Efficient topology optimization for large and dynamically loaded FE models

Wolfgang Witteveen, Klaus Puchner, Karim Sherif, Hans Irschik

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

3 Citations (Scopus)

Abstract

In topology optimization (TopO), a mass distribution in a given design space and for known loads is computed, so that a defined criteria is minimized. TopO of dynamically loaded structures is still a challenge, and publications in the literature have in common that they are quite inefficient for an application to large FE models. The latter, however, is necessary in case of complex design spaces with given edges, bores and the like. This paper introduces a new method derived from the 'equivalent static load (ESL)' - approach, where quasistatic TopO and mode based time integration are performed in a loop. In contrast to the existing methods, damage is used as termination criteria in the loop for the iterative optimization procedure. The ESL for a next TopO step is derived from the most damaged area. Instead of a global optimum search, the FE model is systematically changed until the desired damage level is reached. Numerical examples underline the efficiency of the proposed method even with large FE models.

Original languageEnglish
Title of host publicationIMAC-XXVII
Subtitle of host publicationConference and Exposition on Structural Dynamics - Model Verification and Validation
Publication statusPublished - 2009
Event27th Conference and Exposition on Structural Dynamics 2009, IMAC XXVII - Orlando, FL, United States
Duration: 9 Feb 200912 Feb 2009

Publication series

NameConference Proceedings of the Society for Experimental Mechanics Series
ISSN (Print)2191-5644
ISSN (Electronic)2191-5652

Conference

Conference27th Conference and Exposition on Structural Dynamics 2009, IMAC XXVII
Country/TerritoryUnited States
CityOrlando, FL
Period09.02.200912.02.2009

Fingerprint

Dive into the research topics of 'Efficient topology optimization for large and dynamically loaded FE models'. Together they form a unique fingerprint.

Cite this