Font Size: a A A

Data center power consumption reductions achievable with liquid cooling: A case study

Posted on:2010-11-22Degree:M.SType:Thesis
University:Tufts UniversityCandidate:Hwang, Douglas CFull Text:PDF
GTID:2448390002471111Subject:Engineering
Abstract/Summary:
Data centers consume approximately 2% of the electricity produced in the United States and this consumption continues to increase. This presents an increasing environmental challenge, largely due to the power consumption of the cooling systems needed for their thermal management. A model of data center energy utilization for thermal management based on thermo-fluid first principles was developed. The model is applied to a single rack system that can be used as an experimental platform for its validation. Two models are described, i.e., air-cooling, representing traditional data center thermal management, and a hybrid model which utilizes liquid cooling for the microprocessors and air-cooling for all other components. The air-cooling model considers a single chiller with a variable speed compressor and an off-the-shelf air-to-liquid heat exchanger. Water is used as an intermediate fluid between the refrigerant in the chiller and the air. The hybrid model replaces air cooled heat sinks on the microprocessors with cold plates and includes an additional chiller. Results were compiled to quantify the power consumption of the two cooling schemes over ranges of key variables, such as ambient temperature. For example, the air-cooled model showed a 10% reduction in energy consumption in accessory component cooling (i.e., all hardware aside from the servers) when the chiller setpoint was raised by 5°C, and the water flow rate pumped through the air-to-liquid heat exchanger was increased to accommodate this. In the hybrid model, the results showed a thermal management energy consumption decrease of more than 40% over the parametric ranges investigated relative to traditional air-cooling, illustrating the potential benefit of liquid cooling. The models can be utilized for future cooling system designs and trade-off studies.
Keywords/Search Tags:Cooling, Data center, Consumption, Model, Thermal management
Related items