Abstract
Mechanical metamaterials are artificial structures that possess exceptional mechanical properties that are not naturally occurring. The complex geometrical and topological features of these metamaterials pose significant challenges to both structure design and manufacturing, despite the recent rapid development of additive manufacturing (AM) techniques. Thus, an effective framework for designing 3D metamaterials with desired mechanical properties, while also ensuring AM manufacturability, is urgently needed. In this paper, an AM manufacturability-aware deep generative model-based design framework is proposed for designing 3D metamaterial units for target properties. To accomplish this, we propose using Variational Autoencoder (VAE) as the feature extractor, which maps the 3D metamaterial geometries to a low-dimensional latent feature space. The latent feature space is concurrently linked to discriminators/regressors to predict manufacturability metrics and mechanical properties. We demonstrate that the proposed design framework is capable of designing high-performance metamaterial units with various user-defined manufacturability metrics. To showcase the effectiveness of the proposed design framework, three design cases with different objective functions are presented, and the final optimal designs are validated by comparing them to state-of-the-art designs or the optimal designs obtained by topology optimization methods.
Similar content being viewed by others
References
Alberdi R, Dingreville R, Robbins J, Walsh T, White BC, Jared B, Boyce BL (2020) Multi-morphology lattices lead to improved plastic energy absorption. Mater Des 194:108883
Ashburner J, Friston KJ (2000) Voxel-based morphometry—the methods. Neuroimage 11(6):805–821
Atzeni E, Salmi A (2015) Study on unsupported overhangs of AlSi10Mg parts processed by Direct Metal Laser Sintering (DMLS). J Manuf Process 20:500–506
Babaee S, Shim J, Weaver JC, Chen ER, Patel N, Bertoldi K (2013) 3D soft metamaterials with negative Poisson’s ratio. Adv Mater 25(36):5044–5049
Bastek J-H, Kumar S, Telgen B, Glaesener RN, Kochmann DM (2022) Inverting the structure–property map of truss metamaterials by deep learning. Proc Natl Acad Sci 119(1):e2111505119
Bertrand P, Bayle F, Combe C, Gœuriot P, Smurov I (2007) Ceramic components manufacturing by selective laser sintering. Appl Surf Sci 254(4):989–992
Bintara RD, Aminnudin A, Prasetiyo D, Arbianto FR (2019) The characteristic of overhang object to material usage on FDM 3D printing technology. J Mech Eng Sci Technol (JMEST) 3(1):35–41
Bodaghi M, Damanpack A, Hu G, Liao W (2017) Large deformations of soft metamaterials fabricated by 3D printing. Mater Des 131:81–91
Booth JW, Alperovich J, Chawla P, Ma J, Reid TN, Ramani K (2017) The design for additive manufacturing worksheet. J Mech Des 139(10):100904
Boschetto A, Bottini L (2014) Accuracy prediction in fused deposition modeling. Int J Adv Manuf Technol 73(5):913–928
Bostanabad R, Chan Y-C, Wang L, Zhu P, Chen W (2019) Globally approximate gaussian processes for big data with application to data-driven metamaterials design. J Mech Des 141(11):111402
Cang R, Xu Y, Chen S, Liu Y, Jiao Y, Yi Ren M (2017) Microstructure representation and reconstruction of heterogeneous materials via deep belief network for computational material design. J Mech Des. https://doi.org/10.1115/1.4036649
Cao S, Qiu Y, Wei X-F, Zhang H-H (2015) Experimental and theoretical investigation on ultra-thin powder layering in three dimensional printing (3DP) by a novel double-smoothing mechanism. J Mater Process Technol 220:231–242
Chan Y-C, Ahmed F, Wang L, Chen W (2021) METASET: exploring shape and property spaces for data-driven metamaterials design. J Mech Des 143(3):031707
