Abstract
The paper is concerned with the sensitivity analysis of structural responses in context of linear and non-linear stability phenomena like buckling and snapping. The structural analysis covering these stability phenomena is summarised. Design sensitivity information for a solid shell finite element is derived. The mixed formulation is based on the Hu-Washizu variational functional. Geometrical non-linearities are taken into account with linear elastic material behaviour. Sensitivities are derived analytically for responses of linear and non-linear buckling analysis with discrete finite element matrices. Numerical examples demonstrate the shape optimisation maximising the smallest eigenvalue of the linear buckling analysis and the directly computed critical load scales at bifurcation and limit points of non-linear buckling analysis, respectively. Analytically derived gradients are verified using the finite difference approach.
Similar content being viewed by others
References
Achtziger W, Kočvara M (2008) Structural Topology Optimization with Eigenvalues. SIAM J Optim 18 (4):1129–1164
Arora J (1993) An exposition of the material derivative approach for structural shape sensitivity analysis. Comput Methods Appl Mech Eng 105(1):41–62
Baier H, Seeßelberg C, Specht B (1996) Optimierung in der Strukturmechanik. Vieweg+Teubner, Braunschweig
Barthold FJ (2002) Zur Kontinuumsmechanik inverser Geometrieprobleme. Habilitation, TU Braunschweig
Barthold FJ, Gerzen N, Kijanski W, Materna D (2016) Efficient variational design sensitivity analysis, springer international publishing, Cham, 229–257
Bathe KJ, Dvorkin EN (1983) On the automatic solution of nonlinear finite element equations. Comput Struct 17(5):871–879
Bathe KJ (1996) Finite element procedures. Prentice-Hall, Inc.
Bischoff M, Bletzinger KU, Wall WA, Ramm E (2004) Models and Finite Elements for Thin-Walled Structures, John Wiley & Sons, Ltd
Chan TF (1984) Deflation techniques and block-elimination algorithms for solving bordered singular systems. SIAM J Sci Stat Comput 5(1):121–134
Choi JH (2007) Shape design sensitivity analysis for stability of elastic continuum structures. Int J Solids Struct 44(5):1593–1607
Choi KK, Haug EJ, Lam HL (1982) A numerical method for distributed parameter structural optimization problems with repeated eigenvalues. J Struct Mech 10(2):191–207
Choi KK, Kim NH (2005a) Structural sensitivity analysis and optimization 1–Linear systems mechanical engineering series. Springer, Berlin
Choi KK, Kim NH (2005b) Structural Sensitivity Analysis and Optimization 2–Nonlinear systems and applications Mechanical Engineering Series. Springer, Berlin
Crisfield M (1981) A fast incremental/iterative solution procedure that handles ”snap-through”. Comput Struct 13(1–3):55–62
de Boer H, van Keulen F (2000) Refined semi-analytical design sensitivities. Int J Solids Struct 37:6961–6980
Deml M, Wunderlich W (1997) Direct evaluation of the ’worst’ imperfection shape in shell buckling. Comput Methods Appl Mech Eng 149:201–222. containing papers presented at the Symposium on Advances in Computational Mechanics
El Damatty A, Nassef A (2001) A finite element optimization technique to determine critical imperfections of shell structures. Struct Multidiscipl Optim 23(1):75–87
Elishakoff I, Ohsaki M (2010) Optimization and anti-optimization of structures under uncertainty. World Scientific
Farin G (2002) Curves and surfaces for CAGD, Fifth Edition. Morgan Kaufmann, San Francisco
Fujii F, Ramm E (1997) Computational bifurcation theory: path-tracing, pinpointing and path-switching. Eng Struct 19(5):385–392
