Physical Modeling of Extrusion Process

A .Assempour1, S. Razi2
Center of Excellence in Design, Robotics, and Automation Department of Mechanical Engineering Sharif university of Technology Tehran,Iran E-MAIL: [email protected]

Keywords: extrusion physical - modeling- plexi glass- plasticine 1 Intrduction

In the bulk metal forming processes, a successful and powerful technique for the modeling of deformation is physical modeling using the plasticine as a modeling material. Plasticine is a trade name by Harbut, U.K. and is a modeling clay for children as a playing [1, 2]. Harbuts plasticine is composed of 52.3% vaporable materials such as grease and 47.7% of ash. Mechanical properties of plasticine at room temperature are very similar to those of metals at high temperature [3]. The flow stress of plasticine can be expressed as Eq. (1) [4]:

Ar

ch

ive

Physical modeling of extrusion process has been studied to obtain qualitative and quantitative information. To observe the flow pattern during the process the plasticine specimen was made with layers of different colors (dark and light). The strain distribution was obtained by measuring the thickness of the plasticine layers and considering the axi-symmetric formulations. The stress distribution and subsequently the extrusion load were computed according to the strain distribution. To model the real frictional effect several ring compression tests have been performed. In order to verify the validity of the modeling data, a similarity study between plasticine and Al 2024-T4 was made. The load obtained by this technique was compared with the results of ANSYS modeling and slab method analysis.

&m =k n
where:

of

SI

(1)

1 2

- Assistant Professor - Graduate Student

www.SID.ir

62

Journal of Mechanical Engineering

Vol. 4, No. 1, April 2003

& is effective strain rate, n is work and are effective stress and strain,
hardening exponent and m is strain rate sensitivity exponent. There have been many reports of plasticine modeling in the literature .The results of these modelings have provided researchers varying amounts of plastic flow information. Review of the literature indicates that only qualitative information isobtained by this technique in the bulk deformation processes. The quantitative results have been achieved indirectly by applying the results of the plasticine modeling to the numerical techniques like finite element methods. Chu et al. [5] studied the process sequence design of a large axi-symmetric hot forging product of AISI 4130 in nozzle type by using the plasticine and lead as the model materials .The result of this study was used in finite element modeling for the stress analysis. Kim et al.[6] used layered plasticine for the study of the flow in hot closed die forging . Misilolek [7] used plasticine to understand the role of the metal flow in deformation history in the extrusion process. In this study, modeling of extrusion by plasticine in plexi glass die has been studied. The load and stress - strain distributions were obtained with some range of approximations without applying any numerical method. For the simulation of the real frictional effect, several ring compression tests have been performed on plasticine with different lubricants. The load obtained by this technique compared with load result of ANSYS modeling and slab method analysis.

A 30 tons universal testing machine was used for the experimental works. In order to estimate the frictional shear stress, , the following expression was used [8]: = f . (2) where f is the friction factor. The well-known ring compression test was used to measure the friction factor at die and work piece interface [9]. In this work, the plasticine was formed into rings by plexi glass die set. The Rings were 30 mm inner diameter, 60mm outer diameter and 20mm height. To create similar conditions on all the specimens, the rings were kept in the freezer for two hours and left in the air for 24 hours [10] before performing the experiments. The rings were upset by a pair of flat plexi glass dies. To find similar frictional condition in real hot extrusion process, several lubricants were used in ring tests. Vaseline, baby powder, facial tissue, plastic and soapsuds were the lubricants used in these experiments. The ram velocity used in ring tests was 15mm/min. For every lubricant the friction factor was obtained by measuring the ring dimensional changes and reading the calibration curves [10].According to the ring compression test results, the friction factor varied from 0.05 ( vaseline ),0.3 (soap suds),0.4 (plastic),0.4-0.5 ( facial tissue),and 0.5-0.7 (baby powder). Since the friction factor during the hot working of metal normally falls between 0.3-0.4 [11] soap suds was selected as the lubricant for the modeling experiments in the rest of this work. For the studies of physical modeling, the die set was made with Plexiglas using the lathe Machine. Figure 1 displays the schematic of the die set made in this work. Layered plasticine with two different colors (dark and light) was used for making a billet of 56mm height and 45mm diameter. The temperature during the tests was about 19oC.

