Data center cooling systems consume a significant amount of electricity. The temperature of heat source (hot indoor air) and cold source (outdoor air) is usually ungraded, which results in wasted heat transfer temperature difference. A novel multi-stage cooling system is proposed to match different temperature grades of heat sources with corresponding cold sources. The heat transfer process and energy saving principles of the cooling system are illustrated using T-Q diagrams. A theoretical model is established based on the system heat transfer characteristics. The mismatch coefficient δ and entransy dissipation ΔJ of multi-stage cooling system both reduced by 6.7% compared to single-stage system, and the total temperature difference at natural cooling mode is reduced by 1.3 °C. The multi-stage cooling system offers a new approach to enable greater natural cold sources utilizations and lower power consumptions.
Keywords Data center, Multi-stage cooling, Structure optimization, Temperature grade, Energy saving