【作者】 陈畅；

【导师】 谢吉华； 龙寰；

【作者基本信息】 东南大学 ， 电气工程， 2021， 硕士

【摘要】 随着工业化、城镇化进程加快和消费结构持续升级,我国能源需求特别是电能需求呈现刚性增长趋势。近年来,国家深入推进节能减排工作的开展,促进电力体制改革的不断发展,高度重视电力系统的线损管理工作,致力于打造节约型的现代化绿色电力。目前,配电网线损管理工作一直是供电企业加强营业管理的重点。配电网具有地域分布广、电网规模大、设备种类多、网络连接多样、运行方式多变等鲜明特点,位于电力系统的末端,直接与电力用户相连接,电能损耗一直较高,在整个电网线损中占有相当大的比例。因此,如何对配电网中的异常线损进行准确分析以及有效降低配电网线损成为供电企业亟待解决的重要难题。为了加强配电网线损的治理工作,明确配电网降损的重点方向,制定行之有效的有针对性的降损措施,本文基于天津地区的配电网实际参数针对配电网的线损异常识别以及多措施组合降损优化方面进行了研究,主要研究内容如下:(1)针对现有配电网线损异常识别的研究缺乏对线损的异常类型进行区分的问题,提出基于数据驱动的配电网线损异常类型识别方法,利用K-means、谱聚类、高斯混合模型以及支持向量机算法,学习异常技术线损和异常管理线损的不同波动趋势与模式,判断配电网线损是否存在异常情况并诊断线损的异常类型。(2)针对配电网技术线损异常馈线,提出配电网技术线损异常识别方法。首先通过拓扑匹配,确定最符合实际量测数据的最佳匹配拓扑结构;其次,提出基于老化特性的参数估计算法,并结合特殊牛顿迭代算法,确定考虑老化后的支路参数实际值;最终,通过支路老化评估定位老化严重的配电网支路,为之后配电网馈线的降损改造提供指导方向。(3)针对配电网管理线损异常馈线,提出配电网管理线损异常识别方法,基于馈线正常运行状态下的线损历史数据建立稳态控制图,通过稳态控制图的上下限范围检测管理线损异常馈线的异常波动时间,进一步利用风险评估技术定位相应异常时间中的管理线损具体异常位置,并评估其异常程度的风险等级。(4)为了充分挖掘配电网的降损潜力,提出配电网多措施组合降损优化框架,基于技术线损异常识别结果,通过潮流计算分析配电网馈线的线损薄弱点;从更换配电线路、配变以及无功补偿三个方面生成相应的降损措施;建立考虑降损综合效益的配电网多措施组合降损优化模型;为了解决现有研究多数采用枚举法求解上述模型导致的求解效率较低的问题,提出基于成本效益比的配电网多措施组合降损优化模型的求解方法。更多还原

【Abstract】 With the acceleration of industrialization and urbanization and the continuous upgrading of consumption structure,the energy demand,especially the electricity demand,has shown a growth trend.In recent years,the country has promoted the development of emission reduction work,promoted the continuous development of power system reform,paid great attention to the power loss management of the power system,and is committed to creating an economical modern green power society.At present,power loss management in the distribution network has always been the focus of power supply enterprises to strengthen business management.The distribution network has distinctive features such as wide geographical distribution,large grid-scale,multiple types of equipment,diverse network connections,and changeable operation modes.It is located at the end of the power system and is directly connected to power consumers.The power loss in the distribution network has always been high and occupies a considerable proportion in the whole power system.Therefore,how to accurately analyze the abnormal power loss in the distribution network and effectively reduce the power loss in the distribution network has become a crucial problem to be solved urgently.In order to strengthen the power loss management in the distribution network,clarify the critical direction of power loss reduction in the distribution network,and formulate effective and targeted power loss reduction strategies,the abnormal power loss detection and combined power loss reduction strategy optimization in the distribution network are studied based on the actual parameters of the distribution network in Tianjin.The main research contents are as follows:(1)The abnormal types of power loss are generally rarely considered in the research on the abnormal power loss detection in the distribution network at present.To solve this problem,a data-driven algorithm for identifying abnormal types of power loss in the distribution network is proposed to learn different fluctuation trends and patterns of abnormal technical loss and abnormal non-technical loss,utilizing K-means,spectral clustering,gaussian mixed model,and support vector machine algorithm.(2)For the feeders detected as abnormal technical loss in the distribution network,an algorithm for detecting abnormal technical loss is proposed based on topology verification and parameter estimation in this paper.First,in topology verification,the best-matched topology is determined among the alternative topologies,which best meets the actual measurement data.Then,a novel parameter estimation approach based on the aging characteristics of power equipment is proposed,combined with specialized NewtonRaphson iteration to estimate the actual branch parameters considering aging.Finally,the branch aging assessment is employed to locate the severely aging branch of the distribution network,which provides guidance for the subsequent power loss reduction.(3)For the feeders detected as abnormal non-technical loss,an algorithm for detecting abnormal nontechnical loss is proposed based on the control chart and risk assessment technology in this paper.The control chart is employed to monitor the fluctuation of the power loss and discover the abnormal time.Based on the detected abnormal time,the abnormal position is further located through the risk assessment technology.(4)In order to improve the power loss reduction effect in the distribution network,a framework of combined power loss reduction strategy optimization is proposed in this paper.Based on the detection results of the abnormal technical loss,the weak point of the distribution network is analyzed based on power flow calculation.Corresponding power loss reduction strategies are generated considering three aspects of replacing distribution lines,distribution transformers,and reactive compensation.A combined power loss reduction strategy optimization model considering the comprehensive benefits of power loss reduction is established.In order to solve the problem that the enumeration methods are generally used in most of the existing research to solve the above model,which causes low efficiency of power loss reduction,a method for solving the optimization model based on the cost-benefit ratio is proposed.更多还原