Tribol Lett (2010) 37:353359 DOI 10.

1007/s11249-009-9529-0

ORIGINAL PAPER

Characterization of Frictional Behavior in Cold Forging

K. H. Jung H. C. Lee J. S. Ajiboye Y. T. Im

Received: 25 April 2009 / Accepted: 5 October 2009 / Published online: 15 October 2009 Springer Science+Business Media, LLC 2009

Abstract In the present investigation, tip test was utilized to characterize the effects of surface roughness of the specimen and forming tools, rate of deformation, and type of lubricants on friction in solid and solid contact under high contact pressure at room temperature. For the test, a cylindrical specimen made of aluminum alloy of 6061-O was used and grease, corn oil, VG100, and VG32 were applied as lubricants. Single punch and two counter punch sets with different surface roughness of Ra = 0.08 and 0.63 lm were manufactured in order to investigate a frictional behavior during the test. In addition, two different deformation speeds of 0.1 and 5.0 mm/s were used for the test to check their effect on friction as well. Load levels and tip distances obtained from the test were compared to nd out any correlation between the two. The change of surface topology of the specimen was monitored by optical measurement technique to better understand a frictional behavior at the punch and counter punch interfaces. Present investigation clearly shows that tip test is easy to apply to experimentally characterize the frictional behavior in cold forging under various processing conditions considered. Keywords Friction test method Forging Surface roughness Optical microscopy Forging uids

1 Introduction Friction is generated between two bodies in contact under a normal load and dened as the resistance to relative motion between the two. Conventionally, friction was described by Coulomb friction (s = lp, l = coefcient of friction and p = normal stress) or constant shear model (s = mks, m = shear friction factor and ks = shear yield stress of the material) known as Tresca friction model [1]. For the high contact pressure, the latter works better under the condition that m is characterized correctly [2]. It is well known that frictional behavior depends on processing variables such as contact area, deformation speed, temperature, surface condition, and environmental factors in general. Thus, it is not easy to characterize m correctly for general purpose. So far, ring compression test has been widely used for friction measurement because of its simplicity [3]. In this test, calibration curves introduced by analytical or numerical techniques should be employed to determine the m value by measuring dimensional change of the inner diameter. In spite of the extensive literature available on friction studies, an experimental technique for specically estimating the effects of processing variables such as deformation speed, surface roughness, and the type of lubricants on friction is rather limited [14]. Recently, Im et al. [5, 6] proposed a tip test based on a hybrid of simple compression and backward extrusion in which a radial tip was formed on the extruded end of the specimen to evaluate the friction and lubrication more accurately. A cylindrical specimen, whose diameter was larger than the diameter of the punch and smaller than the diameter of the cylindrical die including the counter punch as shown in Fig. 1, formed the tip in the test. Because of large free surface generation during the backward extrusion, this test can be more suitable for friction

K. H. Jung H. C. Lee J. S. Ajiboye Y. T. Im (&) Department of Mechanical Engineering, National Research Laboratory for Computer Aided Materials Processing, KAIST, 373-1 Gusongdong, Yusonggu, Daejeon 305-701, Korea e-mail: ytim@kaist.ac.kr

123

354

Tribol Lett (2010) 37:353359

Fig. 1 Schematic of the tip test investigated

Fig. 3 Comparison of positions of radial tips and forming loads during punch travel obtained from the simulation and experiment of tip test of AL2024-O with VG100 [7]

Fig. 2 Load versus radial tip distance curves depending on the value of x for the various friction conditions of mf p ranging from 0.0 to 0.9 [6]

measurement in metal forming compared to ring or other tests under the high contact pressure. According to these studies, notable feature of the tip test relies on the linearity among the shear friction factor, radial tip distance, and maximum load measured. In addition, the tip test characterized friction conditions at the punch and counter punch interfaces separately. It was also found out that the friction condition on the punch interface (mf p ) was determined to be higher than the one on the counter punch interface (mf d ) in the case investigated earlier based on numerical calibrations using the nite element analyses. In these works, the slope between the measured load and tip distance was dependent on the ratio x mf d =mf p as reproduced in Fig. 2 based on the result from reference [6]. Thus, the x value was quantitatively determined by matching the slope of the numerically calculated nondimensional load versus tip distance graph with the experimentally determined one in parallel depending on the

deforming material and given surface and processing conditions. Using the test setup developed by Im et al. [5], Kang et al. [7] determined a non-dimensional equation for the shear friction factor mf p at the punch interface according to the test results for the materials of copper, carbon steel, and aluminum alloys under various lubrication conditions with grease, drawing oil, soap, industrial mineral oils, and MoS2: mfp 4:35 d=t 0:95; where d was the tip distance and t the thickness of the extruded part when the surface roughness of the punch and counter punch was in the same order. They also correlated the x value as a function of the strain hardening exponent of the material under the given investigation conditions. In this work, experimentally determined positions of the radial tip from the inner wall of the container (at the ve selected strokes) and forming loads were compared reasonably well for the whole stroke with simulation results by employing different m values for the punch and counter punch interfaces as reproduced in Fig. 3. According to this gure, the present characterization scheme seems to be valid for the whole stroke. In these previous studies carried out so far, the punch velocity and surface conditions were maintained to be constant and similar, respectively. Also, the friction condition at the container was assumed to be the same as the one at the counter punch in numerical simulations [6, 7]. In the present study, an experimental attempt to characterize the inuence of surface roughness, deformation speeds, and type of lubricants on frictional behavior was made at room temperature in which the temperature effect can be negligible. For this study, the downsized tip test setup was designed and manufactured as shown in Fig. 1. The surface roughness of the counter punch was varied