Ar

ch

ive

of

The experimental works

SI

www.SID.ir

Physical Modeling of Extrusion Process

63

The plasticine sample was extruded. Figure 2 shows the deformed sample. To see the flow pattern, the specimen was sectioned and displayed in the figure. 3 Data processing and result

Knowing the flow directions and thickness variations on plasticine layers, the strain and stress distributions and the extrusion load were obtained. It was assumed that the principal directions coincided with the normal to the plasticine layers and the flow directions. Accordingly, the axi- symmetric effective strain and the radial strain 1 are:
(3) where t1 and t 2 are the initial and final layer ickness spectively. Using digital caliper, the thickness variation in figure 2 was measured at different locations and used in equation 3. In figure 2, the minimum thickness was measured in each layer. Figure 3 displays the result of effective strain distribution on the work piece. Using these strain values and applying the Al2024-T4 material properties the stress values were found on the real part.The flow stress data for Al2024-T4 at 2000 C is [11]: & = = ln 1 t2 t1

& 0.05 = 60 0.128

of
.

SI

(4)

To compare the results obtained by the physical modeling, the slab method and ANSYS 5.4 program were used for prediction of the extrusion load. 4.1 Use of slab method & and flow stress , the Considering average values of strain , strain-rate extrusion load F , was calculated by using the following expression [11]:

Ar

F = (1. cos + .sin )dS (5) S where F is the extrusion load, 1 is the normal stress at die neighborhood, is the frictional shear stress, is the angle between normal to die surface and vertical direction and S is the punch and die area. The result of the load calculation with this method was 128 tons. Prediction of axi-symmetric extrusion load from other methods

ch

The required extrusion load was calculated by using the following expression:

ive

Assuming the isothermal conditions, the effective stress distribution in the actual workpiece was computed. In equation (4), the values of effective strains were used according to the results of physical modeling. The strain rates were obtained by simply dividing the effective strains by the total deformation time. Figure 4 displays the results of effective stress.

www.SID.ir

64

Journal of Mechanical Engineering

Vol. 4, No. 1, April 2003

F = 2r0 L c + r02 ln( A0 / A1 ) + r02 d ln( A0 / A1 ) /(sin cos ) + 2r02

2 cot 3 sin (6)

where r0 and L are the initial radius and length of the billet, is the die angle, A0 and A1 are the cross-sectional areas of the billet, d and c are the frictional shear stresses at the die and container surfaces ,respectively.The average strain-rate was obtained by [12],

12V0 r02 tan r03 r13

The flow pattern obtained by the physical modeling in figure 2 is very similar to the grid deformation obtained by ANSYS in figure 6. In both figures it can be observed the faster flow of material at the central zone. Also in both figures, it can be observed the very high elongation after passing the die zones. The trend of effective strain distribution obtained by the physical modeling experiment in figure 3 matches fairly well with the effective strain field calculated by ANSYS in figure 7. However, the predicted strain values by ANSYS are almost 30 percent larger. In addition, the predicted loads by the slab method and ANSYS are larger than that predicted by the physical modeling.

Ar
Discussion

The computational conditions used in simulation of hot extrusion process by ANSYS program [13] are as the following: 1-The flow stress used in simulation was nonlinear plastic model as Eq.(4). 2-The frictional condition used in simulation was f = 0.3 using constant shear friction model. 3-The initial billet geometry used in simulation was 45mm and 56mm for the cylinder diameter and height respectively. 4-The youngs modulus used in simulation was E = 77( MPa) . The initial mesh layout is given in Figure 5 where 144 viscosolid 108 elements were used. Figure 6 displays the result of grid distortion at the final deformation obtained by ANSYS modeling. Figure7 displays the result of effective strain distribution at the final deformation obtained by ANSYS modeling. The result of the load calculation with this method was 162 tons.

ch

ive

of

4.2 USE OF ANSYS PROGRAM

SI

where V0 is the ram velocity and is the average effective strain that was calculated by[12], = ln R where R is the extrusion ratio ( A0 / A1 ) ,.The flow stress of Al2024-T4 at 200 0 C can be calculated by Eq.(4).The result of the load calculation with this method was 167 tons.