Fujii F, Ikeda K, Noguchi H, Okazawa S (2001) Modified stiffness iteration to pinpoint multiple bifurcation points. Comput Methods Appl Mech Eng 190(18–19):2499–2522
Gao X, Ma H (2015) Topology optimization of continuum structures under buckling constraints. Comput Struct 157:142–152
Gerzen N (2014) Analysis and applications of variational sensitivity information in structural optimisation. Dissertation, TU Dortmund
Gerzen N, Barthold F (2013) Variational design sensitivity analysis of a non-linear solid shell with applications to buckling analysis. In: Proceedings of the 10th World Congress on Structural and Multidisciplinary Optimization WCSMO10
Gerzen N, Barthold F, Klinkel S, Wagner W, Materna D (2013) Variational sensitivity analysis of a nonlinear solid shell element. Int J Numer Methods Eng 96(1):29–42
Gu YX, Zhao GZ, Zhang HW, Kang Z, Grandhi RV (2000) Buckling design optimization of complex built-up structures with shape and size variables. Struct Multidiscip Optim 19(3):183–191
Haber RB (1987) Computer aided optimal design Springer, chap A new variational approach to structural shape design sensitivity analysis, 573–587
Hu HT (1994) Buckling optimization of fiber-composite laminate shells considering in-plane shear nonlinearity. Struct Optim 8(2):168–173
Haftka R (1993) Semi-analytical static nonlinear structural sensitivity analysis. AIAA J 31(7):1307–1321
Haug EJ, Rousselet B (1980) Design Sensitivity Analysis in Structural mechanics.II. Eigenvalue Variations. J Struct Mech 8(2):161–186
Haug EJ, Choi KK, Komkov V (1986) Design sensitivity analysis of structural systems. Academic Press, Orlando
Hellweg HB, Crisfield M (1998) A new arc-length method for handling sharp snap-backs. Comput Struct 66(5):704–709
Kegl M, Brank B, Harl B, Oblak MM (2008) Efficient handling of stability problems in shell optimization by asymmetric ’worst-case’ shape imperfection. Int J Numer Methods Eng 73(9):1197–1216
Kemmler R (2004) Stabilität und große Verschiebungen in der Topologie- und Formoptimierung. Dissertation, Universität Stuttgart
Kemmler R, Lipka A, Ramm E (2005) Large deformations and stability in topology optimization. Struct Multidiscipl Optim 30(6):459–476
van Keulen F, Haftka RT, Kim NH (2005) Review of options for structural design sensitivity analysis. part 1: Linear systems. Comput Methods Appl Mech Eng 194:3213–3243
Khot NS, Venkayya VB, Berke L (1976) Optimum structural design with stability constraints. Int J Numer Methods Eng 10(5):1097–1114
Khot NS (1983) Nonlinear analysis of optimized structure with constraints on system stability. AIAA J 21 (8):1181–1186
Khot NS, Kamat MP (1985) Minimum weight design of truss structures with geometric nonlinear behavior. AIAA J 23(1):139–144
Kirikov M, Altus E (2011) Functional gradient as a tool for semi-analytical optimization for structural buckling. Comput Struct 89(17–18):1563–1573
Klinkel S, Wagner W (2008) A piezoelectric solid shell element based on a mixed variational formulation for geometrically linear and nonlinear applications. Comput Struct 86:38–46
Klinkel S, Gruttmann F, Wagner W (2006) A robust non-linear solid shell element based on a mixed variational formulation. Comput Methods Appl Mech Eng 195:179–201
Krenk S (1995) An orthogonal residual procedure for non-linear finite element equations. Int J Numer Methods Eng 38(5):823–839
Kristanič N, Korelc J (2008) Optimization method for the determination of the most unfavorable imperfection of structures. Comput Mech 42(6):859–872
Kwon TS, Lee BC, Lee WJ (1999) An approximation technique for design sensitivity analysis of the critical load in non-linear structures. Int J Numer Methods Eng 45(12):1727– 1736
