Abstract
In sheet metal forming, persistent challenges such as failure, wrinkling, and springback necessitate innovative solutions. This study elucidates a novel methodology tailored to mitigate these prevalent defects. Our primary aim was to bolster formability while concurrently diminishing defects throughout the forming cycle. To realize this objective, we deployed a segmented and variable blank holder force (S-VBHF) trajectory, facilitating precise adjustments to the blank holder force (BHF). In our quest to optimize process parameters, encompassing the S-VBHF, friction coefficient, and drawbead restraining force (DBRF), we amalgamated deep neural network, genetic algorithm, and Monte Carlo simulation techniques, collectively denoted as DNN-GA-MCS. The forming limit diagram (FLD) served as our rigorous evaluative framework, enabling a comprehensive assessment of sheet failure dynamics during the forming phase. Our proposed methodology underwent stringent validation via numerical simulations, with the cylindrical cup from NUMISHEET 2011 (BM1) providing the benchmark. Empirical outcomes underscored substantial enhancements in formed sheet quality: marked reductions of 8.33% in failure, 10.81% in wrinkling, and 5.88% in springback. In summation, our advanced methodology manifests potential in refining sheet metal forming processes, emphasizing its efficacy in substantially curtailing defects.
Similar content being viewed by others
Data availability
Not applicable.
References
Hosford WF, Caddell RM (2011) Metal forming: mechanics and metallurgy. Cambridge University Press
Irthiea IK, Green G (2017) Evaluation of micro deep drawing technique using soft die-simulation and experiments. Int J Adv Manuf Technol 89:2363–2374
Atul ST, Babu ML (2019) A review on effect of thinning, wrinkling and spring-back on deep drawing process. Proc Inst Mech Eng, Part B: J Eng Manuf 233(4):1011–1036
Groover MP (2020) Fundamentals of modern manufacturing: materials, processes, and systems. John Wiley & Sons
Padmanabhan R, Oliveira MC, Alves JL, Menezes LF (2007) Influence of process parameters on the deep drawing of stainless steel. Finite Elem Anal Des 43(14):1062–1067
Raju S, Ganesan G, Karthikeyan R (2010) Influence of variables in deep drawing of AA 6061 sheet. Trans Nonferrous Metals Soc China 20(10):1856–1862
Reddy ACS, Rajesham S, Reddy PR, Kumar TP, Goverdhan J (2015) An experimental study on effect of process parameters in deep drawing using Taguchi technique. Int J Eng Sci Technol 7(1):21–32
Dilmec M, Arap M (2016) Effect of geometrical and process parameters on coefficient of friction in deep drawing process at the flange and the radius regions. Int J Adv Manuf Technol 86:747–759
Kardan M, Parvizi A, Askari A (2018) Influence of process parameters on residual stresses in deep-drawing process with FEM and experimental evaluations. J Braz Soc Mech Sci Eng 40:1–12
Zhang W, Shivpuri R (2009) Probabilistic design of aluminum sheet drawing for reduced risk of wrinkling and fracture. Reliab Eng Syst Saf 94:152–161
Gunnarsson L, Schedin E (2001) Improving the properties of exterior body panels in automobiles using variable blank holder force. J Mater Proc Technol 114(2):168–173
Manabe KI, Soeda K, Shibata A (2021) Effects of variable punch speed and blank holder force in warm superplastic deep drawing process. Metals 11(3):493
Feng Y, Hong Z, Gao Y, Lu R, Wang Y, Tan J (2019) Optimization of variable blank holder force in deep drawing based on support vector regression model and trust region. Int J Adv Manuf Technol 105:4265–4278
