特高压线路工程的工程量组合预测研究  

Combination Forecasting Methods of Ultra High Voltage Transmission Line Project Quantities

作　　者：罗福多 温卫宁 文凯 钟珍[3] 柳瑞禹[3] LUO Fuduo;WEN Weining;WEN Kai;ZHONG Zhen;LIU Ruiyu(Fujian Electric Power Survey and Design Institute,FuzhouFujian 350003,China;State Grid Beijing Institute of Technology and Economy ,Beijing 102209,China;School of Economics and Management, Wuhan University ,Wuhan Hubei 430072,China)

机构地区：[1]福建省电力勘测设计,福建福州350003 [2]国网北京技术经济研究院,北京102209 [3]武汉大学经济与管理学院,湖北武汉430072

出　　处：《湖北电力》2017年第2期1-7,12,共8页Hubei Electric Power

基　　金：国家电网公司软科学研究项目(35-CH7-2015-29)

摘　　要：在分析特高压(ultra high voltage,UHV)线路工程的工程量影响因素基础上,根据已有特高压线路工程相关数据特点,提出支持向量机、BP神经网络以及工程相似度三种工程量预测方法,针对单一预测方法的局限性,为进一步提高预测精度,构建基于偏差平方和最小的组合预测模型,组合预测模型可以多角度搜集数据信息,实现各种预测模型之间的取长补短。通过样本测试表明该组合预测模型明显降低了最大误差,缩小了误差波动范围。同时考虑到不可量化因素对特高压线路工程量的影响,利用数理统计中置信区间的估计得到工程量的区间预测值,为特高压线路工程量管控提供技术支撑。Based on the analysis of influencing factors to UHV Transmission Line Project Quanti?ties,according to the data characteristics of existing UHV transmission line projects,three differ?ent prediction methods are selected:SVM、BP neural network and engineering similarity.Consider?ing the limitation of the single prediction method,in order to improve the prediction accuracy,acombinationforecasting model based on the theory that the sum of deviation squares were minimumis proposed.The model can collect data information from various angles,and realize the comple?ment of each prediction model.The test results show that the combined forecasting model can signifi?cantly reduce the maximum error and the error fluctuation range.At the same time,considering theeffect of the non quantifiable factors on the engineering quantity of UHV transmission line,the in?terval value of the engineering quantity is estimated by using the confidence interval of mathemati?cal statistics.This method can provide technical support for the management and control of extrahigh voltage line projects.

关 键 词：线路工程 相似度 支持向量机 BP神经网络 组合预测 

分 类 号：P258[天文地球—测绘科学与技术]