Development and experimental study of an independent row-based cooling system for improving thermal performance of a data center

Abstract

As the power density of IT equipment increases, the requirements for data center facilities are also becoming more rigorous. In addition to higher IT power density, cooling has become a key issue in data centers where cooling accounts for about 40% of total energy usage. An improved independent row-based data center cooling system has been suggested, and is found to be the better efficient in removing heat output of IT equipment based on the observations and experimental results from two cooling strategies. This study, that is practice-based learning in a legacy data center remodeling project, has evaluated the energy and thermal performance of two data center cooling approaches: an improved independent row-based cooling and a conventional room-based cooling. The air temperature, humidity and air distribution efficiency of an IT environment were examined by using field measurements and six performance indices. The differences and technical features of this experimental study is that two different forms of cooling systems were evaluated based on the same IT environment conditions; location of the IT equipment and the cooling load in same IT server room. The results show that data center cooling systems utilizing the “the fully containment of cold and hot aisles” technique are very effective at removing heat load of IT servers. The advantage of row-based cooling is that it is located close to hot spots and does not experience the heat losses generated in the air delivery path of the room typical of room-based cooling systems. SHI and RHI were improved by 37.1% and 20.0%, and RTI and β were improved by 73.2% and 44.3% respectively. According to the observations and index evaluation results, we suggest that a new independent row-based cooling strategy is more efficient for server cooling. Air management helps to reduce cooling energy by enhancing optimal operation and improving cooling system efficiency. Furthermore, metrics play a significant role in providing performance indices for air management systems.