given their limitations (cost, long run times, and training requirements). For example, in the design and planning stage, even the IT Equipment selections are not yet finalized and guesses must be made as to total rack power and airflow. In such applications, a lighter-duty version of CFD, based on potential flow modeling, may offer significant advantages. Several researchers have developed prototypes and, at least one, vendor offers potential-flow-based software commercially. Most recently, researchers have proposed models for incorporating physical effects like buoyancy into potential flow models thereby extending their useful scope while still retaining the speed and stability that make such models attractive. This presentation will review potential flow modeling technology and its role in data center design.