Two-phase liquid-immersion data center cooling system : experimental performance and thermoeconomic analysis

• Main Authors: Kanbur, Baris Burak, Wu, Chenlong, Fan, Simiao, Tong, Wei, Duan, Fei
• Other Authors: School of Mechanical and Aerospace Engineering
• Format: Article
• Language: English
• Published: 2021
• Subjects: Engineering::Mechanical engineering; Data Center Cooling; Thermoeconomic Analysis
• Online Access: https://hdl.handle.net/10356/154624
• Institution: Nanyang Technological University

The liquid-cooling data center (DC) systems have been becoming important for the rapidly developing high-performance processors since the traditional air-cooled DC systems cannot efficiently manage them due to high heat dissipation rates. Two-phase liquid-immersion cooling is one of the promising direct liquid-cooled DC systems, but its system-scale thermal management performance has not been investigated in detail yet. This study performs the system-scale thermal management of a two-phase liquid-immersion cooling DC system under six different real-time and dynamic operation loads in the range of 3.43-9.17 kW to see the thermodynamic and thermoeconomic performances. The system includes the DC server tank, circulation pump, and dry tower. Results show that the best coefficient of performance (COP) and power usage effectiveness (PUE) values are seen at the highest operation load with 6.67 and 1.15 while the minimum COP and the highest PUE are seen at the lowest operation load with 2.5 and 1.4, respectively. In the component-based assessments, the dry tower is found as the most dominant component according to the exergy destruction ratio analysis, energy costing, and carbon-related costs. The exergy efficiency varies between 8.0% and 18.9% for different operation loads, and the operating temperatures have a crucial impact on the exergetic performances. The thermoeconomic analysis deduces that the levelized product cost is roundly 1.14 S$ · h-1 which means the exergy-related terms (e.g. destruction and loss) increase the costing trends 3.25 times as high compared to the traditional energy-based economic calculations.