Anisotropic Mesh Adaptation, the Method of Moving Asymptotes and the global p-norm Stress Constraint
Preprint
Abstract
The p-norm often used in stress constrained topology optimisation supposedly mimics a delta function and it is thus characterised by a small length scale, and the same should apply for the solid-void transition. We propose to resolve these features using anisotropic mesh adaptation. We use the method of moving asymptotes with interpolation of sensitivities, asymptotes and design variables between iterations. We demonstrate this combination for the L-bracket problem with p=10, and we are able to investigate mesh dependence. Finally, we find that the method of moving asymptotes benefits from a relaxation of the problem statement from a geometry with a sharp corner to one with a rounded corner, even if the corner radius is only 1% of the characteristic length scale.Keywords
All Related Versions
This publication has 8 references indexed in Scilit:
- Conditioning of finite element equations with arbitrary anisotropic meshesMathematics of Computation, 2014
- Sensitivity filtering from a continuum mechanics perspectiveStructural and Multidisciplinary Optimization, 2012
- Optimal topologies derived from a phase-field methodStructural and Multidisciplinary Optimization, 2011
- Continuous Mesh Framework Part I: Well-Posed Continuous Interpolation ErrorSIAM Journal on Numerical Analysis, 2011
- Fully anisotropic goal-oriented mesh adaptation for 3D steady Euler equationsJournal of Computational Physics, 2010
- Large-scale topology optimization in 3D using parallel computingComputer Methods in Applied Mechanics and Engineering, 2001
- ?-relaxed approach in structural topology optimizationStructural and Multidisciplinary Optimization, 1997
- The method of moving asymptotes—a new method for structural optimizationInternational Journal for Numerical Methods in Engineering, 1987