Skip to main content
Log in

Efficient fixed point and Newton–Krylov solvers for FFT-based homogenization of elasticity at large deformations

  • Original Paper
  • Published:
Computational Mechanics Aims and scope Submit manuscript

Abstract

In recent years the FFT-based homogenization method of Moulinec and Suquet has been established as a fast, accurate and robust tool for obtaining effective properties in linear elasticity and conductivity problems. In this work we discuss FFT-based homogenization for elastic problems at large deformations, with a focus on the following improvements. Firstly, we exhibit the fixed point method introduced by Moulinec and Suquet for small deformations as a gradient descent method. Secondly, we propose a Newton–Krylov method for large deformations. We give an example for which this methods needs approximately 20 times less iterations than Newton’s method using linear fixed point solvers and roughly \(100\) times less iterations than the nonlinear fixed point method. However, the Newton–Krylov method requires 4 times more storage than the nonlinear fixed point scheme. Exploiting the special structure we introduce a memory-efficient version with 40 % memory saving. Thirdly, we give an analytical solution for the micromechanical solution field of a two-phase isotropic St.Venant–Kirchhoff laminate. We use this solution for comparison and validation, but it is of independent interest. As an example for a microstructure relevant in engineering we discuss finally the application of the FFT-based method to glass fiber reinforced polymer structures.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13

Similar content being viewed by others

Notes

  1. www.itwm.fraunhofer.de/FeelMath

  2. www.geodict.com/ElastoDict

  3. It is possible to work with three deformation gradients but only at the expense of two matrix operations per iteration.

  4. One should not confuse the position variable \(X\) with the solution vector \(X\) from (38).

  5. www.geodict.com

  6. www.fftw.org

References

  1. Advani SG, Tucker III, CL (1987) The use of tensors to describe and predict fiber orientation in short fiber composites. J Rheol 31(8), 751–784, doi:10.1122/1.549945. URL http://link.aip.org/link/?JOR/31/751/1

  2. Agoras M, Castañeda PP (2012) Multi-scale homogenization of semi-crystalline polymers. Phil Mag 92(8):925–958. doi:10.1080/14786435.2011.637982

    Article  Google Scholar 

  3. Axelsson O, Kaporin IE (2000) On the sublinear and superlinear rate of convergence of conjugate gradient methods. Numer Algorithm 25(1–4):1–22. doi:10.1023/A:1016694031362

    Article  MATH  MathSciNet  Google Scholar 

  4. Bertram A, Böhlke T, Šilhavý M (2007) On the rank 1 convexity of stored energy functions of physically linear stress-strain relations. J Elast 86(3):235–243. doi:10.1007/s10659-006-9091-z

    Article  MATH  Google Scholar 

  5. Boyd JP (1989) Chebyshev and Fourier spectral methods. Springer-Verlag, Berlin

  6. Braess D (2007) Finite elements: theory. Fast solvers and applications in solid mechanics. Cambridge University Press, Cambridge

    Book  Google Scholar 

  7. Brighi B, Bousselsal M (1995) On the rank-one-convexity domain of the Saint Venant-Kirchhoff stored energy function. Rendiconti del Seminario Matematico della Università di Padova 94, 25–45. URL http://eudml.org/doc/108375

  8. Brisard S, Dormieux L (2010) FFT-based methods for the mechanics of composites: A general variational framework. Comput Mater Sci 49(3), 663–671. doi:10.1016/j.commatsci.2010.06.009. URL: http://www.sciencedirect.com/science/article/pii/S0927025610003563

  9. Brisard S, Dormieux L (2012) Combining Galerkin approximation techniques with the principle of Hashin and Shtrikman to derive a new FFT-based numerical method for the homogenization of composites. Comput Methods Appl Mech Eng 217–220(0):197–212. doi:10.1016/j.cma.2012.01.003. URL http://www.sciencedirect.com/science/article/pii/S0045782512000059

  10. Castañeda PP (1996) Exact second-order estimates for the effective mechanical properties of nonlinear composite materials. J Mech Phys Solids 44(6):827–862

    Article  MATH  MathSciNet  Google Scholar 

  11. Castañeda PP, Suquet P (1998) Nonlinear composites. Adv Appl Mech 34(998):171–302

    Google Scholar 

  12. Ciarlet PG (1988) Mathematical elasticity: three-dimensional elasticity, vol I. Elsevier, Amsterdam

    MATH  Google Scholar 

  13. Eisenlohr P., Diehl, M., Lebensohn, R., Roters, F.: A spectral method solution to crystal elasto-viscoplasticity at finite strains. Int J Plast 46(0), 37–53 (2013). doi:10.1016/j.ijplas.2012.09.012. URL http://www.sciencedirect.com/science/article/pii/S0749641912001428

