Damage identification using static and dynamic responses based on topology optimization and lasso regularization

Ryo Sugai, Akira Saito, Hidetaka Saomoto

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper presents a damage identification method based on topology optimization and Lasso regularization. The method uses static displacements or dynamic responses to identify damages of structures. The method has the potential to identify damages with high fidelity, in comparison with ordinary damage identification method based on optimization with parameterized geometry of the damages. However, it is difficult to precisely detect damage using topology optimization due mostly to the large number of design variables. Therefore, supposing that the damage is sufficiently small, we propose a method adding Lasso regularization to the objective functions to suppress active design variables during topology optimization process. To verify the effectiveness of the proposed method, we conducted a set of numerical experiments for static and dynamic problems. We have succeeded in suppressing active design variables and delete artificially generated damages and the location and shape of damage have been precisely identified.

Original languageEnglish
Title of host publication32nd Conference on Mechanical Vibration and Noise (VIB)
PublisherAmerican Society of Mechanical Engineers (ASME)
ISBN (Electronic)9780791883969
DOIs
Publication statusPublished - 2020
EventASME 2020 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, IDETC-CIE 2020 - Virtual, Online
Duration: 17 Aug 202019 Aug 2020

Publication series

NameProceedings of the ASME Design Engineering Technical Conference
Volume7

Conference

ConferenceASME 2020 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, IDETC-CIE 2020
CityVirtual, Online
Period17/08/2019/08/20

Fingerprint

Dive into the research topics of 'Damage identification using static and dynamic responses based on topology optimization and lasso regularization'. Together they form a unique fingerprint.

Cite this