首页|期刊导航|综合智慧能源|绿电直连模式下风光储驱动的数据中心经济性分析

绿电直连模式下风光储驱动的数据中心经济性分析OA

Economic analysis of wind-photovoltaic-storage-driven data centers under green power direct connection mode

中文摘要英文摘要

随着"双碳"目标深入推进与数字经济迅猛发展,数据中心能耗持续攀升,绿电直连模式已成为其低碳转型的重要路径.以宁夏中卫地区为研究对象,基于美国国家航空航天局逐时气象数据与混合优化多能源电力模型软件(HOMER Pro)平台,建立风电与光伏出力时序模型,并构建由光伏发电、风力发电、储能电池、电网及单相浸没式液冷数据中心(IT额定功率为1 000 kW)组成的绿电直连系统综合模型.系统采用"自发自用、余电上网、储能调节"运行策略,配套建立电价、碳减排及财务评价模型.以动态投资回收期(DPP)最小化为优化目标,在绿电占比不低于30%、绿电自用率不低于80%、余电上网比例不超过20%的政策约束下,对风光储容量进行离散优化.在电能利用效率(PUE)为1.10的基准工况下,最优配置为光伏2 150 kW、风机4台(共1 000 kW)、储能2 520 kW·h,对应DPP为9.71 a,20 a累计净现值(NPV)为442.2万元.全年能量平衡分析显示,数据中心年用电总量为7 199 MW·h,自发自用绿电量为3 998 MW·h,绿电占比均值约为55.5%,余电上网率约为7.8%,均满足政策约束.冬、夏典型周逐时分析验证了风光出力的互补特性,储能系统有效平抑了出力波动.对光伏投资、风电投资、储能投资、分时电价整体水平、余电上网电价及碳交易价格6项参数开展±40%敏感性分析,影响程度由大到小排序为:分时电价整体水平、光伏单位投资、风电单位投资、碳交易价格、储能单位投资、余电上网电价.其中:分时电价为主导因素,下降40%将导致项目无法回收投资;光伏与风电投资合计占初始投资的86.7%;碳交易价格、储能投资及上网电价影响相对有限.PUE为1.10,1.30和1.60的3种工况对比表明,PUE越低项目经济性越优.PUE由1.10升至1.60时,20 a累计NPV下降约9.36%,DPP延长约6.18%,主因是峰值负荷提高导致接入电网容量增加,年容量电费大幅上升.

With the advancement of the"dual carbon"goals and the rapid development of the digital economy,the energy consumption of data centers continues to rise.The green power direct connection mode has become a vital pathway for data centers to achieve low-carbon transition.Taking Zhongwei in Ningxia as the research setting,time-series output models were established for wind and photovoltaic(PV)power generation based on NASA hourly meteorological data and the HOMER Pro platform.A comprehensive green power direct connection system model was constructed,consisting of PV power generation,wind power generation,battery energy storage,the public power grid,and a single-phase immersion liquid-cooled data center with an IT rated power of 1 000 kW.The system operated under the strategy of"self-generation for self-consumption,surplus electricity fed to the grid,and energy storage for regulation".Electricity price,carbon emission reduction,and financial evaluation models were established.With the minimization of the dynamic payback period(DPP)as the optimization objective,discrete optimization of wind,PV,and energy storage capacities was carried out subject to policy constraints that the proportion of green power should not be lower than 30%,the self-consumption rate of green power should not be less than 80%,and the proportion of surplus electricity fed to the grid should not exceed 20%.Under the baseline condition with a power usage effectiveness(PUE)of 1.10,the optimal configuration included 2 150 kW PV capacity,4 wind turbines(totaling 1 000 kW),and 2 520 kW·h energy storage capacity.The corresponding DPP was 9.71 years,and the 20-year cumulative net present value(NPV)reached 4.422 million yuan.Annual energy balance analysis showed that the total annual power consumption of the data center was 7.199 million kW·h,the self-consumed green power generated on-site was 3.998 million kW·h,the average green power proportion was approximately 55.5%,and the surplus electricity feed-in rate was around 7.8%,all of which complied with the policy constraints.Hourly analysis of typical weeks in winter and summer revealed the complementary operation characteristics of wind and PV power,and energy storage effectively mitigated output fluctuations.Sensitivity analysis with a variation range of±40%was conducted on six parameters:PV investment,wind power investment,energy storage investment,overall time-of-use electricity price level,surplus electricity feed-in tariff,and carbon trading price.The influencing factors ranked in descending order of influence were:overall time-of-use electricity price level,unit PV investment,unit wind power investment,carbon trading price,unit energy storage investment,and surplus electricity feed-in tariff.The time-of-use electricity price was the dominant influencing factor.A 40%reduction would result in the project being unable to recover its investment.PV and wind power investments together accounted for 86.7%of the initial investment.Carbon trading price,energy storage investment,and feed-in tariff had relatively limited impacts.Further comparison among three operating conditions with PUE values of 1.10,1.30,and 1.60 indicated that a lower PUE led to better project economic performance.When PUE increased from 1.10 to 1.60,the 20-year cumulative NPV decreased by approximately 9.36%,and the DPP increased by roughly 6.18%.This was mainly due to the substantial increase in annual capacity charges resulting from larger grid connection capacity required to accommodate elevated peak loads.These findings provide a quantitative reference for capacity planning and investment decision-making of data centers adopting the green power direct connection mode.

程云龙;刘芸彤;崔诵笈;王星皓;李雪强;张志强;刘圣春

新疆维吾尔自治区和田地区科学技术局,新疆维吾尔自治区 和田 848000||天津商业大学 天津市制冷技术重点实验室,天津 300134天津商业大学 天津市制冷技术重点实验室,天津 300134天津商业大学 天津市制冷技术重点实验室,天津 300134天津提尔科技有限公司 先进液冷技术及系统联合实验室,天津 300450天津商业大学 天津市制冷技术重点实验室,天津 300134||天津提尔科技有限公司 先进液冷技术及系统联合实验室,天津 300450天津商业大学 天津市制冷技术重点实验室,天津 300134天津商业大学 天津市制冷技术重点实验室,天津 300134

能源科技

算电协同绿电直连数据中心容量优化动态投资回收期经济敏感性分析电能利用效率

computing-electricity coordinationgreen power direct connectiondata centercapacity optimizationdynamic payback periodeconomic sensitivity analysispower usage effectiveness

《综合智慧能源》 2026 (9)

120-130,11

国家自然科学基金项目(52476085)河北省自然科学基金项目(E2026404005)天津市自然科学基金项目(25JCYBJC00030) National Natural Science Foundation of China(52476085)Natural Science Foundation of Hebei Province(E2026404005)Tianjin Natural Science Foundation Project(25JCYBJC00030)

10.3969/j.issn.2097-0706.2026.09.010

评论