  14. Eyre DJ, Milton GW (1999) A fast numerical scheme for computing the response of composites using grid refinement. Eur Phys J 6(01):41–47

    Google Scholar 

  15. Feyel, F., Chaboche, J.L.: FE2 multiscale approach for modelling the elastoviscoplastic behaviour of long fibre sic/ti composite materials. Comput Methods Appl Mech Eng, 183(3–4), 309–330 (2000). doi:10.1016/S0045-7825(99)00224-8. URL http://www.sciencedirect.com/science/article/pii/S0045782599002248

  16. Francfort G (1983) Homogenization and linear thermoelasticity. SIAM J Math Anal 14(4):696–708. doi:10.1137/0514053

    Article  MATH  MathSciNet  Google Scholar 

  17. Gélébart, L., Mondon-Cancel, R.: Non-linear extension of FFT-based methods accelerated by conjugate gradients to evaluate the mechanical behavior of composite materials. Comput Mater Sci 77(0), 430–439 (2013). doi:10.1016/j.commatsci.2013.04.046. URL http://www.sciencedirect.com/science/article/pii/S0927025613002188

  18. Hestenes MR, Stiefel E (1952) Methods of conjugate gradients for solving linear systems. J Res Nat Bureau Stand 49:409–436

    Article  MATH  MathSciNet  Google Scholar 

  19. Hill R (1972) On constitutive macro-variables for heterogeneous solids at finite strain. Proc R Soc Lond A 326:131–147

    Article  MATH  Google Scholar 

  20. Johnson SG, Frigo M (2007) A modified split-radix FFT with fewer arithmetic operations. Signal Process IEEE Trans on 55(1):111–119

  21. Kocks UF, Tome CN, Wenk HR (1998) Texture and anisotropy: preferred orientations in polycrystals and their effect on materials properties. Cambridge University Press, Cambridge

    MATH  Google Scholar 

  22. Krawietz A (1986) Materialtheorie. Springer-Verlag, Berlin

    Book  MATH  Google Scholar 

  23. Kröner E (1971) Statistical continuum mechanics. Springer, Wien

    Book  Google Scholar 

  24. Kröner, E.: Bounds for effective elastic moduli of disordered materials. J Mech Phys Solids 25(2), 137–155 (1977). doi:10.1016/0022-5096(77)90009-6. URL http://www.sciencedirect.com/science/article/pii/0022509677900096

  25. Lahellec, N., Michel, J.C., Moulinec, H., Suquet, P.: Analysis of inhomogeneous materials at large strains using fast Fourier transforms. In: C. Miehe (ed.) IUTAM Symposium on computational mechanics of solid materials at large strains, Solid mechanics and its applications, vol. 108, pp. 247–258. Springer, Netherlands (2003). doi:10.1007/978-94-017-0297-3_22. URL http://dx.doi.org/10.1007/978-94-017-0297-3_22

  26. Michel JC, Moulinec H, Suquet P (2001) A computational scheme for linear and non-linear composites with arbitrary phase contrast. Int J Numer Methods Eng 52(12):139–160. doi:10.1002/nme.275

    Article  Google Scholar 

  27. Milton GW (2002) The theory of composites. Cambridge University Press, Cambridge

    Book  MATH  Google Scholar 

  28. Monchiet V, Bonnet G (2012) A polarization-based FFT iterative scheme for computing the effective properties of elastic composites with arbitrary contrast. Int J Numer Methods Eng 89(11):1419–1436. doi:10.1002/nme.3295

    Article  MATH  MathSciNet  Google Scholar 

  29. Monchiet, V., Bonnet, G.: Numerical homogenization of nonlinear composites with a polarization-based FFT iterative scheme. Comput Mater Sci 79(0), 276–283 (2013). doi:10.1016/j.commatsci.2013.04.035. URL http://www.sciencedirect.com/science/article/pii/S0927025613002073

  30. Moulinec H, Suquet P (1994) A fast numerical method for computing the linear and nonlinear mechanical properties of composites. Comptes Rendus de l’Académie des Sciences. Série II, Mécanique, Physique, Chimie, Astronomie 318(11):1417–1423