Leon SE, Paulino GH, Pereira A, Menezes IFM, Lages EN (2014) A Unified Library of Nonlinear Solution Schemes. Appl Mech Rev 64(4):040803–1–26
Leon SE, Lages EN, de Araújo CN, Paulino GH (2014) On the effect of constraint parameters on the generalized displacement control method. Mech Res Commun 56:123–129
Lindgaard E, Lund E, Rasmussen K (2010) Nonlinear buckling optidmization of composite structures considering ”worst” shape imperfections. Int J Solids Struct 47(22–23):3186–3202
Lindgaard E, Lund E (2010) Nonlinear buckling optimization of composite structures. Comput Methods Appl Mech Eng 199(37–40):2319–2330
Lund E, Olhoff N (1994) Shape design sensitivity analysis of eigenvalues using exact numerical differentiation of finite element matrices. Struct Optim 8(1):52–59
Lund E (2009) Buckling topology optimization of laminated multi-material composite shell structures. Compos Struct 91(2):158– 167
Mateus H, Soares C, Soares C (1997) Buckling sensitivity analysis and optimal design of thin laminated structures. Comput Struct 64(1–4):461–472. computational Structures Technology
Mróz Z, Piekarski J (1998) Sensitivity analysis and optimal design of non-linear structures. Int J Numer Methods Eng 42(7):1231– 1262
Nocedal J, Wright SJ (2006) Numerical optimization. Springer, New York
Noguchi H, Hisada T (1993) Sensitivity analysis in post-buckling problems of shell structures. Comput Struct 47(4–5):699–710
Ohsaki M (2002) Maximum loads of imperfect systems corresponding to stable bifurcation. Int J Solids Struct 39(4):927–941
Ohsaki M (2005) Design sensitivity analysis and optimization for nonlinear buckling of finite-dimensional elastic conservative structures. Comput Methods Appl Mech Eng 194(30–33):3331– 3358
Ohsaki M, Uetani K (1996) Sensitivity analysis of bifurcation load of finite dimensional symmetric systems. Int J Numer Methods Eng 39(10):1707–1720
Ohsaki M, Uetani K, Takeuchi M (1998) Optimization of imperfection-sensitive symmetric systems for specified maximum load factor. Comput Methods Appl Mech Eng 166(3–4):349– 362
Parente E, Vaz LE (2003) On evaluation of shape sensitivities of non-linear critical loads. Int J Numer Methods Eng 56(6):809–846
Park JS, Choi KK (1990) Design sensitivity analysis of critical load factor for nonlinear structural systems. Comput Struc 36(5):823–838
Pedersen NL (2000) Maximization of eigenvalues using topology optimization. Struct Multidisc Optim 20 (1):2–11
Powell G, Simons J (1981) Improved iteration strategy for nonlinear structures. Int J Numer Methods Eng 17(10):1455–1467
Ramm E (1981) Strategies for tracing nonlinear response near limit points. In: Wunderlich W, Stein E, Bathe K J (eds) Nonlinear Finite Element Analysis in Structural Mechanics, Springer and Berlin and Heidelberg, pp 63–89
Reitinger R (1994) Stabilität und optimierung imperfektionsempfindlicher tragwerke. Dissertation, Universität Stuttgart
Reitinger R, Bletzinger KU, Ramm E (1994) Shape optimization of buckling sensitive structures. Comput Syst Eng 5(1):65–75
Reitinger R, Bletzinger KU, Ramm E (1994) Buckling and imperfection sensitivity in the optimization of shell structures. Thin-Walled Struct 23(1):159–177
Riks E (1972) The application of newton’s method to the problem of elastic stability. J Appl Mech 39 (4):1060–1065
Riks E (1979) An incremental approach to the solution of snapping and buckling problems. Int J Solids Struct 15(7):529–551
Ritto-Corrêa M, Camotim D (2008) On the arc-length and other quadratic control methods: established, less known and new implementation procedures. Comput Struct 86(11–12):1353–1368
Schweizerhof K, Wriggers P (1986) Consistent linearization for path following methods in nonlinear fe analysis. Comput Methods Appl Mech Eng 59(3):261–279
