Comparative study for waste heat recovery in immersion cooling data centers with district heating and organic Rankine cycle (ORC)

Immersion cooling has received widespread interest for the advantages of high heat dissipation density and excellent temperature uniformity. The outlet water temperature of the immersion cooling cabinet could reach up to 60 degree celsius, and thus the waste heat can be recovered. Waste heat recovery methods in data center with district heating and power generation based on organic Rankine cycle (ORC) were discussed. Besides, the cases coupled with water chiller (WC) and heat pump (HP) were investigated to increase the ORC efficiency. Local factors like ambient temperatures, electricity prices, carbon emission costs, and heating costs were considered for performance comparison in different cities, including Beijing (North China), Shanghai (East China), Shenzhen (South China), and Qingyang (Northwest China). The influence of operation temperatures on system energy-saving performance was studied comprehensively. The results show that by introducing the WC and HP subsystems, the ORC temperature difference between the evaporation and condensation increases, thus improving the ORC efficiency. However, the additional power consumption of the WC and HP subsystems is larger than the increase in ORC power output, thus deteriorating the system energy-saving performance. With the ORC subsystem, the power usage effectiveness (PUE) of the data center in Beijing, Qingyang, Shanghai, and Shenzhen could be decreased from 1.105 to 1.044, 1.042, 1.057, and 1.071, respectively. Comparing district heating with the ORCwaste heat recovery system (ORC-WHS), it is found that civil heating is suitable for Beijing and Qingyang because of the large heating demand and the low initial investment cost. The ORC-WHS is the most cost-effective in Shenzhen because of its relatively high electricity price and low heating demand. For Shanghai, civil heating is recommended if the heating transport distance is less than 10 km. Otherwise, ORC-WHS is recommended. In addition, R600a has the best comprehensive performance due to its relatively low price, low global warming potential, low density and favorable energy-saving performance. This paper may provide guidance for reducing energy consumption and cutting CO2 emissions toward green data centers under different local conditions.