    MATH  Google Scholar 

  31. Moulinec H, Suquet P (1998) A numerical method for computing the overall response of nonlinear composites with complex microstructure. Comput Methods Appl Mech Eng 157(1–2):69–94. doi:10.1016/s0045-7825(97)00218-1

    Article  MATH  MathSciNet  Google Scholar 

  32. Moulinec, H., Suquet, P.: Comparison of FFT-based methods for computing the response of composites with highly contrasted mechanical properties. Physica B: Condens Matter 338(1–4), 58–60 (2003). doi:10.1016/S0921-4526(03)00459-9. URL http://www.sciencedirect.com/science/article/pii/S0921452603004599. Proceedings of the Sixth International Conference on Electrical Transport and Optical Properties of Inhomogeneous Media

  33. Mura T (1987) Micromechanics of defects in solids. Mechanics of elastic and inelastic solids, 2nd edn. Martinus Nijhoff Publishers, Dordrecht

  34. Nemat-Nasser S (1993) Micromechanics: overall properties of heterogeneous materials, North-Holland series in applied mathematics and mechanics. Elsevier Science Publishers B.V, Amsterdam

  35. Ortega JM (1968) The Newton-Kantorovich theorem. Am Math Mon 75(6):658–660. doi:10.2307/2313800

    Article  MATH  Google Scholar 

  36. Ortega JM, Rheinboldt W (1970) Iterative solution of nonlinear equations in several variables. Academic Press, New York

    MATH  Google Scholar 

  37. Paige CC, Saunders MA (1975) Solution of sparse indefinite systems of linear equations. SIAM J Numer Anal 12(4):617–629. doi:10.1137/0712047

    Article  MATH  MathSciNet  Google Scholar 

  38. Schladitz K, Peters S, Reinel-Bitzer D, Wiegmann A, Ohser J (2006) Design of acoustic trim based on geometric modeling and flow simulation for non-woven. Comput Mater Sci 38(1), 56–66. doi:10.1016/j.commatsci.2006.01.018. URL http://www.sciencedirect.com/science/article/pii/S092702560600019X

  39. Schneider M (2014) Convergence of FFT-based homogenization for strongly heterogeneous media. Mathematical methods in the applied sciences n/a(n/a), n/a-n/a. doi:10.1002/mma.3259. URL http://dx.doi.org/10.1002/mma.3259

  40. Šilhavý M (1997) The mechanics and thermodynamics of continuous media. Springer, New York

    MATH  Google Scholar 

  41. Smit RJM, Brekelmans WAM, Meijer HEH (1998) Prediction of the mechanical behavior of nonlinear heterogeneous systems by multi-level finite element modeling. Comput Methods Appl Mech Eng 155(1–2):181–192. doi:10.1016/S0045-7825(97)00139-4. URL http://www.sciencedirect.com/science/article/pii/S0045782597001394

  42. Spahn J, Andrä H, Kabel M, Müller R (2014) A multiscale approach for modeling progressive damage of composite materials using fast Fourier transforms. Compu Methods Appl Mech Eng 268(0):871–883. doi:10.1016/j.cma.2013.10.017. URL http://www.sciencedirect.com/science/article/pii/S0045782513002697

  43. Truesdell C, Noll W (1965). The non-linear field theories of mechanics, encyclopedia of physics, vol. III. Springer URL http://books.google.de/books?id=dp84F_odrBQC

  44. Vinogradov V, Milton GW (2008) An accelerated FFT algorithm for thermoelastic and non-linear composites. Int J Numer Methods Eng 76(11):1678–1695. doi:10.1002/nme.2375

    Article  MATH  MathSciNet  Google Scholar 

  45. Vondrejc J, Zeman J, Marek I (2011) Analysis of a fast Fourier transform based method for modeling of heterogeneous materials. In: Lirkov I, Margenov S, Wasniewski J (eds) LSSC Lecture Notes Computer Science. Springer, Berlin, pp 515–522. doi:10.1007/978-3-642-29843-1_58

  46. Vondřejc J (2013) FFT-based method for homogenization of periodic media: theory and applications. Ph.D. thesis, Department of Mechanics, Faculty of Civil Engineering, Czech Technical University, Czech Republic, Prague (2013).