Seyranian A, Lund E, Olhoff N (1994) Multiple eigenvalues in structural optimization problems. Struct Optim 8(4):207–227
Thompson J (1972) Optimization as a generator of structural instability. Int J Mech Sci 14:627–629
Tortorelli DA, Michaleris P (1994) Design sensitivity analysis: Overview and review. Inverse Prob Eng 1 (1):71–105
Tortorelli DA, Zixian W (1993) A systematic approach to shape sensitivity analysis. Int J Solids Struct 30 (9):1181–1212
Wagner W (1991) Zur behandlung von stabilitätsproblemen der elastostatik mit der methode der finiten elemente. Habilitation, Universität Hannover
Wempner GA (1971) Discrete approximations related to nonlinear theories of solids. Int J Solids Struct 7 (11):1581–1599
Wriggers P (2008) Nonlinear finite element methods springer. Berlin, Heidelberg
Wriggers P, Simo J (1990) A general procedure for the direct computation of turning and bifurcation points. Int J Numer Methods Eng 30(1):155–176
Wriggers P, Wagner W, Miehe C (1988) A quadratically convergent procedure for the calculation of stability points in finite element analysis. Comput Methods Appl Mech Eng 70(3):329– 347
Wu C, Arora J (1988) Design sensitivity analysis of non-linear buckling load. Comput Mech 3(2):129–140
Yang YB, Shieh MS (1990) Solution method for nonlinear problems with multiple critical points. AIAA J 28(12):2110–2116
Özakça M, Tayşi N, Kolcu F (2003) Buckling analysis and shape optimization of elastic variable thickness circular and annular plates—ii. shape optimization. Eng Struct 25(2):193–199
Acknowledgments
We gratefully acknowledge the support of the German Research Foundation (DFG) under grant no. BA 1828/5-1.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix: A: Structural analysis for solid shells
A brief summary of the element formulation presented by Klinkel et al. (2006) and Klinkel and Wagner (2008) for structural analysis is given. Additional sensitivity quantities by means of the pseudo load and sensitivity matrices presented by Gerzen et al. (2013) and (Gerzen 2014) are summarised as well. The shell continuum Ω R is divided in element domains Ω R e , which can be expressed with the assembly over all elements \({\Omega }_{R}=\overset {\text {nel}}{\underset {{e=1}}{{{ A}}}}{\Omega }_{Re}\). A low order hexahedral solid shell element is build. Isoparametric trilinear approximations of geometry and displacement field result in an eight node solid shell element with three displacement degrees of freedom per node. The superscript h indicates the field variables after discretisation. The subscript e indicates quantities on element level.
1.1 A.1 Approximation of kinematic quantities
Geometry, displacements and its variations are interpolated in the same manner
The discrete nodal coordinates, displacements and variations are arranged in the vectors \( \hat {\boldsymbol {X}}_{e}\in \mathbb {R}^{24\times 1}\), \( \hat {\boldsymbol {u}}_{e}\in \mathbb {R}^{24\times 1}\) and \(\delta \hat {\boldsymbol {u}}_{e}\in \mathbb {R}^{24\times 1}\), respectively. The shape functions for the nodes I=1,2,...,8
with −1≤ξ i≤+1 are organised in the interpolation matrix N=[N 1,...,N 8] with the submatrix N I =diag[N I ,N I ,N I ]. The Cartesian coefficients of the Green-Lagrangian strain tensor E are ordered in the vector E=[E 11,E 22,E 33,2E 12,2E 13,2E 23]T in Voigt notation. The covariant basis vectors in discrete form are
with N ,i containing the derivatives of the shape functions with respect to the convective coordinates ξ i. The necessary derivatives of the shape functions with respect to the global coordinates
can be computed.
