Dimensionless Parameters for Evaluation of Thermal Design and Performance of Large-scale Data Centers

Large-scale data centers (~20,000m) will be the major energy consumers of the next generation. The trend towards deployment of computer systems in large numbers, in very dense configurations in racks in a data center, has resulted in very high power densities at room level. Due to high heat loads (~3MWs) in an interconnected environment, data center design based on simple energy balance with zones, is inadequate. Energy consumption of data centers can be severely increased by inadequate air handling systems and rack layouts that allow the hot and cold air streams to mix. In this paper, for the first time, we formulate nondimensional parameters to evaluate the thermal design and performance of large-scale data centers. The parameters, based on temperature and flow data, reflect the convective heat transfer and fluid flow inside the data center. These parameters have been formulated as indices that are scalable from rack level to data center level. To provide a proof of concept, computational fluid dynamic models of data centers are used to validate and demonstrate these indices. A first level design of experiment study is carried out to understand the effect of geometry and data center workload on the parameters. Different data center configurations are also investigated to understand the effectiveness of these parameters in specific cases. These parameters will not only provide an invaluable tool to understand convective heat transfer in large data centers but also suggest means to improve energy efficiency in data centers.