基于Levenberg-Marquardt算法的应力集中预测  

Prediction of Stress Concentration Based on Levenberg-Marquardt Algorithm

作　　者：刘志明[1] 王钟羡[1] 苏虹[1] 朱玉萍[1] 

机构地区：[1]江苏理工大学理学院,江苏镇江212013

出　　处：《江苏理工大学学报（自然科学版）》2001年第6期84-87,共4页Journal of Jiangsu University of Science and Technology(Natural Science)

基　　金：江苏省自然科学基金资助项目 (BK9910 9)

摘　　要：应用应变电测的基本原理及技术 ,测量所研究试件应力集中附近处的应变值 以神经网络技术为基础 ,采用单隐层网络结构 ,以Levenberg -Marquardt算法来训练网络 笔者在收集已有相关资料的基础上 ,对应力集中处的最大值进行了预测 通过与常用的应力集中预测模型进行数值计算比较 ,表明该模型预测精度高 ,可靠性强 模型中所采用的数据预处理方法 。Stresses near a region of stress concentration are measured by using the fundamental theory and technology of electric measurement of strain gauge. A single hidden hierarchical neural network is applied to build up a model that uses a Levenberg Marquardt algorithm to train the neural network. Based on collected data in question, prediction of the maximum stress in terms of stresses measured has been developed. The results indicate that it has higher precision and reliability in prediction when compared through numerical computation with other typical stress concentration models. In addition, the pre processing method of data adopted in this paper greatly enhances the scope of application of the model.

关 键 词：应力集中 应变电测 人工神经网络 LEVENBERG-MARQUARDT算法 预测模型 

分 类 号：O343.4[理学—固体力学] TP183[理学—力学]