A posteriori error estimates for sequential laminates in shape optimization

DOI10.3934/DCDSS.2016055MaRDI QIDQ727511zbMATH OpenOpenAlexFDO

Authors Benedict Geihe, Martin Rumpf

Publication date 7 December 2016

Published in Discrete and Continuous Dynamical Systems. Series S (Search for Journal in Brave)

Copyright license Creative Commons Attribution 4.0 International

Full work available at URL https://arxiv.org/abs/1501.07461

zbMATH Keywords

adaptive meshes elastic shape optimization linear elasticity nested laminates two scale optimization

Mathematics Subject Classification ID

Homogenization in context of PDEs; PDEs in media with periodic structure (35B27) Numerical methods involving duality (49M29) Optimization of shapes other than minimal surfaces (49Q10) Error bounds for boundary value problems involving PDEs (65N15) Finite element, Rayleigh-Ritz and Galerkin methods for boundary value problems involving PDEs (65N30) Classical linear elasticity (74B05) Compliance or weight optimization in solid mechanics (74P05) Homogenization in equilibrium problems of solid mechanics (74Q05) Finite element methods applied to problems in solid mechanics (74S05)

Abstract: A posteriori error estimates are derived in the context of two-dimensional structural elastic shape optimization under the compliance objective. It is known that the optimal shape features are microstructures that can be constructed using sequential lamination. The descriptive parameters explicitly depend on the stress. To derive error estimates the dual weighted residual approach for control problems in PDE constrained optimization is employed, involving the elastic solution and the microstructure parameters. Rigorous estimation of interpolation errors ensures robustness of the estimates while local approximations are used to obtain fully practical error indicators. Numerical results show sharply resolved interfaces between regions of full and intermediate material density.

Review text

The authors consider a 2D linear elastic shape optimization problem which consists to minimize a cost functional among mixtures between hard and soft elastic materials with a given amount of hard material in a circumjacent domain \(D\subset \mathbb{R}^{2}\): \(\int_{D}\chi dx=\Theta \), where \(\chi \in L^{\infty }(D,\{0,1\})\). The hard material fills a subdomain \(\Omega \subset D\), the material is held fixed on \(\Gamma _{D}\subset \partial \Omega \cap \partial D\) and tractions \(g\) are applied on \(\Gamma _{N}\subset \partial \Omega \cap \partial D\), leading to the set \(\Sigma =\{\sigma :D\rightarrow \mathbb{R}_{sym}^{2\times 2}\mid div\{\sigma \}=0\) in \(D\), \(\sigma n=g \) on \(\Gamma _{N}\), \(u=0\) on \(\Gamma _{D}\), \(\sigma =C[q]\varepsilon [ u]\}\) of feasible stresses. The cost functional is taken as \(J[u[\chi ];\chi ]=\int_{\Gamma _{N}}g\cdot u[\chi ]da(x)=l(u[\chi ])=a(q;u[\chi ],u[\chi ])\) where \(l\) is a Lagrange multiplier associated to the amount of hard material and \(a(q;u,\varphi )=\int_{D}C[q]\varepsilon [ u]:\varepsilon [ \varphi ]dx\), with \(C[q]=\chi A[q]+(1-\chi )B[q]\), \( A[q]\) and \(B[q]\) being the elasticity tensors associated to the hard and soft materials, respectively. It is known that this shape optimization problem may be ill-posed with the onset of microstructures. The authors thus consider the relaxed problem \(\min_{\sigma \in \Sigma }\int_{D}\min_{0\leq \theta \leq 1}\min_{C^{\ast }[q]\in G_{\theta }^{B}}C^{\ast }[q]^{-1}\sigma :\sigma +l\theta dx\), where \(G_{\theta }^{B}\) is the set of homogenized tensors. They consider sequential laminate microstructures with a main laminate direction \(\alpha (x)\), a ratio of material spent in the second lamination stage \(m(x)\) and the overall local material density \(\theta (x)\). They give the expressions of the homogenized elasticity tensors in this case. The main purpose of the paper is to prove a posteriori estimates for a finite element discretization of this shape optimization problem. The authors introduce a triangulation of a polygonally bounded domain \(D\) and piecewise bi-quadratic and continuous vector-valued functions with vanishing trace on \(\Gamma _{D}\). The authors first prove that the local elastic energy is invariant with respect to \(\alpha (x)\). The main result of the paper proves an estimate for the difference \(\left| J[q;u]-J[q_{h};u_{h}]\right| \), and the paper ends with the presentation of numerical simulations.

Recommendations

Cites work

Cited in

(2)

This page was built for publication: A posteriori error estimates for sequential laminates in shape optimization

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q727511)