  47. Zeller R, Dederichs PH (1973) Elastic constants of polycrystals. Physica Status Solidi (b) 55(2):831–842. doi:10.1002/pssb.2220550241

    Article  Google Scholar 

  48. Zeman J, Vondřejc J, Novák J, Marek I (2010) Accelerating a FFT-based solver for numerical homogenization of periodic media by conjugate gradients. J Comput Phys 229:8065–8071. doi:10.1016/j.jcp.2010.07.010. URL http://www.sciencedirect.com/science/article/pii/S0021999110003931

Download references

Acknowledgments

The authors benefited from many fruitful discussions with Heiko Andrä and Andreas Günnel.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Matthias Kabel.

Additional information

This work was conducted while MK enjoyed a sabbatical leave at the Karlruhe Institute of Technology (KIT). MS gratefully acknowledges financial support by the German Research Foundation (DFG), Federal Cluster of Excellence EXC \(1075\) “MERGE Technologies for Multifunctional Lightweight Structures”.

Green’s operator for large deformations in Fourier space

Green’s operator for large deformations in Fourier space

An efficient implementation of the basic scheme (Algorithm 1) needs two main ingredients. A fast implementation of the discrete Fourier transformations as provided for example by the FFTWFootnote 6 library [20] and an explicit formula for the Fourier coefficients \(\hat{\varGamma }^0\) of Green’s operator.

The operators \(\nabla ,\) \(\mathrm{Div}\) and \(G^0\) are understood in Fourier space via

$$\begin{aligned}&\left[ \widehat{\nabla d}(\xi )\right] _{kl} = i \xi _l \hat{d}_k(\xi ),\end{aligned}$$
(54)
$$\begin{aligned}&\left[ \widehat{\mathrm{Div}D}(\xi )\right] _{k} = i \xi _l \hat{D}_{kl}(\xi ),\end{aligned}$$
(55)
$$\begin{aligned}&\left[ \widehat{G^0 d}(\xi )\right] _{k} = \hat{G}^0(\xi )_{kl}\hat{d}_{l}(\xi ), \end{aligned}$$
(56)

for \(k,l = 1,\ldots ,d\), where \(\xi = 2\pi z/L\), \(z \in {\mathbb {Z}}^d\), denotes a wave vector with the short notation \(z/L=(z_1/L_1,\ldots z_d/L_d)\). Therefore, in Fourier space the constitutive Eq. (4) and the equilibrium condition (1) have the form

$$\begin{aligned}&\hat{P}_{kL}(\xi ) = i {\mathbb {C}}^0_{kLmN} \xi _N \hat{u}_m(\xi ) + \hat{\tau }_{kL}(\xi ), \end{aligned}$$
(57)
$$\begin{aligned}&i \hat{P}_{kL}(\xi ) \xi _L = 0, \end{aligned}$$
(58)

Eliminating \(\hat{P}_{kL}\) yields

$$\begin{aligned} {\mathbb {C}}^0_{kLmN} \xi _L \xi _N \hat{u}_m(\xi ) = i \hat{\tau }_{kL}(\xi ) \xi _L. \end{aligned}$$
(59)

Due to the definition of the solution operator \(G^0\) this implies

$$\begin{aligned}&(\hat{G}^0)^{-1}_{km}(\xi ) = {\mathbb {C}}^0_{kLmN} \xi _L \xi _N, \end{aligned}$$
(60)
$$\begin{aligned}&\hat{u}_m(\xi ) = \hat{G}^0_{mk}(\xi ) i \hat{\tau }_{kL}\xi _L. \end{aligned}$$
(61)

Using (9) yields additionally

$$\begin{aligned} \hat{\varGamma }^0_{kLmN}(\xi ) = \xi _L \xi _N \hat{G}^0_{km}(\xi ). \end{aligned}$$
(62)

In the case of an isotropic reference material \({\mathbb {C}}^0\) with Lamé moduli \(\lambda _0\) and \(\mu _0\), i.e. \({\mathbb {C}}^0=\lambda _0 I \otimes I + 2 \mu _0 {\mathbb {I}}^S\) or \({\mathbb {C}}^0_{kLmN} = \lambda _0 \delta _{kL}\delta _{mN} + \mu _0 (\delta _{km}\delta _{LN} + \delta _{kN}\delta _{Lm})\) for \(k,L,m,N = 1,\ldots ,d\), the Fourier coefficients \(\hat{G}^0\) of the solution operator read (cf. [33])

$$\begin{aligned}&(\hat{G}^0)^{-1}_{km}(\xi ) \!=\! \lambda _0 \delta _{kL}\delta _{mN} \xi _L \xi _N \!+\! \mu _0 (\delta _{km}\delta _{LN} \!+\! \delta _{kN}\delta _{Lm})\xi _L \xi _N, \end{aligned}$$
(63)
$$\begin{aligned}&(\hat{G}^0)^{-1}(\xi ) = (\lambda _0 + \mu _0) \xi \otimes \xi + \mu _0 |\xi |^2 I,\end{aligned}$$
(64)
$$\begin{aligned}&\hat{G}^0(\xi ) = \frac{(\lambda _0 + 2 \mu _0) |\xi |^2 I - (\lambda _0 + \mu _0) \xi \otimes \xi }{\mu _0(\lambda _0+2\mu _0)|\xi |^4}. \end{aligned}$$
(65)

