Immersion two-phase liquid cooling for data center thermal management – energy & exergy analysis

• Main Author: Teo, Aik Peng
• Other Authors: Fei Duan
• Format: Final Year Project
• Language: English
• Published: 2018
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/75052
• Institution: Nanyang Technological University

The astounding growth in technology in recent years increases the volume of information to be stored in data centers. More organisations are relying on these data centers for their daily operations. This leads to running the data centers all year round which drastically increases the energy consumption to cool down the servers. The excessive heat generated by the servers will cause detrimental effect on the electronic components. The current prevalent method of cooling, the traditional air cooling consumes large amount of energy and is ineffective in removing the excess heat in a short period of time. Alternative cooling method is introduced to solve this problem. Direct two-phase immersion liquid cooling has drawn attention due to its high heat removal capability. In this report, two-phase liquid immersion cooling will be carried out by submerging the servers in a pool of Novec 7100 dielectric coolant. The liquid immersion cooling system has a design load of 32 kW. Energy and exergy analysis will be evaluated under three different loading conditions such as zero-load, half-load and full-load. Results from the experiments show that the Coefficient of Performance (COP) increases by 0.87 from zero-load to half-load, and 1.73 from half-load to full-load. The exergy efficiency at zero-load, half-load and full-load is 0.45, 0.71 and 0.68, respectively. Based on the three loading conditions, the pump has shown to have the largest exergy destruction ratio as compared to the heat exchanger and the dry tower. This indicates that the pump is the main component to focus on in order to improve the exergy efficiency of the system. Further studies can be done on using different coolants in the cooling system. The server tank can be further customized to be more compact which allows for more servers to be installed. Higher loading conditions can be evaluated in both energy and exergy analysis.