基于神经网络的钢管混凝土拱桥可靠度计算  

Reliability calculation of concrete filled steel tubular arch bridge based on neural network

作　　者：余晓琳[1] 颜全胜[1] 李炳秋[1] 丁发兴[2] 

机构地区：[1]华南理工大学土木与交通学院,广州510640 [2]中南大学土木建筑学院,长沙410075

出　　处：《沈阳工业大学学报》2011年第5期582-587,共6页Journal of Shenyang University of Technology

基　　金：国家自然科学基金资助项目(50808180)

摘　　要：由于在大型复杂结构可靠度分析时存在的功能函数不能显式表达,因此采用基于神经网络的响应面法来研究大跨度钢管混凝土拱桥结构的可靠度.运用均匀设计方法确定试验样本点,进行确定性有限元分析,通过获得的训练样本进行神经网络训练,确定神经网络的结构和参数,并结合常规的可靠度分析方法在神经网络响应面上进行可靠度分析.计算结果表明,引入了均匀设计方法的神经网络可以很好地模拟复杂结构的真实响应,该方法运用于大跨度钢管混凝土拱桥结构的可靠度分析是可行的,且具有较高的效率和精度.Since the performance function existing in the reliability analysis of large complex structures can not be expressed explicitly,the reliability of long-span concrete filled steel tubular arch bridge structure was studied with the response surface method based on neural network.The experimental sample points were determined with uniform design method,and the deterministic finite element analysis was carried out.The neural network training was conducted with the obtained training samples,and the structure and parameters of neural network were determined.In addition,in combination with the conventional methods,the reliability analysis was performed on the neural network response surface.The calculated results show that the neural network with introducing the uniform design method can well simulate the real response of complex structures.The present method is feasible in the reliability analysis of long-span concrete filled steel tubular arch bridge structure,and has higher efficiency and accuracy.

关 键 词：神经网络 钢管混凝土 拱桥 拱肋 可靠度 均匀设计 响应面法 串联体系 

分 类 号：TU997[建筑科学—市政工程]