Kitayama S, Koyama H, Kawamoto K, Miyasaka T, Yamamichi K, Noda T (2017) Optimization of blank shape and segmented variable blank holder force trajectories in deep drawing using sequential approximate optimization. Int J Adv Manuf Technol 91:1809–1821
Wang WR, Chen GL, Lin ZQ, Li SH (2007) Determination of optimal blank holder force trajectories for segmented binders of step rectangle box using PID closed-loop FEM simulation. Int J Adv Manuf Technol 32:1074–1082
Kitayma S, Srirat J, Arakawa M, Yamazaki K (2013) Sequential approximate multi-objective optimization using radial basis function network. Struct Multidiscip Optim 48(3):501–515
Gao Y, Li H, Zhao D, Wang M, Fan X (2023) Advances in friction of aluminium alloy deep drawing. Friction 1–32. https://doi.org/10.1007/s40544-023-0761-7
Shivpuri R, Zhang W (2009) Robust design of spatially distributed friction for reduced wrinkling and thinning failure in sheet drawing. Mater Des 30(6):2043–2055
Alavala CR (2016) Effect of temperature, strain rate and coefficient of friction on deep drawing process of 6061 aluminum alloy. Int J Mech Eng 5(6):11–24
Folle LF, Schaeffer L (2019) Effect of surface roughness and lubrication on the friction coefficient in deep drawing processes of aluminum alloy aa1100 with fem analysis 1. Matéria (Rio de Janeiro) 24. https://doi.org/10.1590/S1517-707620190001.0635
Wei L, Yuying Y (2008) Multi-objective optimization of sheet metal forming process using Pareto-based genetic algorithm. J Mater Process Technol 208(1–3):499–506
Kitayama S, Shimizu K, Kawamoto K (2021) Numerical optimization of blank shape and sloped variable blank holder force trajectory for an automotive part. J Adv Mech Des, Syst Manuf 15(3):JAMDSM0027–JAMDSM0027
Abbasi M, Bagheri B, Abdollahzadeh A, Moghaddam AO (2021) A different attempt to improve the formability of aluminum tailor welded blanks (TWB) produced by the FSW. IntJ Mater Form 14:1189–1208
Kitayama S, Natsume S, Yamazaki K, Han J, Uchida H (2016) Numerical investigation and optimization of pulsating and variable blank holder force for identification of formability window for deep drawing of cylindrical cup. Int J Adv Manuf Technol 82:583–593. https://doi.org/10.1007/s00170-015-7385-7
Xie Y, Tang W, Zhang F, Pan BB, Yue Y, Feng M (2019) Optimization of variable blank holder force based on a sharing Niching RBF neural network and an improved NSGA II Alg
Bagheri B, Abbasi M, Hamzeloo R (2021) Comparison of different welding methods on mechanical properties and formability behaviors of tailor welded blanks (TWB) made from AA6061 alloys. Proc Inst Mech Eng C J Mech Eng Sci 235(12):2225–2237
Abbasi M, Hamzeloo SR, Ketabchi M, Shafaat MA, Bagheri B (2014) Analytical method for prediction of weld line movement during stretch forming of tailor-welded blanks. Int J Adv Manuf Technol 73:999–1009
Abbasi M, Bagheri B, Ketabchi M, Haghshenas DF (2012) Application of response surface methodology to drive GTN model parameters and determine the FLD of tailor welded blank. Comput Mater Sci 53(1):368–376
Choudhari CS, Khasbage SS (2021) Experimental investigation of forming parameters for square cup deep drawing process. Mater Today: Proc 44:4261–4267
Wifi AS, Abdelmaguid TF, & El-Ghandour AI (2007) A review of the optimization techniques applied to the deep drawing process. In Proceedings of the 37th international conference on computers and industrial engineering 97–107.
Chinchanikar S, & Kolte Y (2022) A review on experimental and numerical studies on micro deep drawing considering size effects and key process parameters. Australian Journal of Mechanical Engineering, 1–14.