www.SID.ir

Physical Modeling of Extrusion Process

65

In general, the physical modeling could provide good qualitative results. The quantitative results obtained by the physical modeling were under estimated comparing with those predicted by the other methods. By reviewing the experimental procedure and the assumptions in data processing, the followings may be explained as the major causes of under estimation in the quantitative results: 1- At the die and material interface where friction stress exists, the principal strain direction could deviate from the normal to the plasticine layer. Perhaps, this is one of the major influencing factors on the result of load estimation. One could improve the results by estimating the friction stress and correcting the principal stress and strain directions at die and workpiece interface. 2-Application of the total deformation theory in large deformations could be too simple for extracting the quantitative results in one-step. To achieve better results, increments of strain components could be measured at the intermediate steps. 6 Conclusion

8- REFRENCES

[1] Erman and S.L. Semiatin, Introdution to physical modeling of metalworking processes Physical Modeling of Metal Working Processes, Proceeding of a Symposium by the TMS-AIME annual meeting in Denver, pp24-27(1987).

[4] H. Kim, J. Kim and N. Kim, Physical and numerical modeling of hot closeddie forging to reduce Forging Load and Die Wear, J. of Material Processing Technology, Vol. 42, PP401-420, (1994). [5] E.Chu,,Y.T. Kim and N.Lee, Process Sequence design of large axisymmetric forging product AISI 4130 in nozzle type, J. of Material Processing Technology, Vol. 48, No. 1-4,pp 143-149 (1995). [6] H. Kim Y. Kim and J. Kim, Physical and numerical modeling of hot closeddie forging to reduce forging load and die wear, J. of Material Processing Technology, Vol. 42, pp. 401-420, (1994).

Ar

[3] L.E. Matson and H.L. Gegel, physical modeling of metalworking processes: determination of large plastic strains Physical Modeling of Metal Working Processes; Proceeding of a Symposium by the TMS-AIME annual meeting in Denver, pp 115-121(1987).

ch

[2] ASM "Metal Handbook", Ninth Edition (1999).

ive

of

It was demonstrated that physical modeling of Al2024-T4 as an axisymmetric part was useful in obtaining some information about the trend of strain distribution and pattern of material flow pattern. Comparisons of the quantiatative results obtained by physical modeling and those obtained by finite element and slab methods show some range of under estimations. Improvement of the quantitative results could be achieved by reviewing the experimental procedure and the assumptions in data processing.

SI

www.SID.ir

66

Journal of Mechanical Engineering

Vol. 4, No. 1, April 2003

[7] W.Z. Misiolek, Material physical response in the extrusion process, J. of Material Processing Technology, Vol. 60, No. 1-3, pp.117-124, (1996). [8] H. Sofuoglu and J. Rasty Three-dimentional analysis of extrusion process utilizing the physical modeling technique, Transaction of ASME Vol. 115 J. of Energy Resources pp.132-40 (1993). [9] W.Z. Misiolek, Material physical response in the extrusion process, J. of Material Processing Technology, Vol. 60, No. 1-3, pp.117-124 ( 1996). [10] M. Tanaka, S. Onoand and T.Iwadate, Control of friction for simulation of forging, using, Physical Modeling of Metal Working Processes, Proceeding of a Symposium by the TMS-AIME annual meeting in Denver, pp.31-39, (1987).

[13] ANSYS user's manual Vo1.I-IV, Swanson Analysis Systems.Inc. (1992).

Figures:

Ar

ch
Figure.1 Plexiglass die set for physical modeling eperiments

ive

of
www.SID.ir

SI

[12] W. Johnson and P.B. Mellor, Engineering Plasticity, New York, Van Nostrand Reinhold, pp. 392-410, 416-417,418-419 (1973).

[1]] T. Altan, S.I. Oh, L.G. Harold, Metal Forming, AMERICAN SOCITETY FOR METALS, Metals park, OH 44073 (1983).

Physical Modeling of Extrusion Process

67

Figure 2 The billet and deformed plasticine specimen

Figure 3 The result of effective strain distribution by the physical modeling

Ar
Figure 4 The result of effective stress (MPa) distribution by the physical modeling

ch

ive

of
www.SID.ir

SI

68

Journal of Mechanical Engineering

Vol. 4, No. 1, April 2003

Figure 5 The initial mesh layout used in ANSYS modeling

Figure 6 Grid distortion at the final deformation obtained from ANSYS modeling

Ar
Figure 7 Effective strain field calculated by ANSYS program

ch

ive

of
www.SID.ir

SI

Physical Modeling of Extrusion Process

69

. . . . . ANSYS .

Ar

ch

ive

of
www.SID.ir

SI