PUE values calculated in thisstudy account for roughly half of these savings and may be attrib-utable to different assumptions in airflow and control systemperformance. Total energy savings potential is constrained becauseeconomizers have limited utility for use in server closets and serverrooms,which together are estimated to account for about 30% of allIT energy demand.The modeled PUE values improve, most significantly withmid-sized data centers, when the return air temperature set-points are increased in the Economizer Plus scenario. This changereduces cooling demand by allowing the economizers to operatefor more hours of the year. In addition, increasing the temper-ature difference between the supply and return air allows forreduced airflow rates, which would limit the increase in fanenergy associated with economizer use. This finding illustratesthe value of operating data centers at the highest temperaturethat can be tolerated without compromising equipment reli-ability. Future studies of data centers might fruitfully exploredifferent temperature setpoints to optimize the balance betweenchiller and fan energy. The energy use reductions calculated forthe economizer designs highlight the increased contribution ofthe data center components for which no efficiency measureswere assumed, such as UPS and lighting. Efficiency improve-ments to such components could further improve overall datacenter energy efficiency not only by reducing the energy useddirectly to provide these services but also by reducing theadditional energy required to satisfy the cooling load associatedwith thermally inefficient device operation. The same amplifi-cation applies to efficiency improvements in the IT equipmentitself.Overall, PUE results show that larger data centers are typicallymore efficient than smaller data centers, and this disparityincreases in the Economizer and Economizer Plus scenarios. Conse-quently, the trend toward increased data center consolidation [29]portends potential efficiency improvements in overall data centerenergy use. Incorporating economizer use in larger data centersalso increases the differences in energy efficiency among climateregions. As data center services increase and if the buildings thathouse these services become more energy efficient, then the loca-tions of data centers will have a greater effect on their overallenergy demand. Regional differences become evenmore importantin managing greenhouse-gas emissions. The overall energydemand and environmental impact of future data center develop-ment can be strongly influenced by the use of the building energyefficiency measures analyzed here along with the climate andenergy supply characteristics of the site location.AcknowledgmentsWe thank Kim Traber for his help and insight in developingthe data center case studies and John Bruschi for his assistancewith the energy modeling. This work was partially performed atLBNL under the US Department of Energy Contract No. DE-AC02-05CH11231.Appendix. Supplementary materialSupplementary material associated with this article can befound in the online version, at doi:10.1016/j.buildenv.2010.10.023.References[1] Brown R,Masanet E, Nordman B, Tschudi B, Shehabi A, Stanley J, et al. Report tocongress on server and data center energy efficiency: Public law 109-431.Berkeley, CA: Lawrence BerkeleyNational Laboratory; 2007. Report LBNL-363E.[2] Koomey JG. Estimating total power consumption by servers in the U.S. and theworld. Oakland, CA: Analytics Press, http://sites.amd.com/us/Documents/svrpwrusecompletefinal.pdf; 2007.[3] Koomey JG. Worldwide electricity used in data centers. EnvironmentalResearch Letters 2008;3. Art. 034008.[4] Masanet ER, Brown RE, Shehabi A, Koomey JG, Nordman B. The energy effi-ciency potential of U.S. data centers. Proceedings of the Institute of Electricaland Electronics Engineers, submitted for publication.[5] EIA. Emissions from energy consumption at conventional power plants andcombined-heat-and-power plants. Energy Information Administration, Inde-pendent Statistics and Analysis, http://www.eia.doe.gov/cneaf/electricity/epa/epat3p9.html; 2010.[6] Tschudi W, Xu T, Sartor D, Nordman B, Koomey J, Sezgen O. Energy efficientdata centers. Berkeley, CA: Lawrence Berkeley National Laboratory; 2004.Report LBNL-54163.[7] Stanley JR, Brill KG, Koomey J. Four metrics define data center “greenness”enabling users to quantify energy efficiency for profit initiatives. Santa Fe, NM:The Uptime Institute; 2007.[8] Bailey M, Eastwood M, Grieser T, Borovick L, Turner V, Gray RC. Special study:data center of the future. New York, NY: IDC; 2007. Report #06C4799.[9] IDC. IDC’s worldwide installed base forecast, 2007e2010. Framingham, MA:International 建筑与环境空调英文文献和中文翻译(7):http://www.youerw.com/fanyi/lunwen_32446.html