基于BP神经网络的输电线路工程造价预测模型  

BP Neural Network Based Cost Prediction Model for Transmission Projects

作　　者：凌云鹏 阎鹏飞[2] 韩长占[2] 杨晨光[2] 

机构地区：[1]河北省电力公司电力经济技术研究院,河北石家庄050021 [2]河北省电力公司,河北石家庄050021

出　　处：《中国电力》2012年第10期95-99,共5页Electric Power

摘　　要：为解决在输电线路工程前期决策阶段的方案比选过程中,如何利用较少的工程信息,快速、准确地预测比较各方案工程造价,评估工程方案的合理性,本文提出一种基于BP神经网络的输电线路工程造价预测方法,分析并找出输电线路工程造价的影响因素,并将这些因素作为BP神经网络输入变量,输电线路工程造价作为输出变量,建立3层结构的BP神经网络预测模型,并用实际工程数据训练和验证预测模型。计算实例表明,该模型能够准确、快捷地预测输电线路工程造价,适合于工程方案比选过程中的各方案造价预测对比,评估工程方案的合理性。A BP neural network based method for predicting the cost of transmission projects is proposed to quickly and accurately compare the costs and assess the reasonableness of different plans with limited information during the early stage of the decision-making process.The factors influencing the project cost are also analyzed.Furthermore,by putting these factors into BP neural network as input variables and the project cost as an output variable,a BP neural network model of three-tier structure is constructed using actual project data for data training and prediction model verification.The example using this method shows it can forecast the project cost quickly and accurately,which is suitable for comparison of project cost and assessing the reasonableness of different plans.

关 键 词：输电线路 工程造价预测 BP神经网络 

分 类 号：TM752[电气工程—电力系统及自动化]