By applying (62) we arrive at an explicit formula for the Fourier coefficients of the Green’s operator

$$\begin{aligned}&\hat{\varGamma }^0(\xi )\hat{\tau } = \frac{\hat{\tau }\xi \otimes \xi }{\mu _0 |\xi |^2} -\frac{\lambda _0+\mu _0}{\mu _0\left( \lambda _0+2\mu _0\right) } \frac{\hat{\tau }\xi \cdot \xi }{|\xi |^4} \xi \otimes \xi , \end{aligned}$$
(66)
$$\begin{aligned}&\hat{\varGamma }^0_{kLmN}(\xi ) = \frac{\delta _{km}\xi _L\xi _N}{\mu _0 |\xi |^2} - \frac{\lambda _0+\mu _0}{\mu _0\left( \lambda _0+2\mu _0\right) } \frac{\xi _k \xi _L \xi _m \xi _N}{|\xi |^4}. \end{aligned}$$
(67)

Symmetrizing \(\hat{\varGamma }^0\) gives the Fourier coefficients of Green’s operator for linear elasticity.

If the isotropic reference material \({\mathbb {C}}^0\) with Lamé moduli \(\lambda _0\) and \(\mu _0\) is not symmetrized (cf. Sects. 3.1 and 3.2.5), i.e. \({\mathbb {C}}^0=\lambda _0 I \otimes I + 2 \mu _0 {\mathbb {I}}\) or \({\mathbb {C}}^0_{kLmN} = \lambda _0 \delta _{kL}\delta _{mN} + 2\mu _0 \delta _{km}\delta _{LN}\) for \(k,L,m,N = 1,\ldots ,d\), the Fourier coefficients \(\hat{G}^0\) of the solution operator read

$$\begin{aligned}&(\hat{G}^0)^{-1}_{km}(\xi ) = \lambda _0 \delta _{kL}\delta _{mN} \xi _L \xi _N + 2 \mu _0 \delta _{km}\delta _{LN}\xi _L \xi _N, \end{aligned}$$
(68)
$$\begin{aligned}&(\hat{G}^0)^{-1}(\xi ) = \lambda _0 \xi \otimes \xi + 2\mu _0 |\xi |^2 I,\end{aligned}$$
(69)
$$\begin{aligned}&\hat{G}^0(\xi ) = \frac{(\lambda _0 + 2 \mu _0) |\xi |^2 I - \lambda _0 \xi \otimes \xi }{2\mu _0(\lambda _0+2\mu _0)|\xi |^4}. \end{aligned}$$
(70)

By applying (62) we arrive at an explicit formula for the Fourier coefficients of the Green’s operator

$$\begin{aligned}&\hat{\varGamma }^0(\xi )\hat{\tau } = \frac{\hat{\tau }\xi \otimes \xi }{2\mu _0 |\xi |^2} - \frac{\lambda _0}{2\mu _0\left( \lambda _0+2\mu _0\right) } \frac{\hat{\tau }\xi \cdot \xi }{|\xi |^4} \xi \otimes \xi , \end{aligned}$$
(71)
$$\begin{aligned}&\hat{\varGamma }^0_{kLmN}(\xi ) = \frac{\delta _{km}\xi _L\xi _N}{2\mu _0 |\xi |^2} - \frac{\lambda _0}{2\mu _0\left( \lambda _0+2\mu _0\right) } \frac{\xi _k \xi _L \xi _m \xi _N}{|\xi |^4}. \end{aligned}$$
(72)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Kabel, M., Böhlke, T. & Schneider, M. Efficient fixed point and Newton–Krylov solvers for FFT-based homogenization of elasticity at large deformations. Comput Mech 54, 1497–1514 (2014). https://doi.org/10.1007/s00466-014-1071-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00466-014-1071-8

Keywords

Mathematics Subject Classification

Navigation