Hino R, Yoshida F, Toropov VV (2006) Optimum blank design for sheet metal forming based on the interaction of high-and lowfidelity FE models. Arch Appl Mech 75(10):679–691. https://doi.org/10.1007/s00419-006-0047-3
Liu Y, Chen W, Ding L, Wang X (2013) Response surface methodology based on support vector regression for polygon blank shape optimization design. Int J Adv Manuf Technol 66:1397–1405. https://doi.org/10.1007/s00170-012-4417-4
Feng Y, Lu R, Gao Y, Zheng H, Wang Y, Mo W (2018) Multi-objective optimization of VBHF in sheet metal deep-drawing using Kriging, MOABC, and set pair analysis. Int J Adv Manuf Technol 96:3127–3138
Gantar G, Kuzman K (2005) Optimization of stamping processes aiming at maximal process stability. J Mater Proc Tech 167:237–243
Zhang, W., Li, Y., & Li, J. (2011). Reliability-based process design and optimization. In The Proceedings of 2011 9th International Conference on Reliability, Maintainability and Safety (pp. 1096–1100). IEEE.
Marretta L, Di Lorenzo R (2010) Influence of material properties variability on springback and thinning in sheet stamping processes: a stochastic analysis. Int J Adv Manuf Technol 51:117–134
Marretta L, Ingarao G, Di Lorenzo R (2010) Design of sheet stamping operations to control springback and thinning: a multi-objective stochastic optimization approach. Int J Mech Sci 52(7):914–927. https://doi.org/10.1016/j.ijmecsci.2010.03.008
Li H, Wang Q, He F, & Zheng Y (2019) An intelligent prediction strategy of the maximum thinning rate of cylindrical part with flange during sheet metal drawing process. In 2019 Chinese Control And Decision Conference (CCDC) (pp. 2876–2881). IEEE.
Tran MT, Shan Z, Lee HW, Kim DK (2021) Earing reduction by varying blank holding force in deep drawing with deep neural network. Metals 11(3):395
Kitayama S, Koyama H, Kawamoto K, Noda T, Yamamichi K, Miyasaka T (2017) Numerical and experimental case study on simultaneous optimization of blank shape and variable blank holder force trajectory in deep drawing. Struct Multidiscip Optim 55:347–359. https://doi.org/10.1007/s00158-016-1484-4
Kitayama S, Yamada S (2017) Simultaneous optimization of blank shape and variable blank holder force of front side member manufacturing by deep drawing. Int J Adv Manuf Technol 91:1381–1390. https://doi.org/10.1007/s00170-016-9837-0
Kitayama S, Yokoyama M, Kawamoto K, Noda T, Miyasaka T, Echigo Y (2018) Practical approach of simultaneous optimization of variable blank holder force and variable slide velocity trajectory in sheet metal forming. Int J Adv Manuf Technol 98:2693–2703
Kitayama, S., Ishizuki, R., Yokoyaka, M., Kawamoto, K., Natsume, S., Adachi, K., ... & Ohtani, T. (2019). Numerical optimization of variable blank holder force trajectory and blank shape for twist springback reduction using sequential approximate optimization. The International Journal of Advanced Manufacturing Technology, 103, 63–75.
Zhai, J., Zhang, Q., Zhang, Z., Sun, Y., Qin, X., & Chen, X. (2019, May). Optimal design for springback of automotive panel forming with kriging model. In 2019 International Conference on Advances in Construction Machinery and Vehicle Engineering (ICACMVE) (pp. 23–28). IEEE.
Barlat F, Aretz H, Yoon JW, Karabin M, Brem JC, Dick R (2005) Linear transfomation-based anisotropic yield functions. Int J Plast 21(5):1009–1039
Miettinen KM (1998) Nonlinear multiobjective optimization. Kluwer Academic Publishers
Taguchi, G., & Konishi, S. (1987). Taguchi methods: orthogonal arrays and linear graphs; tools for quality engineering. ASI press.
