基于BP神经网络的配电网工程造价预测模型建立与应用

Establishment and application of cost prediction model for distribution network project based on BP neural network

针对规模不断增长的配电网工程所出现的造价管理问题,文中以现行配电网工程项目线路划分为基础,分析造价的主要影响因素,提出一种基于BP神经网络的配电网工程造价预测模型,可以用较少的信息快速准确的对配电网工程造价进行预测。该模型首先对造价的影响因素使用因子分析法进行度量,并将主要因子作为输入,通过3层结构的神经网络得到工程造价输出。文中最后通过运用国内已有的配电网工程中的75段输电线路工程数据对所提出的BP神经网络预测模型进行验证,实验结果表明所提出的预测模型准确性高,具有很好的实用性与可行性,并为日益严峻的配电网工程造价预测问题提供了新的可行的解决方案。

In view of the cost management problem of the growing distribution network project, this paper is based on the existing line division of the distribution network project,analyzes the main influencing factors of cost, and proposes a cost forecasting model of distribution network project based on BP neural network,which can be used to predict the cost of distribution network quickly and accurately with less information. The model first uses the factor analysis method to measure the influence factors of the cost,and takes the main factors as input,and obtains the engineering cost output through the neural network of the 3 layer structure. Finally,the proposed BP neural network prediction model is verified by using the existing 75 section transmission line engineering data of the existing distribution network project in China. The experimental results show that the prediction model proposed in this paper has high accuracy,good practicability and feasibility,and provides a new and feasible solution for the increasingly serious cost prediction of distribution network project.