×

Multiscale topology optimization using neural network surrogate models. (English) Zbl 1440.74319

Summary: We are concerned with optimization of macroscale elastic structures that are designed utilizing spatially varying microscale metamaterials. The macroscale optimization is accomplished using gradient-based nonlinear topological optimization. But instead of using density as the optimization decision variable, the decision variables are the multiple parameters that define the local microscale metamaterial. This is accomplished using single layer feedforward Gaussian basis function networks as a surrogate models of the elastic response of the microscale metamaterial. The surrogate models are trained using highly resolved continuum finite element simulations of the microscale metamaterials and hence are significantly more accurate than analytical models e.g. classical beam theory. Because the derivative of the surrogate model is important for sensitivity analysis of the macroscale topology optimization, a neural network training procedure based on the Sobolev norm is described. Since the SIMP method is not appropriate for spatially varying lattices, an alternative method is developed to enable creation of void regions. The efficacy of this approach is demonstrated via several examples in which the optimal graded metamaterial outperforms a traditional solid structure.

MSC:

74P15 Topological methods for optimization problems in solid mechanics
74M25 Micromechanics of solids

Software:

L-BFGS; hypre; Ipopt
PDFBibTeX XMLCite
Full Text: DOI Link

References:

[1] Bendsoe, M. P., Optimal shape design as a material distribution problem, Struct. Optim., 1, 193-202 (1989)
[2] Bendsoe, M. P.; Sigmund, O., Topology Optimization Theory, Methods, and Applications (2003), Springer · Zbl 1059.74001
[3] Rozvany, G., Aims, scope, methods, history, and unified terminology of computer aided optimization in structural mechanics, Struct. Multidiscip. Optim., 21, 2, 90-108 (2001)
[4] Reitz, A., Sufficiency of a finite exponent in SIMP (power law) methods, Struct. Multidiscip. Optim., 21, 159-163 (2001)
[5] Bruns, T.; Tortorelli, D., Topology optimization of nonlinear elastic structures and compliant mechanisms, Comput. Methods Appl. Mech. Engrg., 190, 26-27 (2001) · Zbl 1014.74057
[6] Bruns, T.; Tortorelli, D., An element removal and rein- troduction strategy for the topology optimization of structures and compliant mechanisms, Internat. J. Numer. Methods Engrg., 57, 10, 1413-1430 (2003) · Zbl 1062.74589
[7] Lazarov, B.; Sigmund, O., Filters in topology optimization based on helmholtz-type differential equations, Int. J. Numer. Methods Eng., 86, 765-781 (2011) · Zbl 1235.74258
[8] Zowe, J.; Kocvara, M.; Bendsoe, M., Free material optimization via mathematical programming, Math. Program., 79, 445-466 (1997) · Zbl 0886.90145
[9] Kocvara, M.; Stingl, M.; Zowe, J., Free material optimization: Recent progress, Optimization, 57, 79-100 (2008) · Zbl 1132.74035
[10] Allaire, G.; Bonnetier, E.; Francfort, G.; Jouve, F., Shape optimization by the homogenization method, Numer. Math., 76, 27-68 (1997) · Zbl 0889.73051
[11] Watts, S.; Tortorelli, D., A geometric projection method for designing three-dimensional open lattices with inverse homogenization, Internat. J. Numer. Methods Engrg., 112, 15641588 (2017)
[12] Stoer, J.; Bulirsch, R., Introduction to Numerical Analysis (1992), Springer-Verlag · Zbl 0423.65002
[13] Buhmann, M., Radial Basis Functions: Theory and Implementations (2003), Cambridge University Press · Zbl 1038.41001
[14] Schaback, R., Error estimates and condition numbers for radial basis function interpolation, Adv. Comput. Math., 3, 251-264 (1995) · Zbl 0861.65007
[15] Hassoun, M., Fundamentals of Artificial Neural Networks (1995), MIT Press · Zbl 0850.68271
