The rapidly increasing electricity demand for data center operation has motivated efforts to betterunderstand current data center energy use and to identify strategies that reduce the environmentalimpact of these buildings. This paper builds on previous data center energy modeling efforts by char-acterizing local climate and mechanical equipment differences among data centers and then evaluatingtheir consequences for building energy use. Cities in the United States with significant data centeractivity are identified. Representative climate conditions for these cities are applied to data center energymodels for several different prototypical space types. Results indicate that widespread, effective econ-omizer use in data centers could reduce energy demand for data centers by about 20e25%, equivalent toan energy efficiency resource in the US of ∼13e17 billion kWh per year. Almost half of the potentialsavings would result from better airflow management and proper control sequences. The total energysavings potential of economizers, although substantial, is constrained by their limited potential for use inserver closets and server rooms, which together are estimated to account for about 30% of all data centerenergy demand. Incorporating economizer use into the mechanical systems of larger data centers wouldincrease the variation in energy efficiency among geographic regions, indicating that as data centerbuildings become more energy efficient, their locations will have an increasing effect on overall energydemand. Differences among regions become even more important when accounting for greenhouse-gasemissions. Future data center development could consider site location, along with efficiency measures,to limit the environmental impact attributable to this increasingly prominent economic sector. 34789
1. IntroductionIncreased reliance on the storage, transfer, and processing ofdigital information throughout all aspects of society has causedsignificant growth in data center energy use. These buildings houseinformation technology (IT) equipment, such as computer servers,as well as storage and network devices. In the United States, datacenter energy use doubled between 2000 and 2006 to about 60billion kWh annually and is expected to continue to rapidlyincrease [1e3]. While growth projections made before the currenteconomic downturn may overestimate near-term growth in datacenter activity, a recent evaluation showed that significant growthcontinued at least through 2008 with data centers consuming ∼70 billion kWh during that year [4]. That level of energy use corre-sponds to emissions of ∼1.2 1013gy 1of fossil carbon (42 Mt/y ofCO2), based on the average carbon intensity of 160 gC/kWh for USelectricity production [5].Non-IT components in data centersdheating ventilation andair-conditioning (HVAC) equipment, uninterruptible powersupplies (UPS), and building lightingdaccount for approximatelyhalf of data center electricity demand [6]. Total data center energydemand is often characterized with a simple metric, the powerusage effectiveness (PUE) [7], which is defined as the ratio of thetotal data center building load to the data center IT equipment load.Based on industry consensus regarding data center practices,previous estimates have applied a PUE of 2.0 to represent currentaverage data center energy efficiency, implying that IT and non-ITenergy use are equal [1,2,4]. However, IT equipment operates indata centers of significantly different types and sizes, which couldaffect actual PUE values. Masanet et al. [4] apportioned estimatesof national IT energy use among different data center space types based on International Data Corporation (IDC) data for 2005 USinstalled servers [1,8,9]. In that analysis, five space types were usedto categorize the locations of IT energy use: server closets, serverrooms, localized data centers,mid-tier data centers, and enterprise-class data centers. Table 1 shows that each space type representsa significant portion of total IT energy use in the United States.Along with differences in data center space type, the location ofan operating data center could affect the energy demand of non-ITequipment. 建筑与环境空调英文文献和中文翻译:http://www.751com.cn/fanyi/lunwen_32446.html