面向绿电消纳的交通-算力负荷灵活性可调潜力量化综述OA
Review of flexibility and adjustable potential quantification of transportation and computing loads for green power accommodation
高比例风电、光伏接入使绿电消纳由电量平衡转向源荷时空匹配与灵活调节问题,电动汽车、换电站、电动重卡补能站和数据中心等新型负荷凭借时空转移与功率调节能力,成为支撑绿电消纳的重要需求侧资源.梳理了3类交通负荷的灵活性来源、约束边界及量化方法:普通电动汽车基于荷电状态(SOC)安全边界和停车时间窗口形成概率性聚合能力,换电站依托电池库存和充换电解耦形成站端负荷转移空间,电动重卡在矿区、港口等场景提供集中式调节能力;三者均可从功率上下限、可转移电量和SOC状态边界刻画可调潜力,但须以车辆到离站时间、运输任务和站点容量等约束修正响应持续时间.算力侧从任务时间转移、空间迁移和设备侧功率调节3个层面分析了数据中心负荷灵活性,可延迟批处理任务、跨数据中心任务迁移和动态电压频率调节分别提供时间、空间和功率维度的调节空间,但均受服务等级协议、网络带宽和设备安全等约束限制.进一步归纳了协同量化思路:基于虚拟储能模型将交通补能时移与算力任务延迟统一为等效充放电行为并刻画时间灵活性边界;基于节点间可迁移负荷容量刻画空间灵活性;进而提出交通-算力交叉补偿模型,显式描述一类负荷受限时另一类负荷的补充调节关系,采用拉丁超立方抽样与K-medoids聚类生成代表场景,在给定置信水平下定义可信可执行功率、折减系数和响应可靠率,实现理论潜力的可信量化.研究表明,交通负荷适合提供短时功率响应和局部配网支撑,算力负荷在持续调节和跨区域迁移方面更具优势,二者协同可形成"短时快速响应-中长期持续调节-跨节点负荷迁移"的复合柔性资源,有效提升绿电出力与负荷需求的时空匹配度.理论可调潜力须经运行约束和不确定性修正后才能作为电网可调用的柔性资源.最后指出,现有研究存在统一建模不足、可信可调潜力量化不足、指标体系不完善和典型场景验证不充分等问题,未来应发展机理-数据融合的可信潜力在线计算技术与安全隐私约束下的分布式智能调控技术.
The high penetration of wind and photovoltaic power has shifted green power accommodation from an energy balance issue to the challenges of spatiotemporal source-load matching and flexible regulation.Emerging loads,including electric vehicles,battery swapping stations,energy replenishment stations for electric heavy-duty trucks,and data centers,have become important demand-side resources for supporting green power accommodation due to their spatiotemporal transfer and power regulation capabilities.The flexibility sources,constraint boundaries,and quantification methods for three types of transportation loads are reviewed.Conventional electric vehicles exhibit probabilistic aggregation capabilities based on safe state of charge(SOC)boundaries and parking time windows.Battery swapping stations provide station-level load transfer flexibility through battery inventories and decoupling of charging and swapping processes.Electric heavy-duty trucks provide centralized regulation capabilities in mines,ports,and other scenarios.The adjustable potential of all three load types can be characterized by upper and lower power limits,transferable energy,and SOC boundaries.However,the response duration should be corrected by considering constraints such as vehicle arrival and departure times,transport tasks,and station capacity.On the computing side,data center load flexibility is analyzed from three aspects:task time transfer,spatial migration,and equipment-side power regulation.Deferrable batch processing tasks,task migration across data centers,and dynamic voltage and frequency scaling provide regulation flexibility in the temporal,spatial,and power dimensions,respectively.However,all of these approaches are constrained by service-level agreements,network bandwidth,and equipment safety.The approaches for coordinated quantification are further summarized.A virtual energy storage model is used to unify the time transfer of transportation energy replenishment and the delay of computing tasks as equivalent charging and discharging behaviors,and to characterize temporal flexibility boundaries.Spatial flexibility is characterized based on transferable load capacity between nodes,and a cross-compensation model between transportation and computing loads is then proposed to explicitly describe the complementary regulation relationship between the two types of loads when one type is constrained.Representative scenarios are generated using Latin hypercube sampling and K-medoids clustering.Under a specified confidence level,credible executable power,reduction coefficients,and response reliability rate are defined to enable credible quantification of theoretical potential.Existing studies indicate that transportation loads are suitable for providing short-term power response and local distribution network support,whereas computing loads have advantages in sustained regulation and cross-regional migration.Their coordination can form a composite flexible resource that integrates short-term rapid response,medium-and long-term continuous regulation,and cross-node load migration,thereby effectively improving the spatiotemporal matching between green power output and load demand.Theoretical adjustable potential can serve as a flexible resource available for grid scheduling only after correction for operating constraints and uncertainties.Existing research faces problems including inadequate unified modeling,insufficient quantification of credible adjustable potential,incomplete indicator systems,and inadequate validation in typical scenarios.Future efforts should focus on developing online technologies for credible potential calculation based on mechanism-data fusion and distributed intelligent regulation technologies under security and privacy constraints.
朱虹谕;张蕊;武晓栋;陈拓宇;陈敏;TAN Minghui;张冬冬
内蒙古工业大学 新能源学院,内蒙古 鄂尔多斯 017010||内蒙古工业大学风能太阳能利用技术教育部重点实验室,呼和浩特 010051||内蒙古工业大学大规模储能技术教育部工程研究中心,呼和浩特 010051内蒙古工业大学 新能源学院,内蒙古 鄂尔多斯 017010内蒙古工业大学 新能源学院,内蒙古 鄂尔多斯 017010||内蒙古工业大学风能太阳能利用技术教育部重点实验室,呼和浩特 010051湖南大学 电气与信息工程学院,长沙 410082北京交通大学 电气工程学院,北京 100044泰莱大学 工程学院,吉隆坡 47500,马来西亚内蒙古工业大学 新能源学院,内蒙古 鄂尔多斯 017010||内蒙古工业大学风能太阳能利用技术教育部重点实验室,呼和浩特 010051||内蒙古工业大学大规模储能技术教育部工程研究中心,呼和浩特 010051||内蒙古工业大学内蒙古自治区新能源与储能技术重点实验室,呼和浩特 010051
能源科技
绿电消纳交通负荷算力负荷可调潜力虚拟储能时空协同需求响应算电协同
green power accommodationtransportation loadcomputing loadadjustable potentialvirtual energy storagespatiotemporal coordinationdemand responsecomputing-electricity coordination
《综合智慧能源》 2026 (9)
14-28,15
国家重点研发计划项目(2026YFE0211200)内蒙古自治区自然科学基金项目(2026QC0364,2026QC0321)国家自然科学基金项目(52677064) National Key R&D Program of China(2026YFE0211200)Inner Mongolia Natural Science Foundation(2026QC0364,2026QC0321)National Natural Science Foundation of China(52677064)
评论