Chen H, Chan CT (2007) Acoustic cloaking in three dimensions using acoustic metamaterials. Appl Phys Lett 91(18):183518
Chen D, Skouras M, Zhu B, Matusik W (2018) Computational discovery of extremal microstructure families. Sci Adv 4(1):eaao7005
Claeys C, de Melo Filho NGR, Van Belle L, Deckers E, Desmet W (2017) Design and validation of metamaterials for multiple structural stop bands in waveguides. Extreme Mech Lett 12:7–22
Evans KE (1991) Auxetic polymers: a new range of materials. Endeavour 15(4):170–174
Feng S, Kamat AM, Sabooni S, Pei Y (2021) Experimental and numerical investigation of the origin of surface roughness in laser powder bed fused overhang regions. Virtual Phys Prototyping 16(sup1):S66–S84
Gao J, Li H, Gao L, Xiao M (2018) Topological shape optimization of 3D micro-structured materials using energy-based homogenization method. Adv Eng Softw 116:89–102
Garland AP, Adstedt KM, Casias ZJ, White BC, Mook WM, Kaehr B, Jared BH, Lester BT, Leathe NS, Schwaller E (2020) Coulombic friction in metamaterials to dissipate mechanical energy. Extreme Mech Lett 40:100847
Grasso M, Colosimo BM (2017) Process defects and in situ monitoring methods in metal powder bed fusion: a review. Meas Sci Technol 28(4):044005
Grima JN, Winczewski S, Mizzi L, Grech MC, Cauchi R, Gatt R, Attard D, Wojciechowski KW, Rybicki J (2015) Tailoring graphene to achieve negative Poisson’s ratio properties. Adv Mater 27(8):1455–1459
Gu GX, Chen C-T, Richmond DJ, Buehler MJ (2018) Bioinspired hierarchical composite design using machine learning: simulation, additive manufacturing, and experiment. Mater Horiz 5(5):939–945
Gurbuz C, Kronowetter F, Dietz C, Eser M, Schmid J, Marburg S (2021) Generative adversarial networks for the design of acoustic metamaterials. J Acoust Soc Am 149(2):1162
He L, Song X (2018) Supportability of a high-yield-stress slurry in a new stereolithography-based ceramic fabrication process. Jom 70:407–412
Hunter LW, Brackett D, Brierley N, Yang J, Attallah MM (2020) Assessment of trapped powder removal and inspection strategies for powder bed fusion techniques. Int J Adv Manuf Technol 106:4521–4532
Hutmacher DW, Schantz T, Zein I, Ng KW, Teoh SH, Tan KC (2001) Mechanical properties and cell cultural response of polycaprolactone scaffolds designed and fabricated via fused deposition modeling. J Biomed Mater Res 55(2):203–216
Järvinen J-P, Matilainen V, Li X, Piili H, Salminen A, Mäkelä I, Nyrhilä O (2014) Characterization of effect of support structures in laser additive manufacturing of stainless steel. Phys Procedia 56:72–81
Jha D, Singh S, Al-Bahrani R, Liao W-K, Choudhary A, De Graef M, Agrawal A (2018) Extracting grain orientations from ebsd patterns of polycrystalline materials using convolutional neural networks. Microsc Microanal 24(5):497–502
Jiang J, Xu X, Stringer J (2018) Support structures for additive manufacturing: a review. J Manuf Mater Process 2(4):64
Jin Y, Du J, He Y (2017) Optimization of process planning for reducing material consumption in additive manufacturing. J Manuf Syst 44:65–78
Kerbrat O, Mognol P, Hascoët J-Y (2011) A new DFM approach to combine machining and additive manufacturing. Comput Ind 62(7):684–692
Kingma DP, Welling M (2014) Stochastic gradient VB and the variational auto-encoder. In: Second international conference on learning representations, ICLR
Kruth J-P, Vandenbroucke B, Van Vaerenbergh J, Mercelis P (2005) Benchmarking of different SLS/SLM processes as rapid manufacturing techniques. In: Proceedings of the International conference polymers & moulds innovations PMI 2005
Kumar S, Tan S, Zheng L, Kochmann DM (2020) Inverse-designed spinodoid metamaterials. Npj Comput Mater 6(1):73
