Proceedings of the fourth International Symposium on Mechanics, Aerospace and Informatics Engineering 2009 Changwon Exhibition Convention Center (CECO), Korea, September 10 ∼ 12, 2009 ISMAI04-MF-01

-89-

An Intelligent Metal Forming Simulator AFDEX and its Applications

Man-Soo Joun1)*, Min-Cheol Lee2), Jae-Gun Eom3), Qiu-Shi Li4), Shang-Hyun Shim4)

Abstract

We present a concept of intelligent metal forming simulation technique and its basic requirements and introduce an intelligent forging simulator AFDEX with emphasis on its fulfillment of the requirements. The intelligent forging simulator is determined by the adaptive and optimal mesh generation technique and many intelligent application-oriented special functions which minimize the user-intervention during forging simulation. Of course, the solution accuracy should be optimized in the intelligent simulation. We have developed AFDEX to meet the requirements on the intelligent metal forming simulation. Its characteristics are introduced with the help of typical application examples. To verify the simulated results, we carry out tensile test simulation and several forging processes including a rotor pole forging process, a crank shaft forging process and a bevel gear enclosed die forging process. The predictions are compared with the experiments.

Key Words : Intelligent Metal Forming Simulator, Mesh Quality, Experiment, Mesh Generator, Forging Simulator

1. Introduction

Recently application of metal forming simulation technologies finds rapidly their ways into the related industries from the large scaled companies to the small sized companies. For example, most leading small or medium sized metal forming companies are now using various forging simulators(1-6). The metal forming simulation technologies have resulted in innovation of metal forming process design. Now their applications have deep influence even in the quality control of the metal formed products as well as the cost reduction and the productivity enhancement. In the near future, major consumers of automobile companies may impose submission of simulation results on their manufacturers before final acceptance of the metal formed products.

However, the speed of advance in metal forming simulation technologies is slow relative to the demand of the users in the industries. The speed is rather being decelerated because the simulator is so complicated and the users’ demands are so creative. It is needless to say that many users have various sorts of discontents against their metal forming simulators. Experiences make us say that many problems come out of their own special situations. However, it is an undeniable fact that computational time, unpredicted simulation stop, incredibility due to bad mesh quality and solution inaccuracy may be bothering the users. We believe that all the above factors are deeply related to mesh generation technique. Mesh quality, that is, mesh regularity and mesh normality over the workpiece-die interface can affect directly or indirectly the computational time. Especially mesh quality over the workpiece-die interface in metal forming is very important. The authors(7) showed that

the bad normality of quadrilateral elements along the workpiece-die interface can worsen the problem of numerical oscillation of normal contact stress. Thus, in order to obtain more accurate solution, the mesh quality over the contact area should be enhanced. It is, however, believed that most researchers on mesh generation for application to metal forming simulation have their interests not in mesh quality but in meshing generality of their mesh generation techniques although many researchers have studied academically the adaptive mesh generation techniques(8-12).

Recently, Lee et al.(13-15) developed a tetrahedral element generation technique which is excellent in mesh quality and applied it to an intelligent metal forming simulator, AFDEX 3D (1). In this paper, we present a concept of intelligent metal forming simulation technique and its basic requirements and then the characteristics of an intelligent metal forming simulator, AFDEX 3D, are introduced with emphasis on their fulfillment of the requirements

2. Intelligent forging simulation And AFDEX

Intelligent forging simulation in this paper is defined as the forging simulation for higher accuracy with minimized user intervention. The intelligent forging simulation technologies should satisfy the following requirements:

• The solution should be accurate. • The intelligent meshing or remeshing capabilities should be

supported to minimize the loss of solution accuracy during remeshing. • Optimized or adaptive meshing density capabilities should be

adopted to obtain the solution within reasonable computational time without loss of solution accuracy. Also, complex geometries can be meshed without failure. • Characteriestic boundaries or edges and workpiece-die

contact boundaries should be accurately traced during simulation and remeshing, which are most important in metal forming simulation. • Automatic simulation of multi-stage forging processes should

be made to reduce the total simulation time including both

1)* School of Mechanical and Aerospace Engineering,

Gyeongsang National University 900 Gajwa-dong, Jinju, Gyeongnam, 660-701, Korea TEL: +82-55-751-5316 FAX: +82-55-751-5316 E-mail: msjoun@gnu.ac.kr

2) 2nd BK21, Gyeongsang National University, Korea 3) Researcher of TIC of Gyeongsang National University 4) Graduate Student of Gyeongsang National University

-90-

computational time and user’s processing time between the stages. • The usage should be convenient and user-friendly.