Yan Z, Zhou H, Zhang X, Liu J, Wang C, Lu X, Sui X (2022) Interactive effect between WS2 films with different structures and space oils for improvement of tribological performance. Tribology International 170
Adebogun A, Hudson R, Matthews A, Withers PJ (2020) Industrial gear oils: influence of bulk oil temperature and contact pressure on tribological performance and subsurface changes. Tribol Lett 68:1–20
Birleanu C, Pustan M, Pop G, Cioaza M, Popa F, Lazarescu L, Contiu G (2022) Experimental investigation of the tribological behaviors of carbon fiber reinforced polymer composites under boundary lubrication. Polymers 14(18):3716
Hamilton, A., Tran, T., Mckay, M. B., Quiring, B., & Vassilevski, P. S. (2019). Dnn approximation of nonlinear finite element equations (No. LLNL-TR-791918). Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States).
Jamli MR, Farid NM (2019) The sustainability of neural network applications within finite element analysis in sheet metal forming: a review. Measurement 138:446–460. https://doi.org/10.1016/j.measurement.2019.02.034
Kingma, D. P., & Ba, J. (2014). Adam: a method for stochastic optimization. arXiv preprint arXiv:1412.6980 .
Reddi S J, Kale S, Kumar S (2019) On the convergence of adam and beyond. arXiv preprint arXiv:1904.09237. https://doi.org/10.48550/arXiv.1904.09237
Goyal P, Dollár P, Girshick R, Noordhuis P, Wesolowski L, Kyrola A, He K (2017) Accurate, large minibatch sgd: training imagenet in 1 hour. arXiv preprint arXiv:1706.02677. https://doi.org/10.48550/arXiv.1706.02677
Keskar, N. S., Mudigere, D., Nocedal, J., Smelyanskiy, M., & Tang, P. T. P. (2016). On large-batch training for deep learning: generalization gap and sharp minima. arXiv preprint arXiv:1609.04836 .
Han, J., Yamazaki, K., Makino, S., & Shirasawa, T. (2013, May). Optimization of deep drawing process for circular cup forming. In 10th World congress on structural and multidisciplinary optimization May 19, 24.
Chen Z, Zhao J, Fang G (2019) Finite element modeling for deep-drawing of aluminum alloy sheet 6014–T4 using anisotropic yield and non-AFR models. Int J Adv Manuf Technol 104:535–549. https://doi.org/10.1007/s00170-019-03921-w
He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR).
Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556 .
Szegedy C, Liu W, Jia Y, Sermanet P, Reed S, Anguelov D, Rabinovich A (2015) Going deeper with convolutions. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). https://doi.org/10.1109/CVPR.2015.7298594
Cui, B., Guo, H., & Zhou, Z. H. (2016). Multi-task deep neural networks for non-linear regression. In Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence (AAAI).
Zhang W, Wu X, Liu T (2018) A comparative study of deep neural networks for non-linear regression. J Comput Sci Technol 33(3):478–496
Glorot, X., Bordes, A., & Bengio, Y. (2011). Deep sparse rectifier neural networks. In Proceedings of the 14th International Conference on Artificial Intelligence and Statistics (AISTATS).
Nair, V., & Hinton, G. E. (2010). Rectified linear units improve restricted Boltzmann machines. In Proceedings of the 27th International Conference on Machine Learning (ICML).
Acknowledgements
The authors are appreciated the support from the Ministry of Trade, Industry & Energy (MOTIE, Korea) in establishing this work.
Funding
This research is supported by the Automobile Industry Technology Development Program ((20019142) funded by the Ministry of Trade, Industry & Energy (MOTIE, Korea).