Lam TF, Xiong Y, Dharmawan AG, Foong S, Soh GS (2020) Adaptive process control implementation of wire arc additive manufacturing for thin-walled components with overhang features. Int J Adv Manuf Technol 108:1061–1071
Lee J, Lee K (2017) Block-based inner support structure generation algorithm for 3D printing using fused deposition modeling. Int J Adv Manuf Technol 89(5):2151–2163
Lei M, Hong W, Zhao Z, Hamel C, Chen M, Lu H, Qi HJ (2019) 3D printing of auxetic metamaterials with digitally reprogrammable shape. ACS Appl Mater Interfaces 11(25):22768–22776
Li M, Wang Z (2020) Reliability-based multifidelity optimization using adaptive hybrid learning. ASCE-ASME J Risk Uncertain Eng Syst B 6(2):021005
Li S, Hassanin H, Attallah MM, Adkins NJ, Essa K (2016) The development of TiNi-based negative Poisson’s ratio structure using selective laser melting. Acta Mater 105:75–83
Li C, Liu J, Fang X, Guo Y (2017) Efficient predictive model of part distortion and residual stress in selective laser melting. Addit Manuf 17:157–168
Li H, Luo Z, Gao L, Qin Q (2018) Topology optimization for concurrent design of structures with multi-patch microstructures by level sets. Comput Methods Appl Mech Eng 331:536–561
Liu R, Agrawal A, Liao W, Choudhary A, De Graef M (2016) Materials discovery: understanding polycrystals from large-scale electron patterns. In: 2016 IEEE International conference on big data (Big Data). IEEE
Liu Z, Xu H, Zhu P (2020) An adaptive multi-fidelity approach for design optimization of mesostructure-structure systems. Struct Multidisc Optim 62:375–386
Lores A, Azurmendi N, Agote I, Zuza E (2019) A review on recent developments in binder jetting metal additive manufacturing: materials and process characteristics. Powder Metall 62(5):267–296
Meyer PP, Bonatti C, Tancogne-Dejean T, Mohr D (2022) Graph-based metamaterials: deep learning of structure-property relations. Mater Des 223:111175
Mirabolghasemi A, Akbarzadeh A, Rodrigue D, Therriault D (2019) Thermal conductivity of architected cellular metamaterials. Acta Mater 174:61–80
Moylan S, Slotwinski J, Cooke A, Jurrens K, Donmez MA (2023) Proposal for a standardized test artifact for additive manufacturing machines and processes. In: 2012 International solid freeform fabrication symposium. University of Texas at Austin
Oliveira JP, LaLonde A, Ma J (2020) Processing parameters in laser powder bed fusion metal additive manufacturing. Mater Des 193:108762
Paszke A, Gross S, Massa F, Lerer A, Bradbury J, Chanan G, Killeen T, Lin Z, Gimelshein N, Antiga L (2019) Pytorch: an imperative style, high-performance deep learning library. In: Advances in neural information processing systems, vol 32
Pham M-S, Liu C, Todd I, Lertthanasarn J (2019) Damage-tolerant architected materials inspired by crystal microstructure. Nature 565(7739):305–311
Qi Z, Zhang N, Liu Y, Chen W (2019) Prediction of mechanical properties of carbon fiber based on cross-scale FEM and machine learning. Compos Struct 212:199–206
Qian X (2017) Undercut and overhang angle control in topology optimization: a density gradient based integral approach. Int J Numer Methods Eng 111(3):247–272
Qian J, Cheng Y, Zhang A, Zhou Q, Zhang J (2021) Optimization design of metamaterial vibration isolator with honeycomb structure based on multi-fidelity surrogate model. Struct Multidisc Optim 64:423–439
Reddy KSN, Ferguson I, Frecker M, Simpson TW, Dickman CJ (2016) Topology optimization software for additive manufacturing: a review of current capabilities and a real-world example. In: International design engineering technical conferences and computers and information in engineering conference. 2016. American Society of Mechanical Engineers
Shi Y, Zhang Y, Baek S, De Backer W, Harik R (2018) Manufacturability analysis for additive manufacturing using a novel feature recognition technique. Comput-Aided Des Appl 15(6):941–952