1.2 A.2 Approximation of strains and their variations.
The local convective strain components are approximated with
The corresponding strains are computed via ANS interpolation. A transformation to Cartesian coordinates
can be done using the transformation matrix \(\boldsymbol {T}_{S}^{-T}\). It is defined via \(\boldsymbol {T}_{S}=\boldsymbol {T}(\bar {a},\bar {b})\) with \(\bar {a}=2\), \(\bar {b}=1\) and
and \(J_{ik}=\boldsymbol {e}_{i}\cdot \boldsymbol {G}_{k}^{h}\). The vectors \(\boldsymbol {G}_{k}^{h}\) are the well known convective tangent vectors and e i are the orthogonal unit base vectors of Cartesian space. On element level the approximation of the virtual and incremental Green-Lagrangian strains reads
respectively. For the approximation the interpolation matrix reads
The submatrix B L I at I-th node is given by
with
The superscripts A,B,C,D denote collocation points of assumed natural strain (ANS) interpolation for the treatment of transverse shear locking A=(−1,0,0), B=(0,−1,0), C=(1,0,0) and D=(0,1,0) in convective coordinates ξ i. To overcome curvature thickness locking the collocation points i=(−1,−1,0), i i=(1,−1,0), i i i=(1,1,0) and i v=(−1,1,0) in convective coordinates ξ i are chosen. They are denoted with superscript L = i,i i,i i i,i v. Details on how to choose these collocation points can be found in Klinkel et al. (2006) and references therein. Due to the ANS interpolations from now on the element formulation is not isotropic any more and ξ 3 denotes the thickness direction. The quantity \(\Delta \delta \boldsymbol {E}:{\hat {\boldsymbol {S}}}\) from the linearisation of the weak form (120) is approximated in the following way The first quantity K e is obtained by the discretisation of
given by the submatrices G I J =diag[G I J ,G I J ,G I J ] for the node combination I, J defined by the scalar
The required matrix B I J is known from (35).
1.3 A.3 Approximation of assumed strain fields
The strain tensor \({\bar {\boldsymbol {E}}}\) is additively decomposed
The components of the strain fields \(\hat {\boldsymbol {E}}\) and \(\tilde {\boldsymbol {E}}\) are interpolated in local convective co-ordinates and transformed to Cartesian coordinates using Voigt Notation. The transformation of the contravariant components \(\hat {E}^{ij}\) is done by the transformation matrix \(\boldsymbol {T}_{E}=\boldsymbol {T}(\bar {a},\bar {b})\) with \(\bar {a}=1\), \(\bar {b}=2\), cf. (102). The approximation of the strain field in vector notation is
with
Quantities evaluated at the centre of the element are denoted with the superscript 0. \(\boldsymbol {I}\in \mathbb {R}^{6\times 6}\) is the identity matrix. The interpolation matrices in natural coordinates are
and
Klinkel et al. (2006) call \(\tilde {\boldsymbol {E}}\) the enhanced assumed strain field. Its covariant components are interpolated and transformed to Cartesian space using the relations
and
\(\boldsymbol {J}=[\boldsymbol {G}_{1}^{h},\boldsymbol {G}_{2}^{h},\boldsymbol {G}_{3}^{h}]^{T}\) is the Jacobian matrix. The interpolation matrix for the enhanced assumed strain field in natural coorinates reads
The interpolation of the total strain can now be expressed as
Derivatives of strain energy function yield the 2nd Piola-Kirchhoff stresses and the material matrix
respectively.
1.4 A.4 Approximation of assumed stress fiels
The interpolation of the stress field \( \hat {S}^{h}\) reads
The same procedure is used for the virtual stresses \(\delta \hat {S}^{h}\) and the incremental stresses \(\delta \hat {S}^{h}\). The transformation to global co-ordinates is done by the transformation matrix \(\boldsymbol {T}_{S}^{0}=\boldsymbol {T}(\bar {a},\bar {b})\) with \(\bar {a}=2\), \(\bar {b}=1\). The superscript 0 denotes quantities evaluated at the element centre. The transformation matrix T is given by (102).
