Application of Neural Network Technique to Predict the Formability in Incremental Forming Process

Abstract:

Nowadays, several researchers all over the world on Incremental Sheet Forming are focused on formability analysis and, in particular, on the development of a model able to predict material response also in the design phase. Focalising the attention on this aspect and taking into account the literature review, it is possible to establish that both numerical and analytical approaches were utilised to justify the experimental evidences. Anyway, many efforts are spent in order to define new models to predict material breakage during incremental forming of simply shaped components. On the contrary, the availability of a predictive model able to detect failure insurgence, based on the products geometry, is a relevant issue to improve industrial application of incremental forming process. At the same time, it still represents an open point of the current scientific research. According to this aim, a preliminary analysis was executed to detect the factors that deeply impact on the process performance. In a subsequent analysis, a robust Neural Network was built in order to define a more reliable tool for calculating the allowable component height when a multi-slope trajectory is imposed. The reliability of the proposed approach was tested executing some experimental results. All the details are accurately discussed in the paper.

• Periodical: Key Engineering Materials (Volumes 410-411)
• Pages: 381-389
• DOI: https://doi.org/10.4028/www.scientific.net/KEM.410-411.381
• Online since: March 2009
• Authors: Giuseppina Ambrogio, Luigino Filice
• Keywords: Incremental Forming, Neural Network Design, Taguchi’s Design of Experiment

References

[1] Y. Kim, J. Park: Journal of Material Processing and Technology vol. 130-131 (2002), pp.42-46.

[2] L. Fratini, G. Ambrogio, R. Di Lorenzo, L. Filice, F. Micari, Annals of the CIRP vol. 53/1/ (2004), pp.207-210.

DOI: 10.1016/s0007-8506(07)60680-5

[3] G. Hirt, S. Junk and N. Witulsky, Incremental sheet forming: quality evaluation and process simulation, Proc. of the 7th ICTP Conference (2002), pp.925-930.

[4] G. Ambrogio, L. Filice, L. De Napoli, M. Muzzupappa, Journal of Engineering Manufacture . Part B (2005), p.823. 830.

[5] J. Jesweit, E. Hagan, Rapid Proto-typing of a Headlight with Sheet Metal, Proceedings of Shemet, (2001), pp.165-170.

[6] G. Ambrogio, L. De Napoli, L. Filice, F. Gagliardi, M. Muzzupappa, Journal of Materials Processing and Technology, vol. 162-163 (2005), pp.156-162.

DOI: 10.1016/j.jmatprotec.2005.02.148

[7] G. Ambrogio, L. Filice, L. Manco, F. Micari, A depth dependent analytical approach to determine material breaking in SPIF, AIP Conference Proceedings of the 10 th Esaform (2007), vol. 907, pp.331-336.

DOI: 10.1063/1.2729534

[8] W. Sukthomya, J. Tannock, Neural Network Computing and Application vol. 14 (2005), pp.337-344.

[9] F.C. Khaw, B.S. Lim, L.E.N. Lim, Neurocomputing vol. 7 (1995), pp.225-245.

[10] K.N. Hornik, M. Stinchcombe,H. White, Neural Network vol. 2 n. 5 (1990), pp.359-566.

[11] A.J. Maren, D. Jones, S. Franklin, in: Handbook of Neural Computing Applications, edited by Academic Press (1990).

[12] M. Hagan, H. Demuth, M. Beale, in: Neural Network Design, edited by Citic Publishing House (1996).

[13] J. Freeman, D. Skapura, in Neural Networks Algorithms, Applications and programming Techniques, edited by Addison-Wesley (1999).

[14] V. Kecman, in Learning and soft computing, edited by MIT press (2001).

[15] H. Demuth, M. Beale, M. Hagan, in Neural Network Toolbox User's Guide, edited by The MathWorks Inc. (2005).