Strano G, Hao L, Everson R, Evans K (2013) A new approach to the design and optimisation of support structures in additive manufacturing. Int J Adv Manuf Technol 66:1247–1254
Takezawa A, Koizumi Y, Kobashi M (2017) High-stiffness and strength porous maraging steel via topology optimization and selective laser melting. Addit Manuf 18:194–202
Tao W, Leu MC (2016) Design of lattice structure for additive manufacturing. In: 2016 International symposium on flexible automation (ISFA). IEEE
Tedia S, Williams CB (2016) Manufacturability analysis tool for additive manufacturing using voxel-based geometric modeling. In: 2016 International solid freeform fabrication symposium. University of Texas at Austin
Telea A, Jalba A (2011) Voxel-based assessment of printability of 3D shapes. In: Mathematical morphology and its applications to image and signal processing: 10th international symposium, ISMM 2011, Verbania-Intra, Italy, July 6–8, 2011. Proceedings 10. Springer
Thompson MK, Moroni G, Vaneker T, Fadel G, Campbell RI, Gibson I, Bernard A, Schulz J, Graf P, Ahuja B (2016) Design for additive manufacturing: trends, opportunities, considerations, and constraints. CIRP Ann 65(2):737–760
Thore C-J, Grundström HA, Torstenfelt B, Klarbring A (2019) Penalty regulation of overhang in topology optimization for additive manufacturing. Struct Multidisc Optim 60:59–67
Tuncer N, Bose A (2020) Solid-state metal additive manufacturing: a review. Jom 72(9):3090–3111
Vogiatzis P, Chen S, Wang X, Li T, Wang L (2017) Topology optimization of multi-material negative Poisson’s ratio metamaterials using a reconciled level set method. Comput Aided Des 83:15–32
Vogiatzis P, Chen S, Gu XD, Chuang C-H, Xu H, Lei N (2018) Multi-material topology optimization of structures infilled with conformal metamaterials. In: International design engineering technical conferences and computers and information in engineering conference. American Society of Mechanical Engineers
Walton D, Moztarzadeh H (2017) Design and development of an additive manufactured component by topology optimisation. Procedia Cirp 60:205–210
Wang Y, Luo Z, Zhang N, Kang Z (2014) Topological shape optimization of microstructural metamaterials using a level set method. Comput Mater Sci 87:178–186
Wang C, Xu B, Meng Q, Rong J, Zhao Y (2020a) Numerical performance of Poisson method for restricting enclosed voids in topology optimization. Comput Struct 239:106337
Wang L, Chan Y-C, Liu Z, Zhu P, Chen W (2020b) Data-driven metamaterial design with Laplace-Beltrami spectrum as “shape-DNA.” Struct Multidisc Optim 61:2613–2628
Wang Z, Xian W, Baccouche MR, Lanzerath H, Li Y, Xu H (2021) A Gaussian mixture variational autoencoder-based approach for designing phononic bandgap metamaterials. In: International design engineering technical conferences and computers and information in engineering conference. American Society of Mechanical Engineers
Wang Z, Xian W, Baccouche MR, Lanzerath H, Li Y, Xu H (2022a) Design of phononic bandgap metamaterials based on Gaussian mixture beta variational autoencoder and iterative model updating. J Mech Des 144(4):041705
Wang Z, Zhuang R, Xian W, Tian J, Li Y, Chen S, Xu H (2022b) Phononic metamaterial design via transfer learning-based topology optimization framework. In: International design engineering technical conferences and computers and information in engineering conference. American Society of Mechanical Engineers
Watts S, Arrighi W, Kudo J, Tortorelli DA, White DA (2019) Simple, accurate surrogate models of the elastic response of three-dimensional open truss micro-architectures with applications to multiscale topology design. Struct Multidisc Optim 60(5):1887–1920