1.5 A.5 Element matrices and vectors
Element matrices are
and element vectors are
Appendix: B: Sensitivity analysis for solid shells
Quantities for design sensitivity analysis by means of the pseudo load and sensitivity matrices derived by Gerzen et al. (2013) are summarised briefly.
1.1 B.1 Variational relations
For the computation of total derivatives of objectives and constraints after (43) pseudo load and sensitivity operator are desired. The pseudo load operator is obtained as variation of the physical residual with respect to the design
with the partial derivatives
1.2 B.2 Discretised relations
For the discretisation of the pseudo load matrix, derivatives of transformation matrices and strains as well as the approximation of a divergence are desired. The term Divδ X is approximated with
and
The derivative of the transformation matrix T from (102) with respect to \( \hat {\boldsymbol {X}}_{e}\) is denoted with \({\mathbb {T}}\in \mathbb {R}^{6\times 6\times 24}\). Its coefficients are
with I=1,...,8 . Submatrices are
with
and
including
with
and
including
and finally
with
and
including
The transposed \({\mathbb {T}}^{T}\) is defined as \({\mathbb {T}}^{T}_{ijk}={\mathbb {T}}_{jik}\). The derivatives of applied transformation matrices read for the stresses \({\mathbb {T}}_{S}={\mathbb {T}}(\bar {a},\bar {b})\) with \(\bar {a}=2\), \(\bar {b}=1\) and for strains \({\mathbb {T}}_{E}={\mathbb {T}}(\bar {a},\bar {b})\) with \(\bar {a}=1\), \(\bar {b}=2\). Note that the notation means \(T_{ij,\hat {X}_{k}}={\mathbb {T}}_{ijk}\). The derivative of the transformation matrix T M with respect to \( \hat {\boldsymbol {X}}_{e}\) is \({\mathbb {T}}_{M}\in \mathbb {R}^{6\times 6\times 24}\). With \((T_{M})_{ij,\hat {X}_{k}}=({\mathbb {T}}_{M})_{ijk}\) the derivative \((\boldsymbol {T}_{M})_{,\hat {X}_{k}}\) can be computed with J=[G 1 G 2 G 3] and \(\boldsymbol {J}^{0}=[\boldsymbol {G}_{1}^{0} \boldsymbol {G}_{2}^{0} \boldsymbol {G}_{3}^{0}]\) resulting in
The scalar product (⋅:⋅) is applied to matrices, as it is defined for tensors. The computation of \(\boldsymbol {J}^{0}_{,\hat {X}_{k}}\) and \(\boldsymbol {J}_{,\hat {X}_{k}}\) is straightforward, details are omitted here. The first derivative of local strains \(\boldsymbol {E}^{h}_{L}\) with respect to \( \hat {X}_{e}\) is \((\boldsymbol {E}^{h}_{L})_{,\hat {X}}=\boldsymbol {Q}=[\boldsymbol {Q}_{1},...,\boldsymbol {Q}_{8}]\) with
and
1.3 B.3 Sensitivity and pseudo load matrices
The parts of the pseudo load are approximated as
with matrices
and
The pseudo load and sensitivity matrices P v and S v are obtained by assembly over all finite elements of
of the element quantities
respectively. To compute the sensitivity matrix their relation
is used. Static condensation yields
with
and the element quantity
using the abbreviation \(\boldsymbol {P}_{{E}}^e=\boldsymbol {P}_{{\hat {E}}}^{e}-\boldsymbol {A}_{e}^{12}(\boldsymbol {A}_{e}^{22})^{-1}\boldsymbol {P}_{{\tilde {E}}}^e\). After solving the unknown displacement sensitivities, the sensitivities of stresses and strains can be calculated on element level as follows
Rights and permissions
About this article
Cite this article
Radau, L., Gerzen, N. & Barthold, FJ. Sensitivity of structural response in context of linear and non-linear buckling analysis with solid shell finite elements. Struct Multidisc Optim 55, 2259–2283 (2017). https://doi.org/10.1007/s00158-016-1639-3
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00158-016-1639-3