摘要
目前在役的装配式梁桥一部分已出现了不同形式和程度的病害,横向连接损伤是最为普遍的病害之一。利用横向连接损伤对应力横向分布系数的影响规律,提出一种基于BP神经网络的装配式梁桥横向连接损伤识别方法,采用应力横向分布系数作为神经网络的输入参数,横向连接损伤位置和程度作为输出参数。并针对BP神经网络存在的缺点,采用遗传算法优化BP神经网络的权值和阈值。使用训练后的神经网络进行吊水岩大桥横向连接损伤定位和损伤程度预测,结果表明该损伤识别方法识别精度较高,可应用于实际工程。
Some fabricated bridges in service have shown different kinds of damage with different levels,one of the most popular diseases is transverse connection damage.Employing the influence of transverse connection damage on transverse distribution coefficient of stress,an approach to identify the transverse connection damage of fabricated bridges based on BP neural network technique was presented in this paper,in which the transverse distribution coefficient is input parameter,damage position and level of transverse connection are output parameters.To overcome the disadvantage of BP neural network,genetic algorithm was utilized to optimize the weight and threshold of the BP neural network.Damage position and level identification for DiaoShuiYan Bridge revealed that this damage identification method possesses high accuracy and can be applicable to practical fabricated bridges.
作者
鄢光显
张龙
谢官模
YAN Guang-xian;ZHANG Long;XIE Guan-mo(School of Science,Wuhan University of Technology,Wuhan 430070,China;China Railway Major Bridge(Nanjing)Bridge and Tunnel Inspect&Retrofit Co Ltd,Nanjing 210061,China)
出处
《武汉理工大学学报》
CAS
北大核心
2019年第6期53-57,共5页
Journal of Wuhan University of Technology
基金
中央高校基本科研业务费专项资金资助(2016IB001)
关键词
装配式梁桥
横向连接损伤识别
应力横向分布系数
遗传算法
BP神经网络
fabricated bridge
transverse connection damage identification
transverse distribution coefficient of stress
genetic algorithm
BP neural network