Wei X, Geng G, Zhang Y (2016) Steady and low consuming supporting for fused deposition modeling. Zidonghua Xuebao/acta Automatica Sinica 42(1):98–106
Wei C, Chueh Y-H, Zhang X, Huang Y, Chen Q, Li L (2019) Easy-to-remove composite support material and procedure in additive manufacturing of metallic components using multiple material laser-based powder bed fusion. J Manuf Sci Eng. https://doi.org/10.1115/1.4043536
White T (2016) Sampling generative networks. arXiv preprint arXiv:1609.04468
Wilt JK, Yang C, Gu GX (2020) Accelerating auxetic metamaterial design with deep learning. Adv Eng Mater 22(5):1901266
Xia Z, Zhang Y, Ellyin F (2003) A unified periodical boundary conditions for representative volume elements of composites and applications. Int J Solids Struct 40(8):1907–1921
Xiong Y, Yao S, Zhao Z-L, Xie YM (2020) A new approach to eliminating enclosed voids in topology optimization for additive manufacturing. Addit Manuf 32:101006
Xu H, Liu Z (2019) Control variate multifidelity estimators for the variance and sensitivity analysis of mesostructure–structure systems. ASCE-ASME J Risk Uncertain Eng Syst B 5(2):020907
Xu L, Hoffman N, Wang Z, Xu H (2022) Harnessing structural stochasticity in the computational discovery and design of microstructures. Mater Des 223:111223
Yadollahi A, Shamsaei N (2017) Additive manufacturing of fatigue resistant materials: challenges and opportunities. Int J Fatigue 98:14–31
Yang KV, Rometsch P, Jarvis T, Rao J, Cao S, Davies C, Wu X (2018) Porosity formation mechanisms and fatigue response in Al-Si-Mg alloys made by selective laser melting. Mater Sci Eng A 712:166–174
Yasuda H, Yang J (2015) Reentrant origami-based metamaterials with negative Poisson’s ratio and bistability. Phys Rev Lett 114(18):185502
Zein I, Hutmacher DW, Tan KC, Teoh SH (2002) Fused deposition modeling of novel scaffold architectures for tissue engineering applications. Biomaterials 23(4):1169–1185
Zeng Q, Zhao Z, Lei H, Wang P (2022) A deep learning approach for reverse design of gradient mechanical metamaterials. Int J Mech Sci 240:107920
Zeng Q, Zhao Z, Lei H, Wang P (2023) A deep learning approach for inverse design of gradient mechanical metamaterials. Int J Mech Sci 240:107920
Zhang G, Khandelwal K (2020) Topology optimization of dissipative metamaterials at finite strains based on nonlinear homogenization. Struct Multidisc Optim 62:1419–1455
Zhang Y, Yang S, Zhao YF (2020) Manufacturability analysis of metal laser-based powder bed fusion additive manufacturing—a survey. Int J Adv Manuf Technol 110(1):57–78
Zhang K, Cheng G, Wang Y (2022a) Structural topology optimization subject to overhang angle constraint with overhang length relaxation in additive manufacturing. Sci China Technol Sci 65(6):1213–1231
Zhang W, Wang C, Zhou L, Gao T (2022b) Three-dimensional topology optimization considering overhang constraints with B-spline parameterization. Comput Struct 269:106823
Zheng X, Lee H, Weisgraber TH, Shusteff M, DeOtte J, Duoss EB, Kuntz JD, Biener MM, Ge Q, Jackson JA (2014) Ultralight, ultrastiff mechanical metamaterials. Science 344(6190):1373–1377
Zhou L, Zhang W (2019) Topology optimization method with elimination of enclosed voids. Struct Multidisc Optim 60:117–136
Acknowledgements
This material is based upon works supported by the National Science Foundation (NSF) under Grant No. IIP-#1822157 (Phase I IUCRC at University of Connecticut: Center for Science of Heterogeneous Additive Printing of 3D Materials, SHAP3D) and NSF Grant No. CMMI-2142290. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation or the sponsors.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflicts of interest.
Replication of results
Upon acceptance, the authors will provide access to the data and model required to reproduce the results upon reasonable request.
