Japanese Journal of JSCE
Online ISSN : 2436-6021
Special issue (Applied Mechanics) Paper
ESTIMATION OF CROSS-SECTIONAL CHARACTERISTICS BY MACHINE LEARNING FOR EVALUATION OF ADDITIONAL STRESS DUE TO SHEAR LAG
Hiroki AOKIIsao SAIKIYu OTAKERyohei MITSUI
Author information
JOURNAL FREE ACCESS

2023 Volume 79 Issue 15 Article ID: 22-15009

Details
Abstract

The distribution of bending stress along the direction perpendicular to the bridge axis on the flange of beams with a wide flange is not uniform due to shear lag. In the design of beams, the additional stress due to the shear lag is considered by reducing the bending rigidity by the effective width. However, it has been known that the shear lag is not caused by bending but by cross-sectional deformation associated with shear deformation. In this context, a beam theory with a degree of freedom of cross-sectional deformation due to shear is proposed to evaluate shear lag effect. While the beam theory considering cross-sectional deformation has been known to estimate shear lag effect accurately, a finite element analysis of representative volume of cross-section is required to obtain a couple of additional cross-sectional parameters. In this study, we propose a method to estimate the additional parameters using LASSO regression and Gaussian process regression. The accuracy of the proposed method is confirmed by a set of test data.

Content from these authors
© © 2023 Japan Society of Civil Engineers
Previous article Next article
feedback
Top