Author information
Authors and Affiliations
Contributions
Conceptualization, F. Guo and H. Jeong; methodology, F. Guo and H. Jeong; software, F. Guo and H. Jeong; validation, F. Guo and H. Jeong and N. Kim; formal analysis, F. Guo and H. Jeong; investigation D. Park, F. Guo and H. Jeong; resources, N. Kim and B. Sung; data curation, F. Guo, H. Jeong and D. Park; writing original draft preparation, F. Guo and H. Jeong; writing–review and editing, N. Kim; visualization, F. Guo and H. Jeong; supervision, N. Kim and B. Sung; project administration, N. Kim; funding acquisition, N. Kim; All authors have read and agreed to the published version of the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix 1. Deep neural network (DNN) modeling
Figure 16 illustrates the DNN model design that we used in this study. The input layer has three neurons including VBHF, trajectory, DBRF and the output layer has three neurons with failure, wrinkling and springback. We chose a DNN architecture with five hidden layers containing 128, 256, 512, 256, and 128 neurons, respectively, based on previous research that suggested deeper and wider networks could improve model performance by increasing the capacity for learning complex patterns and relationships [60, 61, 62]. This architecture has also been shown to be effective for non-linear regression tasks, as it allows for the modeling of non-linear relationships between the input and output variables [63, 64].
The rectified linear unit (ReLU) activation functions were employed in the hidden layers of our DNN model to introduce nonlinearity and enable better feature representation and extraction [65, 66]. The DNN was implemented using keras on top of TensorFlow.
To assess the accuracy of the models, two metrics have been adopted: mean absolute error (MAE) metric and mean squared error (MSE) metric were used. MAE metric was employed to assess the performance quality of an approximative model by represented the average magnitude of the residuals in datasets. MSE metric evaluated the variance of the residuals and was used to assess the overall quality of the approximate model. While MSE metric evaluated the average of the squared discrepancies among the original values and their corresponding approximations in the datasets. The values of the terms mentioned were computed using the following mathematical formulas.
where, Ŷi is approximated value of Y, Yi is actual value of Y, n denote total number of samples.
Appendix 2. Design of experiment
This part describes the processing parameter design. To determine the effect of S-VBHF, punch stroke trajectory, DBRF and μs, mixed level and four factors simulation sets design experiments were implemented on deep drawing process. Four factors of the processing parameter determination are shown in the Fig. 7, the level of S-VBHF, punch stroke trajectory, DBRF and μs is selected as 3, 2, 6, and 3. The whole stroke was partitioned into three sub-stroke sections, thus the trajectory case could be separated as arranged as 6 cases where L, M, and H means low level, middle level, and high level in the three sub-stroke sections in Fig. 17a. Considering the configuration of the stroke trajectory that which level should be longer in whole stroke, reference the literature mentioned in the cylindrical cup cases, only two sub-strokes of 10 mm and one 20 mm sub-stroke were used. S-VBHF amplitude configuration was partitioned from 1 to 5kN as shown in Fig. 17c, because of space limitations, three magnitude of the BHF only considering trajectory case L-M-H is shown (Fig. 18).
Appendix 3. Nondominated sorting genetic algorithm-II (NSGA-II)
NSGA-II algorithm was implemented to tackle the challenge of determining optimal process parameters in metal forming processes in this study. The NSGA-II algorithm adheres to a general framework by a genetic algorithm, incorporating a survival selection process that has been modified to suit the optimization problem. The process involves the classification of individuals based on their non-domination status, which reflects the trade-off relationship among the individuals with respect to the objectives. Subsequently, individuals are selected by their crowding distances, which quantify the diversity of solutions in the population. Each individual in the population consists of design variables, including the blank holder force (VBHF), drawbead restraining force (DBRF), friction coefficient (μs) and VBHF trajectory. The evaluation of the individuals was performed in terms of three defects, namely failure, wrinkling and springback.The NSGA-II was implemented using keras on top of TensorFlow (Fig. 19).
Appendix 4. Monte Carlo simulation (MCS) method
Figure 20 present the probability density curves described by statistically normal distributions of process parameters (a) VBHF, (b) trajectory, (c) DBRF, and (d) μs. The MCS method was implemented using a model in MATLAB.
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
Guo, F., Jeong, H., Park, D. et al. Numerical multi-objective optimization of segmented and variable blank holder force trajectories in deep drawing based on DNN-GA-MCS strategy. Int J Adv Manuf Technol 130, 3445–3468 (2024). https://doi.org/10.1007/s00170-023-12846-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00170-023-12846-4