Additional information
Responsible Editor: Palaniappan Ramu
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
1.1 Details of the microstructure family template-based method
This method is used to generate the first group of microstructure samples in the database. This method is a four-step process:
Step 1: Define rectangular bars in the continuous cubic spatial domain \(\left[ {0,l} \right]^{3}\), where \(l\) represents the length of the spatial domain. The rectangular bars are defined by the coordinates of the pair of reference points \([loc_{i} , loc_{j} \left] = \right[\left( {locx_{i} , locy_{i} ,locz_{i} { }} \right), \left( {locx_{j} , locy_{j} ,locz_{j} } \right)]\) and the side length of the square cross-section of the bar (\(h\)). Therefore, each bar corresponds to 7 design variables. To avoid overly complicated structures, we only create 1 or 2 bars in this step.
Step 2: Voxelate the spatial domain \(\left[ {0,l} \right]^{3}\). In this work, we use \(l = 48\), resulting in a \(48 \times 48 \times 48\) voxel domain. The bars created in Step 1 are also voxelated. The voxels of the rectangular bars are referred as “original bar voxels” in the following steps.
Step 3: Map the original bar voxels by the midpoint of the grid along each axis sequentially, following the relationship \(loc^{\prime} = l - loc\). Through this mirroring process, the original bar voxels on one side of each axis’s midpoint are reflected to the opposite side, creating a symmetric arrangement in all three directions.
Step 4: Record the indices of each voxel to enforce a hierarchical relationship between \(locx_{i} ,{ }locy_{i} ,and locz_{i}\) coordinates. This step map the bar voxels \(\left( {locx_{i} ,{ }locy_{i} ,locz_{i} { }} \right)\) obtained by previous step to new locations \(\left( {locx_{i} {^{\prime}},locy_{i} {^{\prime}},locz_{i} {^{\prime}}} \right)\) by permuting their coordinates while adhering to the relationship \(locx_{i}{\prime} > locy_{i}{\prime} > locz_{i} {^{\prime}}\). This step enforces a cubic symmetry in the structure.
After completing the abovementioned steps, the rectangular bars defined within the spatial domain \(\left[ {0,l} \right]^{3}\) are mirrored to form a voxelated, cubic, symmetric metamaterial unit. For the metamaterial units generated by 1 bar, there are 7 design variables; for the metamaterial units generated by 2 bars, there are in total 14 design variables. Latin hyper sampling (LHS) is used to assign values to the parameters of each bar. In our LHS table, the values regarding the coordinates \((locx_{i} , locy_{i} ,locz_{i} , locx_{j} , locy_{j} ,locz_{j}\)) are generated in the range of \(\left[ {0,l} \right]\) and the parameter of side length \(h\) is generated in the range of \(\left( {0,l} \right]\) (Fig. 12).
1.2 Deep generative model that trains VAE and supervised learning model separately
Here, we present an alternative modeling approach that trains the VAE and supervised learning models separately. The VAE is firstly trained to obtain a low-dimensional latent feature space that solely depends on the geometries of the 3D metamaterial units. Then, we use supervised learning models to establish the relationship between the latent feature variables and the mechanical properties and manufacturability metric values. The accuracy of this model is evaluated based on the reconstruction accuracy and the accuracies of the supervised learning models. The metrics are introduced in Eqs. 10–12. The results are presented in Table 4. Compared to the proposed integrated manufacturability-aware model (Table 1), the strategy of training the VAE and the supervised learning models separately show slightly higher accuracy in structural reconstruction but lower accuracies in predicting properties and manufacturability metric values. The lower prediction accuracies may lead to infeasible or low-performance designs found by optimization.
1.3 Convergence test to determine the dimensions of latent feature space
See Table 5.
1.4 Hyperparameters of the manufacturability-aware deep generative model
See Table 6
1.5 Multi-objective design optimization: manufacturability metric values and true manufacturability of the optimal design candidates
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Wang, Z., Xu, H. Manufacturability-aware deep generative design of 3D metamaterial units for additive manufacturing. Struct Multidisc Optim 67, 22 (2024). https://doi.org/10.1007/s00158-023-03732-4
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s00158-023-03732-4