[16] Park, J.; Sandberg, I., Universal approximation using radial-basis-function networks, Neural Comput., 3, 246-257 (1991)
[17] Hornik, K.; Stinchombre, M.; White, H., Multilayer feedforward networks are universal approximators, Neural Netw., 2, 359-366 (1989) · Zbl 1383.92015
[18] Cao, F.; Zhang, Y.; He, Z. R., Interpolation and rates of convergence for a class of neural networks, Appl. Math. Model., 33, 1441-1456 (2009) · Zbl 1168.41302
[19] Cao, F.; Lin, S.; Lu, Z., Approximation capability of interpolation neural networks, Neurocomputing, 74, 457-460 (2010)
[20] Gallant, A. R.; White, H., On learning derivatives of an unknown function mapping with multilayer feedforward networks, Neural Netw., 5, 129-138 (1992)
[21] Llanas, B.; S, L., Hermite interpolation by neural networks, Appl. Math. Comput., 191, 429-439 (2007) · Zbl 1193.41002
[22] Unger, J.; Konke, C., Neural networks as material models within a multiscale approach, Comput. Struct., 87, 1177-1186 (2009)
[23] Furukawa, T.; Hoffman, M., Accurate cyclic plastic analysis using a neural network material model, Eng. Anal. Bound. Elem., 28, 195-204 (2004) · Zbl 1106.74353
[24] Lefik, M.; Boso, D. P.; Screfler, B. A., Artificial Neural Networks in numerical modelling of composites, Comput. Methods Appl. Mech. Engrg., 198, 1785-1804 (2009) · Zbl 1227.74052
[25] Man, H.; Furukawa, T., Neural network constitutive modelling for non-linear characterization of anisotropic materials, Int. J. Numer. Methods Eng., 85, 8, 939957 (2011) · Zbl 1217.74029
[26] Liang, G.; Chandrashekhara, K., Neural network based constitutive model for elastomeric foams, Eng. Struct., 30, 2002-2011 (2008)
[27] Papadopoulos, V.; Soimiris, G.; Giovanis, D.; Papadrakakis, M., A neural network-based surrogate model for carbon nanotubes with geometric nonlinearities, Comput. Methods Appl. Mech. Engrg., 328, 411-430 (2018) · Zbl 1439.74459
[28] MFEM: Modular finite element methods, mfem.org.; MFEM: Modular finite element methods, mfem.org.
[29] Wächter, A.; Biegler, L. T., On the implementation of a primal-dual interior point filter line search algorithm for large-scale nonlinear programming., Math. Program., 106, 1, 5-57 (2006) · Zbl 1134.90542
[30] Messner, M., Optimal lattice-structured materials, J. Mech. Phys. Solids, 96, 162-183 (2016)
[31] Bower, A., Applied Mechanics of Solids (2010), CRC Press
[32] Falgout, R.; Jones, J.; Yang, U., The design and implementation of hypre, a library of parallel high performance preconditioners, (Bruaset, M. A.; Tveito, A., Numerical Solution of Partial Differential Equations on Parallel Computers (2006), Springer), 267-294 · Zbl 1097.65059
[33] Cardaliaguet, P.; Euvrard, G., Approximation of a function and its derivative with a neural network, Neural Netw., 5, 207-220 (1992)
[34] Mai-Duy, N.; Tran-Cong, T., Approximation of function and its derivatives using radial basis function networks, Appl. Math. Model., 27, 197-220 (2003) · Zbl 1024.65012
[35] Lui, D.; Nocedal, J., On the limited memory BFGS method for large scale optimization, Math. Program., 45, 503-528 (1989) · Zbl 0696.90048
[36] Naumann, U., The Art of Differentiating Computer Programs (2012), SIAM · Zbl 1275.65015
[37] Camastra, F., Data dimensionalityestimation methods: a survey, Pattern Recognit., 36, 2945-2954 (2003) · Zbl 1059.68100
[38] Kirby, M., Geometric Data Analysis: An Empirical Approach to Dimensionality Reduction and the Study of Patterns (2001), Wiley · Zbl 1008.68116
[39] Karpinski, M.; Macintyre, A., Polynomial bounds for VC dimension of sigmoidal and general Pfaffin neural networks, J. Comput. Syst. Sci., 54, 169-176 (1997) · Zbl 0869.68088
[40] Bartlett, P.; Maiorov, V.; Meir, R., Almost linear VC-dimension bounds for piecewise polynomial networks, Neural Comput., 10, 8, 2159-2173 (1998)
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.