Artificial neural network for correction of effects of plasticity in equibiaxial residual stress profiles measured by hole drilling

Author:

Chupakhin Sergey1,Kashaev Nikolai1,Klusemann Benjamin12,Huber Norbert1

Affiliation:

1. Division of Materials Mechanics, The Institute of Materials Research, Helmholtz-Zentrum Geesthacht, Geesthacht, Germany

2. Institute of Product and Process Innovation, Leuphana University of Lüneburg, Lüneburg, Germany

Abstract

The hole drilling method is a widely known technique for the determination of non-uniform residual stresses in metallic structures by measuring strain relaxations at the material surface caused through the stress redistribution during drilling of the hole. The integral method is a popular procedure for solving the inverse problem of determining the residual stresses from the measured surface strain. It assumes that the residual stress can be approximated by step-wise constant values, and the material behaves elastically so that the superposition principle can be applied. Required calibration data are obtained from finite element simulations, assuming linear elastic material behavior. That limits the method to the measurement of residual stresses well below the yield strength. There is a lack of research regarding effects caused by residual stresses approaching the yield strength and high through-thickness stress gradients as well as the correction of the resulting errors. However, such high residual stresses are often introduced in various materials by processes such as laser shock peening, for example, to obtain life extension of safety relevant components. The aim of this work is to investigate the limitations of the hole drilling method related to the effects of plasticity and to develop an applicable and efficient method for stress correction, capable of covering a wide range of stress levels. For this reason, an axisymmetric model was used for simulating the hole drilling process in ABAQUS involving plasticity. Afterward, the integral method was applied to the relaxation strain data for determining the equibiaxial stress field. An artificial neural network has been used for solving the inverse problem of stress profile correction. Finally, AA2024-T3 specimens were laser peened and the measured stress fields were corrected by means of the trained network. To quantify the stress overestimation in the hole drilling measurement, an error evaluation has been conducted.

Publisher

SAGE Publications

Subject

Applied Mathematics,Mechanical Engineering,Mechanics of Materials,Modelling and Simulation

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3