建筑物基础沉降径向基神经网络预测  

Foundation settlement of building and RBF neural networks prediction

作　　者：黄禄文 刘海卿[1] 

机构地区：[1]辽宁工程技术大学建筑工程学院,辽宁阜新123000

出　　处：《辽宁工程技术大学学报（自然科学版）》2016年第8期836-840,共5页Journal of Liaoning Technical University (Natural Science)

基　　金：辽宁省自然科学基金(20102091)

摘　　要：为解决建筑物基础沉降量的安全监测问题,对其进行有效的预测、校核与分析,运用MATLAB软件建立径向基神经网络模型对某市建筑物的基础沉降量进行预测.结果表明:径向基神经网络的结构形式简易,适应能力更强,预测误差比BP网络小,平均约为66.83%,达到预测精准度所需的耗时短、收敛速度更快.径向基神经网络的预测结果与实测结果较为吻合,表明径向基神经网络预测模型适用于建筑工程沉降预测领域之中.In order to solve the problem of the safety monitoring of the building foundation settlement and carry on the effective observation, checking and analysis, this paper used MATLAB software to establish the RBF neural network model to predict the building foundation settlement of a city. The results show that structure of RBF neural networks is simple and has better adaptive capability. RBF neural network prediction error is smaller than that of the BP networks, with the average error about 66.83%, and required time is short for the convergence accuracy. Prediction effect of RBF neural networks are consist with measured effect, which suggests that RBF neural network prediction model are suitable for the field of construction project settlement prediction, which is profound and significant for safety monitoring of the actual project and development.

关 键 词：BP神经网络 径向基神经网络 模型预测 建筑物基础沉降量 安全监测 预测分析 

分 类 号：TU241.8[建筑科学—建筑设计及理论]