International Journal of Advanced Manufacturing Technology manuscript No. (will be inserted by the editor) An Intelligent Process Model: Predicting Springback in Single Point Incremental Forming Muhamad S. Khan · Frans Coenen · Clare Dixon · Subhieh El-Salhi · Mariluz Penalva · Asun Rivero the date of receipt and acceptance should be inserted later Abstract This paper proposes an Intelligent Process Model (IPM), founded on the concept of data mining, for predicting springback in the context of sheet metal forming, in particular, Single Point Incremental Form- ing (SPIF). A limitation with the SPIF process is that the application of the process results in geometric de- viations, which means that the resulting shape is not necessarily the desired shape. Errors are introduced in a non-linear manner for a variety of reasons, but a con- tributor is the geometry of the desired shape. A Local Geometry Matrix (LGM) representation is used that allows the capture of local geometries in such a way that they are suited to input to a classifier generator. It is demonstrated that a rule based classifier can be used to train the classifier and generate a classification model. The resulting model can then be used to predict errors with respect to new shapes so that some correc- tion strategy can be applied. The reported evaluation of the proposed IPM indicates that very promising re- sults can be obtained with regard to reducing the shape deviations due to springback. Keywords Single Point Incremental Forming · Data Mining · Springback Correction M. S. Khan · F. Coenen · C. Dixon · S. El-Salhi Department of Computer Science, University of Liv- erpool, Liverpool, U.K. (tel: (+44) 151 795 4280, fax: (+44) 151 795 4235) E-mail: [email protected], [email protected], [email protected], hssel- [email protected]M. Penalva · A. Rivero Tecnalia, Parque Tecnol´ ogico, San Sebasti´ an, Spain E-mail: [email protected], [email protected]1 Introduction Single Point Incremental Forming (SPIF) is a sheet metal forming process that involves a local and pro- gressive pressing out of the desired shape on a clamped sheet by a round-headed forming tool which follows a continuous path. Flexibility in sheet metal forming has attracted interest in recent years due to the shrinking of the product life cycle, increased demand and cus- tomisation requests [1, 8, 23]. SPIF offers full flexibility since the use of dedicated tooling, required with many sheet metal forming operations, isn’t necessary. How- ever one disadvantage is that the operation time is typ- ically high. Nevertheless, SPIF may still be of use for low volume series, can help increase the flexibility of any forming process in the design and industrialisa- tion phases by providing realistic prototypes and can be used in combination with other forming processes, for instance, to produce part details. Material springback is a phenomenon common to any sheet metal forming process that leads to the geo- metric inaccuracy of the resulting shape. In SPIF the increase in geometric deviations from the design shape because of the absence of tooling is the key deterrent from the widespread industrialisation take up of the process [17]. Typically, the management of geometric deviations due to springback is based on Finite Element (FE) predictions combined with practical expertise on the tooling set up. It is well known that Finite Element Analysis (FEA) requires intensive resource consump- tion. For SPIF the resource consumption represents a severe drawback since the process itself is highly time- consuming and hence numerical simulations are only affordable for small parts requiring short tool paths [4, 14, 31]. On the other hand, the accuracy of numeri-
13
Embed
An Intelligent Process Model: Predicting Springback in ...livrepository.liverpool.ac.uk/3005539/1/intJoAMT_2014-9-23.pdf · any forming process in the design and industrialisa-tion
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
International Journal of Advanced Manufacturing Technology manuscript No.(will be inserted by the editor)
An Intelligent Process Model: Predicting Springback inSingle Point Incremental Forming
Muhamad S. Khan · Frans Coenen · Clare Dixon · Subhieh El-Salhi ·Mariluz Penalva · Asun Rivero
the date of receipt and acceptance should be inserted later
Abstract This paper proposes an Intelligent Process
Model (IPM), founded on the concept of data mining,
for predicting springback in the context of sheet metal
forming, in particular, Single Point Incremental Form-
ing (SPIF). A limitation with the SPIF process is that
the application of the process results in geometric de-
viations, which means that the resulting shape is not
necessarily the desired shape. Errors are introduced in
a non-linear manner for a variety of reasons, but a con-
tributor is the geometry of the desired shape. A Local
Geometry Matrix (LGM) representation is used that
allows the capture of local geometries in such a way
that they are suited to input to a classifier generator.
It is demonstrated that a rule based classifier can be
used to train the classifier and generate a classification
model. The resulting model can then be used to predict
errors with respect to new shapes so that some correc-
tion strategy can be applied. The reported evaluation
of the proposed IPM indicates that very promising re-
sults can be obtained with regard to reducing the shape
From the above reported experiments it is clear that
shape deviation due to springback is reduced with re-
spect both geometries using the corrected CAD cloud
data formed using the proposed IPM.
