基于GA-BP神经网络的钢桁架桥节点结构优化分析
Optimization Analysis of the Joint Structure in Steel Truss Bridge Based on GA-BP Neural Network
张留1
作者信息
- 1. 中铁十五局集团第四工程有限公司 新郑 451100
- 折叠
摘要
为克服有限元法建模工作量大,不能有效反映节点局部应力状态的问题,文中依托钢桁架桥梁,建立GA-BP神经网络节点结构优化预测模型,设置N1横隔板厚度、N2横隔板厚度、节点板厚度、加劲肋厚度、加劲肋高度5个参数、7个水平、共42组样本数据,确定权值和阈值,对节点结构进行优化分析.结果表明,GA-BP神经网络模型可以对节点结构组成部件进行优化,预测的节点最大应力值与有限元分析结果相差较小,验证了提出模型的有效性.
Abstract
In order to overcome the problem of large modeling workload and inability to effectively re-flect the local stress state of nodes in finite element analysis methods,a GA-BP neural network node structure optimization prediction model was established based on steel truss bridges.Five parameters,including N1 diaphragm thickness,N2 diaphragm thickness,node plate thickness,stiffener thick-ness,and stiffener height,were set up at seven levels.A total of 42 sets of sample data were collect-ed,with weights and thresholds determined,and the node structure was optimized and analyzed.The results indicate that the GA-BP neural network model can optimize the structural components of nodes,and the predicted maximum stress value of nodes is similar to the finite element analysis re-sults,which verifies the effectiveness of the proposed model.
关键词
钢桁架桥/节点/GA-BP神经网络/优化Key words
steel truss bridge/joint/GA-BP neural network/optimization引用本文复制引用
基金项目
中铁十五局集团科研项目(CR1504-GL-YX-2021-GC-JS-006)
出版年
2024
NETL
NSTL
万方数据