Isogeometric Topology Optimization Based on Deep Learning

Abstract

Topology optimization plays an important role in a wide range of engineering applications. In this paper, we propose a novel isogeometric topology optimization algorithm based on deep learning. Unlike the other neural network-based methods, the density distributions in the design domain are represented in the B-spline space. In addition, we use relatively novel technologies, U-Net and DenseNet, to form the neural network structure. The 2D and 3D numerical experiments show that the proposed method has an accuracy rate of over 97% for the final optimization results. After training, the new approach can save time greatly for the new topology optimization compared with traditional solid isotropic material with penalization method and IGA method. The approach can also overcome the checkerboard phenomenon.

Keywords

Topology optimization / Deep learning / Isogeometric analysis / B-splines / SIMP

References

[1] Bazilevs Y, Calo VM, Cottrell JA, Evans JA, Hughes TJR, Lipton S, Scott MA, Sederberg TW. Isogeometric analysis using T-splines. Comput. Methods Appl. Mech. Eng.. 2010, 199 5–8 229-263

[2] Bendsøe MP, Kikuchi N. Generating optimal topologies in structural design using a homogenization method. Comput. Methods Appl. Mech. Eng.. 1988, 71 2 197-224

[3] Bendsøe MP, Sigmund O. Material interpolation schemes in topology optimization. Arch. Appl. Mech.. 1999, 69 9–10 635-654

[4] Bendsøe MP, Sigmund O. Topology Optimization: Theory, Method and Applications. 2003 Berlin: Springer

[5] Bengio Y, Simard P, Frasconi P. Learning long-term dependencies with gradient descent is difficult. IEEE Trans. Neural Netw.. 1994, 5 2 157-166

[6] Bourdin B, Chambolle A. Design-dependent loads in topology optimization. ESAIM Control Optim. Calc. Var.. 2003, 9 9 19-48

[7] Cottrell, J.A., Hughes, T.J.R., Bazilevs, Y.: Isogeometric Analysis: Toward Integration of CAD and FEA. Wiley (2009)

[8] Deng H, To AC. Topology optimization based on deep representation learning (DRL) for compliance and stress-constrained design. Comput. Mech.. 2020, 66 449-469

[9] Dunning PD, Kim HA. A new hole insertion method for level set based structural topology optimization. Int. J. Numer. Methods Eng.. 2013, 93 1 118-134

[10] Fu YF, Rolfe B, Chiu LNS, Wang Y, Ghabraie K. Smooth topological design of 3D continuum structures using elemental volume fractions. Comput. Struct.. 2020, 231 106213

[11] Goodfellow, I., Bengio, Y., Courville, A.: Deep Learning. The MIT Press (2015)

[12] Guo X, Zhang W, Zhong W. Doing topology optimization explicitly and geometrically—a new moving morphable components based framework. J. Appl. Mech.. 2014, 81 8 081009

[13] Hassani B, Khanzadi M, Tavakkoli SM. An isogeometrical approach to structural topology optimization by optimality criteria. Struct. Multidiscip. Optim.. 2012, 45 2 223-233

[14] Hornik K, Stinchcombe M, White H. Multilayer feedforward networks are universal approximators. Neural Netw.. 1989, 2 5 359-366

[15] Hou W, Gai Y, Zhu X, Wang X, Zhao C, Xu L, Jiang K, Hu P. Explicit isogeometric topology optimization using moving morphable components. Comput. Methods Appl. Mech. Eng.. 2017, 326 694-712

[16] Huang, G., Liu, Z., Maaten, L.V.D., Weinberger, K.Q.: Densely connected convolutional networks. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), vol. 1, pp. 2261–2269 (2017)

[17] Hughes TJR, Cottrell JA, Bazilevs Y. Isogeometric analysis: CAD, finite elements, NURBS, exact geometry and mesh refinement. Comput. Methods Appl. Mech. Eng.. 2005, 194 39–41 4135-4195

[18] Kang P, Youn SK. Isogeometric topology optimization of shell structures using trimmed NURBS surfaces. Finite Elem. Anal. Des.. 2016, 120 18-40