An Intelligent Process Model: Predicting Springback in Single Point Incremental Forming 11
Fig. 13 Benchmark Pyramid manufactured using CAD cloud data (left); corrected cloud data(right); scale in millimeters
Fig. 14 Modified Pyramid manufactured using CAD cloud (left); corrected cloud data (right); scale in millimeters
5.2 IPM Processing Times
The overall processing time either to generate the clas-
sifier or process a new cloud given a classifier is less
than 10 seconds. We provide a breakdown below for
the Benchmark pyramid. The timings for the Modi-
fied Pyramid are similar. The IPM software was run
on an Apple Mac computer, running Mac OS X Ver-
sion 10.7.5, 2.66 GHz Intel Core i7 processor and 8GB
of RAM. The run time is shown in Table 4.
Classifier Generation Run Time (secs)Error calculation 6.2Error smoothing 0.4Rule generation 1.2PredictionsError prediction 1.9Generating corrected cloud 6.4
Table 4 Timing for Processes in the IPM
6 Conclusion
In this paper a classification based Intelligent Process
Model (IPM) was proposed to predict springback in
12 Muhamad S. Khan et al.
sheet metal forming incurred using SPIF. A Local Ge-
ometry Matrix (LGM) representation was proposed that
allows the capture of local 3-D surface geometries in
such a way that classifier generators can be effectively
applied. The resulting classifier was integrated into the
proposed IPM. The IPM was designed to:
– predict errors with respect to new surfaces to be
manufactured;
– apply corrections to the original CAD cloud;
– conduct appropriate smoothing to the corrected CAD
cloud.
The corrected cloud is ready for use in the definition
of a new corrected tool path that takes the springback
effect into consideration. The operation of the proposed
IPM has been fully described. The paper also presented
an evaluation of the operation of the proposed IPM,
using two fabricated parts that strongly indicates that
the IPM can be successfully used to generate corrected
tool paths. This was illustrated by reporting on exper-
iments that compared the quality of parts fabricated
using uncorrected tool paths with those fabricated us-
ing corrected tool paths. The timings to generate the
classifier or to apply the IPM to a new part were less
than 10 seconds.
Future work includes further development and anal-
ysis of the IPM. It is anticipated that improvements can
be made such as generalisation of the model to differ-
ent materials, metal thickness and other shapes. For
example, we are currently applying the IPM to differ-
ent geometries including those with curved surfaces and
different materials including titanium and Inconel. Re-
garding correction factors, as mentioned in Section 3.4,
we intend to investigate the use of correction factors
when applying the corrections, for example, using a
data mining approach to associate different factors with
different parts of the shape. Additionally, an iterative
version of the IPM that keeps predicting and apply-
ing corrections until they are within a certain tolerance
range (or a certain number of iterations has been car-
ried out) has been proposed and needs further analysis.
Acknowledgements The research leading to the results pre-sented in this paper has received funding from the EuropeanUnion Seventh Framework Programme (FP7/2007-2013) un-der grant agreement number 266208. The authors would liketo thank their project partners, in particular, David Baillyfrom RWTH-IBF (Germany) for his support in the manufac-ture and analysis of the test geometries.
References
1. Allwood JM, Utsunomiya H (2006) A survey
of flexible forming processes in Japan. Interna-
tional Journal of Machine Tools and Manufacture
46(15):1939–1960
2. Ambrogio G, Costantino I, De Napoli L, Filice L,
Fratini L, Muzzupappa M (2004) Influence of some
relevant process parameters on the dimensional ac-
curacy in incremental forming: a numerical and ex-
perimental investigation. Journal of Materials Pro-
cessing Technology 153-154:501–507
3. Ambrogio G, De Napoli L, Filice L (2009) A novel
approach based on multiple back-drawing incre-
mental forming to reduce geometry deviation. In-
ternational Journal of Material Forming 2(1):9–12
4. Bambach M (2010) A geometric model of the
kinematics of incremental sheet forming for the
prediction of membrane strains and sheet thick-
ness. Journal of Materials Processing Technology
210(12):1562–1573
5. Bambach M, Araghi BT, Hirt G (2009) Strategies
to improve the geometric accuracy in asymmetric
single point incremental forming. Production Engi-
neering Research and Development 3(2):145–156
6. Behera AK, Verbert J, Lauwers B, Duflou JR
(2013) Tool path compensation strategies for single
point incremental sheet forming using multivariate
adaptive regression splines. Computer-Aided De-
sign 45(3):575–590
7. Dearden G, Edwardson SP, Abed E, Bartkowiak K,
Watkins KG (2006) Correction of distortion and
design shape in aluminium structures using laser
forming. In: 25th International Congress on Appli-
cations of Lasers and Electro Optics (ICALEO), pp
813–817
8. Dirikolu MH, Akdemir E (2004) Computer aided
modelling of flexible forming processes. Journal of
Materials Processing Technology 148(3):376–381
9. Dunston S, Ranjithan S, Bernold E (1996) Neural
network model for the automated control of spring-
back in rebars. IEEE Expert: Intelligent Systems
and Their Applications 11(4):45–49
10. Edwardson SP, Watkins KG, Dearden G, Magee
J (2001) Generation of 3D shapes using a laser
forming technique. In: Proceedings of International
Congress on Applications of Lasers and Electro Op-
tics ICALEO, pp 2–5
11. Egerton PA, Hall W (1998) Computer graphics:
Mathematical first steps. Simon and Schuster In-
ternational
12. El-Salhi S, Coenen F, Dixon C, Khan MS (2012)
Identification of Correlations Between 3D Sur-
faces Using Data Mining Techniques: Predicting
Springback in Sheet Metal Forming. In: SGAI-
International Conference on Artificial Intelligence,
pp 391–404
An Intelligent Process Model: Predicting Springback in Single Point Incremental Forming 13
13. El-Salhi S, Coenen F, Dixon C, Khan MS (2013)
Predicting Features in Complex 3D Surfaces Us-
ing a Point Series Representation: A Case Study
in Sheet Metal Forming. In: Advanced Data Min-
ing and Applications, Springer, LNCS, vol 8346, pp
505–516
14. Hadoush A, van den Boogaard AH (2012) Effi-
cient implicit simulation of incremental sheet form-
ing. International Journal for Numerical Methods
in Engineering 90(5):597–612
15. Hirt G, Kopp R, Ames J, Bambach M (2004)
Forming strategies and Process Modelling for
CNC Incremental Sheet Forming. CIRP Annals-
Manufacturing Technology 53(1):203–206
16. Inamdar M, Date PP, Narasimhan K, Maiti SK,
Singh UP (2000) Development of an articial neural
network to predict springback in Air Vee bending.
