liu.seSearch for publications in DiVA
Change search
ReferencesLink to record
Permanent link

Direct link
Large-scale robust topology optimization under load-uncertainty
Linköping University, Department of Management and Engineering, Solid Mechanics. Linköping University, Faculty of Science & Engineering.
Linköping University, Department of Management and Engineering, Solid Mechanics. Linköping University, Faculty of Science & Engineering.
Linköping University, Department of Management and Engineering, Solid Mechanics. Linköping University, Faculty of Science & Engineering.
2015 (English)In: Advances in Structural and Multidisciplinary Optimization - Proceedings of the 11th World Congress of Structural and Multidisciplinary Optimization(WCSMO-11) / [ed] Qing Li, Grant P Steven and Zhongpu (Leo) Zhang, 2015, 255-260 p.Conference paper (Refereed)
Abstract [en]

Structures designed by topology optimization (TO) are frequently sensitive to loads different from the ones accounted for in the optimization. In extreme cases this means that loads differing ever so slightly from the ones it was designed to carry may cause a structure to collapse. It is therefore clear that handling uncertainty regarding the actual loadings is important. To address this issue in a systematic  manner is one of the main goals in the field of robust TO. In this work we present a deterministic robust formulation of TO for maximum stiffness design which accounts for uncertain variations around a set of nominal loads. The idea is to find a design which minimizes the maximum compliance obtained as the loads vary in infinite, so-called uncertainty sets. This naturally gives rise to a semi-infinite optimization problem, which we here reformulate into a non-linear, semi-definite program. With appropriate numerical algorithms this optimization problem can be solved at a cost similar to that of solving a standard multiple load-case TO problem with the number of loads equal to the number of spatial dimensions plus one, times the number of nominal loads. In contrast to most previously suggested methods, which can only be applied to small-scale problems, the presented method is – as illustrated by a numerical example – well-suited for large-scale TO problems.

Place, publisher, year, edition, pages
2015. 255-260 p.
Keyword [en]
Robust optimization, Topology optimization, Large-scale optimization, Non-linear semi-definite programming
National Category
Applied Mechanics
Identifiers
URN: urn:nbn:se:liu:diva-123004ISBN: 978-0-646-94394-7OAI: oai:DiVA.org:liu-123004DiVA: diva2:875671
Conference
11th World Congress on Structural and Multidisciplinary Optimization (WCSMO-11), Sydney Australia, 7–12 June
Available from: 2015-12-01 Created: 2015-12-01 Last updated: 2015-12-09Bibliographically approved
In thesis
1. Topology optimization considering stress, fatigue and load uncertainties
Open this publication in new window or tab >>Topology optimization considering stress, fatigue and load uncertainties
2016 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

This dissertation concerns structural topology optimization in conceptual design stages. The objective of the project has been to identify and solve problems that prevent structural topology optimization from being used in a broader sense in the avionic industry; therefore the main focus has been on stress and fatigue constraints and robustness with respect to load uncertainties.

The thesis consists of two parts. The first part gives an introduction to topology optimization, describes the new contributions developed within this project and motivates why these are important. The second part includes five papers.

The first paper deals with stress constraints and a clustered approach is presented where stress constraints are applied to stress clusters, instead of being defined for each point of the structure. Different approaches for how to create and update the clusters, such that sufficiently accurate representations of the local stresses are obtained at a reasonable computational cost, are developed and evaluated.

High-cycle fatigue constraints are developed in the second paper, where loads described by a variable-amplitude load spectrum and material data from fatigue tests are used to determine a limit stress, for which below fatigue failure is not expected. A clustered approach is then used to constrain the tensile principal stresses below this limit.

The third paper introduces load uncertainties and stiffness optimization considering the worst possible loading is then formulated as a semi-definite programming problem, which is solved very efficiently. The load is due to acceleration of point masses attached to the structure and the mass of the structure itself, and the uncertainty concerns the direction of the acceleration. The fourth paper introduces an extension to the formulated semi-definite programming problem such that both fixed and uncertain loads can be optimized for simultaneously.

Game theory is used in the fifth paper to formulate a general framework, allowing essentially any differentiable objective and constraint functions, for topology optimization under load uncertainty. Two players, one controlling the structure and one the loads, are in conflict such that a solution to the game, a Nash equilibrium, is a design optimized for the worst possible load.

Place, publisher, year, edition, pages
Linköping: Linköping University Electronic Press, 2016. 63 p.
Series
Linköping Studies in Science and Technology. Dissertations, ISSN 0345-7524 ; 1730
National Category
Applied Mechanics
Identifiers
urn:nbn:se:liu:diva-123008 (URN)10.3384/diss.diva-123008 (DOI)978-91-7685-883-7 (print) (ISBN)
Public defence
2016-01-15, C3, C-huset, Campus Valla, Linköping, 10:15 (English)
Opponent
Supervisors
Available from: 2015-12-01 Created: 2015-12-01 Last updated: 2015-12-11Bibliographically approved

Open Access in DiVA

No full text

Search in DiVA

By author/editor
Thore, Carl-JohanHolmberg, ErikKlarbring, Anders
By organisation
Solid MechanicsFaculty of Science & Engineering
Applied Mechanics

Search outside of DiVA

GoogleGoogle Scholar
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Total: 59 hits
ReferencesLink to record
Permanent link

Direct link