基于深度神经网络的平行钢丝腐蚀短裂纹疲劳寿命预测研究  

Research on Fatigue Life Prediction of Short Corrosion Cracks in Parallel Steel Wires Based on Deep Neural Network

作　　者：李天翔 刘嘉[1] LI Tian-xiang;LIU Jia(School of Civil Engineering and Architecture,Wuhan University of Technology,Wuhan 430070,China)

机构地区：[1]武汉理工大学土木工程与建筑学院,武汉430070

出　　处：《武汉理工大学学报》2024年第8期53-60,共8页Journal of Wuhan University of Technology

基　　金：国家自然科学基金(51978550,42271453)。

摘　　要：对深度神经网络在针对平行钢丝腐蚀短裂纹疲劳寿命的预测性能进行了研究。根据Voronoi图建立高强钢丝腐蚀坑处的三维晶体代表模型,并基于Tanaka-Mura模型计算不同腐蚀坑以及不同循环荷载大小作用下短裂纹疲劳寿命。根据有限元模拟结果建立数据集,采用深度神经网络以及高斯过程回归对数据集进行预测,结果表明深度神经网络在短裂纹疲劳寿命预测方面有着更好的预测精度。The performance of deep neural network in predicting the fatigue life of short corrosion cracks in parallel steel wires was studied.According to the Voronoi diagram,a three-dimensional crystal representative model of high-strength steel wire corrosion pits was established,and the fatigue life of short cracks under different corrosion pits and different cyclic loads was calculated based on the Tanaka-Mura model.The data set was established according to the finite element simulation results,and the data set was predicted by deep neural network and Gaussian process regression.The results show that the deep neural network has better prediction accuracy in short crack fatigue life prediction.

关 键 词：高强钢丝 Tanaka-Mura模型 机器学习 腐蚀疲劳 疲劳寿命 短裂纹 

分 类 号：U441.4[建筑科学—桥梁与隧道工程]