The prediction of flow and temperature in data center operation is of particular importance in creating better capacity utilization and lower capital costs. However, it is often a time- and computing-intensive task, and would be well served by more expeditious modeling methods than full Computational Fluid Dynamic (CFD) thermofluidic models. A software package named COMPACT (Compact Model of Potential Flow and Convective Transport) was developed to provide one such alternative. Recent versions of COMPACT take under 10 seconds on a commercially available laptop to characterize a 550 square foot data center; the same room modeled with a CFD solver took 8 hours. Having the ability to create velocity and temperature predictions orders of magnitude faster than conventional CFD allow a variety of data center applications for the model, such as use as a first-order design tool, a potential improvement to plant-based controllers, a tool for system-wide assessment of life-cycle efficiency, and as an initial guess for complex CFD solvers. COMPACT applies convective energy transport equations to a computed potential flow field to approximate a flow and temperature field. The results from this model were compared to experimental measurements taken from a data center at Hewlett-Packard Laboratories in Palo Alto, CA. The presence of high localized temperatures in the model led to the conclusion that recirculation and buoyancy were contributing excessively to error in the model. A novel approach was proposed to account for these effects: a non-iterative (to preserve computational resources) method of vortex superposition, in which hot locations in the original model are analyzed and a corrective flow field consisting of Rankine vortices is superimposed on the solution. An updated model using this approach was tested with further experimental measurements taken from the data center at HP Labs, and identical inputs were used to compare COMPACT to commercially available CFD. These newer results showed a marked decrease in mean deviation of the model from measured temperatures, as well as elimination of the highly localized temperatures which afflicted the original COMPACT results. The vortex superposition model was also “tuned”, with vortex strength optimized for multiple test cases at varying levels of recirculation. In addition, the model-generated velocity and temperature fields were used to locate and quantify the destruction of exergy resulting from the mixing of warmer and cooler air in the room; these results are used as a measure of system efficiency.
[1]
Amip J. Shah,et al.
Exergy-Based Environmental Design of a Computer Room Air Conditioning Unit
,
2009
.
[2]
Jeffrey Rambo,et al.
Modeling of data center airflow and heat transfer: State of the art and future trends
,
2007,
Distributed and Parallel Databases.
[3]
C. Bash,et al.
Exergy Analysis of Data Center Thermal Management Systems
,
2008
.
[4]
Amip J. Shah,et al.
Lifetime exergy consumption as a sustainability metric for information technologies
,
2009,
2009 IEEE International Symposium on Sustainable Systems and Technology.
[5]
Van P. Carey,et al.
Experimental Validation of the COMPACT Code in Data Centers
,
2010
.
[6]
Cullen E. Bash,et al.
Optimization of Outside Air Cooling in Data Centers
,
2011
.
[7]
Cullen E. Bash,et al.
Computational Fluid Dynamics Modeling of High Compute Density Data Centers to Assure System Inlet Air Specifications
,
2001
.
[8]
Amip Shah,et al.
Optimizing data center cooling infrastructures using exergothermovolumes
,
2010,
2010 12th IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems.
[9]
Van P. Carey,et al.
Exploration of a Potential-Flow-Based Compact Model of Air-Flow Transport in Data Centers
,
2009
.