[19] Krizhevsky A, Sutskever I, Hinton G. ImageNet classification with deep convolutional neural networks. Commun. ACM. 2017, 60 6 84-90

[20] Lecun Y, Bottou L. Gradient-based learning applied to document recognition. Proc. IEEE. 1998, 86 11 2278-2324

[21] Li X, Wei X, Zhang Y. Hybrid non-uniform recursive subdivision with improved convergence rates. Comput. Methods Appl. Mech. Eng.. 2019, 352 606-624

[22] Liu K, Tovar A. An efficient 3D topology optimization code written in Matlab. Struct. Multidiscip. Optim.. 2014, 50 6 1175-1196

[23] Mcculloch WS, Pitts W. A logical calculus of the ideas immanent in nervous activity. Bull. Math. Biophys.. 1943, 5 4 115-133

[24] Nguyen TH, Paulino GH, Song J, Le CH. A computational paradigm for multiresolution topology optimization (MTOP). Struct. Multidiscip. Optim.. 2010, 41 4 525-539

[25] Nie Z, Jiang H, Kara LB. Stress field prediction in cantilevered structures using convolutional neural networks. J. Comput. Inf. Sci. Eng.. 2019, 20 1 011002

[26] Qian X. Topology optimization in B-spline space. Comput. Methods Appl. Mech. Eng.. 2013, 265 15-35

[27] Rawat, S., Shen, M.H.H.: A novel topology design approach using an integrated Deep Learning network architecture. e-Print Archive. arXiv:1808.02334 (2018)

[28] Rawat, S., Shen, M.H.H.: A novel topology optimization approach using conditional deep learning. e-Print Archive. arXiv:1901.04859 (2019)

[29] Ronneberger O, Fischer P, Brox T. U-Net: convolutional networks for biomedical image segmentation. Med. Image Comput. Comput.-Assist. Interv.. 2015, 2015 234-241

[30] Saurabh, B., Harsh, G., Sanket, B., Sagar, P., Levent, K.: 3D topology optimization using convolutional neural networks. e-Print Archive. arXiv:1808.07440v1 (2018)

[31] Seo YD, Kim HJ, Youn SK. Isogeometric topology optimization using trimmed spline surfaces. Comput. Methods Appl. Mech. Eng.. 2010, 199 49–52 3270-3296

[32] Sethian JA, Wiegmann A. Structural boundary design via level set and immersed interface methods. J. Comput. Phys.. 2000, 163 2 489-528

[33] Sosnovik I, Oseledets I. Neural networks for topology optimization. Russ. J. Numer. Anal. Math. Model.. 2019, 34 4 215-223

[34] Wang M, Qian X. Efficient filtering in topology optimization via B-Splines. J. Mech. Des.. 2015, 137 3 031402

[35] Wang MY, Wang X, Guo D. A level set method for structural topology optimization. Comput. Methods Appl. Mech. Eng.. 2003, 192 1–2 227-246

[36] Wang Y, Benson DJ. Isogeometric analysis for parameterized LSM-based structural topology optimization. Comput. Mech.. 2016, 57 1 19-35

[37] Xie Y, Steven GP. A simple evolutionary procedure for structural optimization. Comput. Struct.. 1993, 49 5 885-896

[38] Xie Y, Yang X, Steven GP, Querin O. The theory and application of evolutionary structural optimization method. Eng. Mech.. 1999, 16 6 70-81

[39] Yu Y, Hur T, Jung J. Deep learning for determining a near-optimal topological design without any iteration. Struct. Multidicip. Optim.. 2019, 59 787-799

[40] Zhang W, Li D, Yuan J, Song J, Guo X. A new three-dimensional topology optimization method based on moving morphable components (MMCs). Comput. Mech.. 2016, 59 4 1-19

[41] Zheng R, Kim C. An enhanced topology optimization approach based on the combined MMC and NURBS-curve boundaries. Int. J. Precis. Eng. Manuf.. 2020, 21 2 1529-1538

Funding

National Key R &D Program of China(2020YFB1708900)

National Science of Foundation of China(No.61872328)