The authors developed a forging simulator called AFDEX to fulfill the above requirements on the intelligent forging simulation.

First, we discuss about the accuracy of AFDEX. Fig. 1 compares experiments of three tensile tests with predictions obtained by AFDEX. The flow stresses with respect to strains used for the predictions were obtained using AFDEX/MAT, which is a material property identification module from the tensile test. The predictions followed the experiments so accurately that the error between them was less than 0.3% from the yielding point to the fracture point(16). This comparison shows the accuracy of AFDEX and the fact that accurate prediction of a tensile test can be obtained. It is believed that the accurate prediction of the tensile test can contribute to the advance in understanding of microstructural and mechanical behaviors nearing fracture point as well as acquisition of flow stress information.

Fig. 2 compares the metal flow lines of experiments and predictions(17). As shown in this figure, they are so close that engineers can remove the costly and time-taking tryouts which had been traditionally essential in developing hot forging processes of bearing parts and the like which require fulfillment of strict metal flow lines.

Fig. 3 and Fig. 4 show optimized quadrilateral element mesh system(7) and tetrahedral element mesh system(15) of AFDEX 2D and AFDEX 3D, respectively. During mesh generation or remeshing, the deviation of mesh density between desired and generated and the number of transition elements were minimized. In addition, mesh quality, that is mesh regularity and normality near the workpiece-die interface, was optimized in order to reduce numerical inaccuracy(7). As shown in Fig. 4, the mesh density of the generated tetrahedrons follows well the desired mesh density.

Mesh density control capability especially near the sharp edge

is of a great importance for precision simulation of metal forming processes. In AFDEX, geometrical aspects including surface curvature, workpiece-die interface, sharp edge and the like as well as state variables including strain, strain-rate, temperature gradient and the like are both used to determinate an optimal mesh density(15). Fig. 5 is a typical tetrahedral element mesh system, that is, the final configuration of a spiral bevel gear forging simulation. During the whole simulation, all teeth of the spiral bevel gear were well described(1), which is sufficient enough to show one of the characteristics of the mesh generator adopted by AFDEX.

Fig. 6 shows robustness of the mesh generator of AFDEX 3D, which is one of essential requirements which an intelligent forging simulator should possess.

Fig. 7 shows the edges generated in cold forging and Fig. 8 shows the workpiece-die boundaries. Such characteristic boundaries or surfaces should be accurately described for carrying out the intelligent simulation.

Intelligent forging simulator should have the automatic simulation capability of a sequence of multi-stage forging processes to minimize the actual total simulation time including the computational time and the user’s intervention time. Fig. 9 and Fig. 10 show the predictions of 5-stage ball-stud forging processes, which were obtained in an automatic manner by AFDEX 2D and AFDEX 3D for the same process, respectively.

The convenience in using the forging simulator from standpoint of process design engineers who have creative ideas is one of the essential requirements of the intelligent forging simulator. AFDEX has been developed based on various advices from the related industries and thus it is believed that our pre and post processors are very friendly to the users even if they are still evolving day after day to meet users’ requirements.

We applied AFDEX to simulate a rotor pole forging process, a crank shaft forging process and a bevel gear enclosed die forging process and compared the predictions with the experiments, as shown in Fig. 11-13, respectively. The comparison shows the close similarity between the predictions and the experiments. Note that the details of the predictions can be obtained from the web site of AFDEX(1).

