摘要
目前对电网覆冰灾害风险评估研究工作,缺乏对资料有效性评估手段,同时现有研究所考虑评价指标不够全面等缺陷。本文以贵州电网460条线路覆冰数据为例,从覆冰资料、微地形因子、气象条件出发,全面考虑与电网覆冰相关的13个评价指标,首先采用k-VNN算法对样本资料质量控制,剔除偏离较大样本;其次建立指标遴选判据,剔除对本文研究影响较小指标,并对可变指标气象因子,采用过程信息处理;最后采用LS-SVM算法建立电网覆冰灾害评估模型。研究表明:5.87%样本偏离程度较大,属于无用样本;山脊、坡向两个指标对电网覆冰影响较小;本文所建电网覆冰风险评估模型与实际观测值资料较为接近,平均相对误差仅为2.13%,质量控制手段能有效提升模型精度12.7%,气象过程信息处理能有效提升计算模型精度4.5%。本文最后以110kV赫韭线覆冰情况为例,与人工观冰结果比对,进一步验证本文所建模型的准确性。本文研究所得结论,能够有效减小冰期人工观冰工作量,指导电网防冰加固等具有重要意义。
At present,the research on the risk assessment of power grid icing disaster lacks the means to evaluate the effectiveness of the data,and the existing research is not comprehensive enough.In this paper,taking 460 lines icing data of Guizhou power grid as an example,starting from icing data,micro terrain factors and meteorological conditions,13 evaluation indexes related to power grid icing are comprehensively considered.Firstly,k-VNN algorithm is used to control the quality of sample data and eliminate large deviation samples.Secondly,the criterion of index selection is established to eliminate the indexes that have little influence on the research of this paper,and to eliminate the meteorological factors of the variable indexes,the process information is processed and LS-SVM algorithm is used to establish the model of icing disaster assessment.The results show that:5.87%of the samples deviate greatly and are useless samples;the ridge and slope direction have little influence on the icing risk of power grid.The average relative error is only 2.13%,the precision of the model can be improved 12.7%by means of quality control and 4.5%by means of meteorological process information processing.Finally,taking the icing condition of 110 kV hejiu line as an example,the accuracy of the model is further verified by comparing with the artificial ice observation results.The conclusions of this paper can effectively reduce the workload of ice observation during the Ice Age and guide the anti-ice reinforcement of the power grid.
作者
吴建蓉
文屹
杨涛
吕黔苏
肖书舟
黄军凯
WU Jianrong;WEN Yi;YANG Tao;LYU Qiansu;XIAO Shuzhou;HUANG Junkai(Electric Power Research Institute of Guizhou Power Grid Co.,Ltd.,Guiyang 550000 Guizhou,China)
出处
《电力大数据》
2021年第11期48-54,共7页
Power Systems and Big Data