International Journal of Advanced Manufacturing
Technology 16(5):376–381
17. Jeswiet J, Geiger M, Engel U, Kleiner M, Schikorra
M, Duflou J, Neugebauer R, Bariani P, Bruschi S
(2008) Metal forming since 2000. CIRP Journal of
Manufacturing Science and Technology 1:2–17
18. Junk S, Hirt G, Chouvalova I (2003) Forming
strategies and tools in incremental sheet forming.
In: Proceedings of the 10th International Confer-
ence on Sheet Metal (SHEMET), pp 57–64
19. Khan MS, Coenen F, Dixon C, El-Salhi S (2012)
Classification based 3-D surface analysis: Pre-
dicting springback in sheet metal forming. Jour-
nal of Theoretical and Applied Computer Science
6(2):45–59
20. Khan MS, Coenen F, Dixon C, El-Salhi S (2012)
Finding correlations between 3-D surfaces: A study
in asymmetric incremental sheet forming. In: Proc.
Machine Learning and Data Mining in Pattern
Recognition (MLDM’12), Springer LNAI 7376, pp
336–379
21. Kim DJ, Kim BM (2000) Application of neural net-
work and FEM for metal forming processes. Inter-
national Journal of Machine Tools and Manufac-
ture 40(6):911–925
22. Kinsey B, Cao J, Solla S (2000) Consistent
and minimal springback using a stepped binder
force trajectory and neural network control. Jour-
nal of Engineering Materials and Technology
122(1113):113–118
23. Li RJ, Li MZ, Qiu NJ, Cai ZY (2014) Surface
flexible rolling for three-dimensional sheet metal
parts. Journal of Materials Processing Technology
214(2):380–389
24. Lu B, Chen J, Ou H, Cao J (2013) Feature-based
tool path generation approach for incremental sheet
forming process. Journal of Materials Processing
Technology 213:1221–1223
25. Manabe K, Yang M, Yoshihara S (1998) Artifi-
cial intelligence identification of process parameters
and adaptive control system for deep drawing pro-
cess. Journal of Materials Processing Technology
80-81:421–426
26. Micari F, Ambrogio G, Filice L (2007) Shape and
dimensional accuracy in single point incremental
forming: State of the art and future trends. Journal
of Materials Processing Technology 191(1-3):390–
395
27. Pathak KK, Panthi S, Ramakrishnan N (2005) Ap-
plication of neural network in sheet metal bending
process. Defence Science Journal 55(2):125–131
28. Rauch M, Hascoet J, Hamman J, Plenel Y (2009)
Tool path programming optimization for incremen-
tal sheet forming application. Computer Aided De-
sign 41:877–85
29. Ruffini R, Cao J (1998) Using neural network
for springback minimization in a channel forming
process. Journal of Materials and Manufacturing
107(5):65–73
30. Skjoedt M, Hancock MH, Bay N (2007) Creating
helical tool paths for single point incremental form-
ing. Key Engineering Materials 344:583–590
31. Thibaud S, Ben Hmida R, Richard F, Malecot P
(2012) A fully parametric toolbox for the simula-
tion of single point incremental sheet forming pro-
cess: Numerical feasibility and experimental vali-
dation. Simulation Modelling Practice and Theory
29:32–43
32. Yin J, Li D (2004) Knowledge discovery from finite
element simulation data. In: Proceedings of 2004
International Conference on Machine Learning and
Cybernetics, pp 1335–1340
33. Zhang S, Luo C, Peng Y, Li D, Yang H (2003)
Study on factors affecting springback and applica-
tion of data mining in springback analysis. Journal