Engineering strain (mm/mm)

Eng

inee

ring

stre

ss(M

Pa)

0 0.1 0.2 0.3 0.40

200

400

600

800

1000

1200

Experiment (SCM435)Analysis (SCM435)Experiment (ESW95)Analysis (ESW95)Experiment (ESW105)Analysis (ESW105)

Fig. 1 Comparison three tensile tests Fig. 2 Comparison of metal flow lines

-91-

Fig. 3 Optimized quadrilateral element Fig. 4 Optimized tetrahedral element

Fig. 5 Intelligent remeshing Fig. 6 Robustness of remeshing

Fig. 7 Edges generated by the workpiece-die contact

Fig. 8 Characteristic lines generated along the die-free free-surface interface

-92-

Fig. 9 2D simulation results obtained by the automatic simulation capability

Fig. 10 3D simulation results obtained by the automatic simulation capability

Ben

ding

pro

cess

Sizi

ng p

roce

ss

Fig. 11 Rotor pole forging simulation and its comparison with the experiment

-93-

Fig. 12 Crank shaft forging simulation and its comparison with the experiment

Fig. 13 Bevel gear enclosed die forging simulation and its comparison with the experiment

3. Conclusions

Forging simulation became an essential activity during

developing a forging process. Owing to advances in computer hardware technologies, the major problem is moving from the computational time to the accuracy of solution. Now we can get the simulation results of a forging process within five minutes which took more than three hours just ten years ago. With much more faster personal computer, we can reduce the computational time to less than 30 seconds. This change of computer hardware

References

(1) AFDEX, http://engine.gsnu.ac.kr/-msjoun/afdex.htm (2) EESY-2-FORM, http://www.cpmgmbh.de (3) DEFORM, http://www.deform.com (4) FORGE, http://www.transvalor.com (5) Superforge, Superform, http://www.mscsoftware.com (6) QForm, http://www.quantor.com (7) Joun, M.S., Lee, M.C. “Quadrilateral finite-element

generation and mesh quality control for metal forming simulation.” Int J Numer Methods Eng, Vol. 40, No. 21, pp. 4059 ~ 4075, 1997

(8) Zhu J, Gotoh M. “Automatic remeshing of 2D quadrilateral elements and its application to continuous deformation simulation: Part I,” Remeshing algorithm. J Mater Process Technol, Vol. 87, No. 1~3, pp. 165 ~ 178, 1999

(9) Kwak D-Y, Cheon J-S, Im Y-T. “Remeshing for metal forming simulating-Part I Two-dimensional quadrilateral remeshing,” Int J Numer Methods Eng, Vol. 53, No. 11, pp. 2463 ~ 2500, 2002

(10) Liu D, Luo ZJ, Gu MX. “The algorithm of automatic local mesh subdivision and its application to finite-element analysis of a large deformation forming process,” J Mater Process Technol, Vol. 83, No. 1~3, pp. 164 ~ 169, 1998

(11) Wan J, Kocak S, Shephard MS. “Automated adaptive 3D forming simulation processes,” Eng Comput, Vol. 21, No. 1, pp. 47 ~ 75, 2005

(12) Choi W-Y, Son I-H, Im Y-T. “Locally refined tetrahedral mesh generation based on advancing front technique with optimization and smoothing scheme,” Commun Numer Methods Eng, Vol.20, No. 9, pp. 681 ~ 688, 2004

(13) M. C. Lee, M. S. Joun, “Adaptive triangular element generation and optimization-based smoothing, Part 1-On the plane,” Adv Eng Softw, Vol. 39, No. 1, pp. 25 ~ 34, 2007

(14) M. C. Lee, M. S. Joun, “Adaptive triangular element generation and optimization-based smoothing, Part 2-On the surface,” Adv Eng Softw, Vol. 39, No. 1, pp. 35 ~ 46, 2007

(15) M.C. Lee, M.S. Joun, J. K. Lee, “Adaptive tetrahedral element generation and refinement to improve the quality of bulk metal forming simulation,” Finite Elem Anal Des, Vol. 42, pp. 788 ~ 802, 2007

(16) M. S. Joun, J. G. Eom, M. C. Lee, “A new method for acquiring true stress-strain curves over a large range of strains using a tensile test and finite element method,” Mech Mater, Vol. 40, No. 7, pp. 586 ~ 593, 2008

(17) M. S. Joun, H. K. Moon, R. Shivpuri, “Automatic simulation of a sequence of hot- former forging processes by a rigid thermo-viscoplastic finite element method,” J Eng Mater Technol-Trans ASME, Vol. 120, No. 4, pp. 291 ~ 296, 1998

Acknowledgements

This work is carried out by the support of the second phase

BK 21 project.