123

Tribol Lett (2010) 37:353359 Fig. 4 Initial surface topologies of the rough and smooth counter punches measured by a whitelight scanning interferometer microscope

355

from Ra = 0.08 to 0.63 lm to examine variations of friction behavior and material ow in the test using AL6061-O depending on surface topology, while the surface roughness of the punch was maintained to be the same as 0.17 lm. Four different lubricants such as grease (AMSOIL GLC-12), cooking corn oil, VG100, and VG32 (Shell Tellus) were used in experiments to determine the effect of viscosity on friction. The ram velocity was also varied from 0.1 to 5 mm/s in experiments to check their effect on friction as well. To measure surface topologies of the billet, punch, and counter punch, optical measurements were made by employing a confocal microscopy (Olympus OLS-3000) and white-light scanning interferometer microscopy, respectively. The average values of measured forming loads and tip distances obtained from ve experiments were compared depending on the test conditions.

2 Experiment The tip test experiments were performed with the device illustrated in Fig. 1. The upper punch was moving downward at the speed of 0.1 or 5 mm/s up to the stroke of 3.1 mm, while the counter punch and container were stationary. The load was applied using a computer-controlled MTS machine (Alliance RT/100) with a maximum load of 100 kN. Commercially available AL6061-O bar with the diameter of 12 mm was fully annealed by heating from room temperature to 415 C and soaked at this temperature for 3 h according to Metals Handbook [8]. Then, the bar was cooled off at a cooling rate of 30 C/h to 265 C and nally exposed to air-cooling until reaching the room temperature, followed by machining into a cylindrical specimen size of 10 mm in diameter and 5 mm in height. Measurement of the dimension of the specimen with vernier calipers (Mitutoyo CD-15CPX) reduced a size irregularity, and the tolerance of specimens was maintained to be less than 0.02 mm. In earlier works [57], the diameter and height of the specimen used in experiments were 30 and 15 mm, respectively.

The punch and counter punch including the container used for the tip test were made of tool steel alloy AISI D2 and were chromium coated and polished. Figure 4 compares surface topologies of the counter punch used for the tip test in terms of arithmetic mean values of Ra = 0.08 or 0.63 lm, respectively. For the measurement, a confocal laser microscope (Olympus OLS-3000) and white-light scanning interferometer microscope were used. The arithmetic mean value of surface roughness of the specimen was also measured at the top (0.65 lm), bottom (0.65 lm), and circumferential (1.05 lm) surfaces by employing Olympus OLS-3000. Before applying each of the four kinds of lubricants of grease (AMSOIL GLC-12), cooking corn oil, VG100, and VG32 (Shell Tellus), surfaces of the punch, counter punch, and die were cleaned by a wiper (Kimtech Wipers) soaked with acetone (Junsei Assay (GC) min. 99%) to avoid adulteration among lubricants. After cleaning, application of a hair dryer blew away residue of acetone and each lubricant was brushed manually. The environmental factors such as temperature and humidity were carefully monitored during the test. For each lubricant, ve experiments were carried out. The tip distance d of the deformed specimen was measured as depicted in Fig. 1 by an optical microscope (Olympus B201), which has an extended focal imaging function to integrate pictures of different focuses. In the tip test, centering should be carefully monitored for achieving an axi-symmetric deformation. Since it is not easy to maintain the axi-symmetry perfectly in experiments, the tip distance was measured at four different points to determine the average value used in the present investigation.

3 Results and Discussion The measured load versus stroke curves is given in Fig. 5 for various experimental conditions. In this gure, the effects of the surface roughness, deformation speed, and type of lubricants were not noticeable at the upsetting stage of deformation. As deformation increases, beginning of the

123

356

Tribol Lett (2010) 37:353359

Fig. 5 Load versus stroke curves depending on the surface roughness of the counter punch and deforming speed for the lubricants of a grease, b corn oil, c VG100, and d VG32

backward extrusion indicated in an arrow in this gure occurred earlier for the rough counter punch case than in the smooth counter punch case due to the friction effect of surface roughness in terms of barreling. In the same gure, the case with VG32 and rough counter punch at the lower speed requires the highest load requirement as expected. For both surface roughness cases of the counter punch, load requirements were the lowest under the lubrication case with grease and VG100 for the deformation speed of 0.1 and 5.0 mm/s, respectively, in the present experiments. According to Fig. 5a, load levels were about the same for all the strokes for the grease irrelevant of the experimental conditions. For the higher ram velocity of 5 mm/s load levels were measured to be lower at the later stage of backward extrusion than those for the lower deformation speed of 0.1 mm/s in Fig. 5b, c, and d for the lubricants of corn oil, VG100, and VG32, more vividly in VG100 and VG32 when the stroke was higher than approximately 2.25 mm. This kind of load drop might be due to the larger hydrodynamic support of lubrication pockets for liquid lubricants at the boundary interfaces, especially largest for VG100.

According to these gures, measured load levels were interestingly almost the same between the rough and smooth counter punch cases for all the lubricants for higher and lower deformation speeds except for the cases of VG100 and 32 of lower deformation speed of 0.1 mm/s. This shows that load measurement alone might not be a good parameter to determine friction levels depending on the interfacial and processing conditions. In order to gure out load conversion for the case of VG32, the changes of surface topologies at the bottom surface of the specimen during two deformation stages are compared with the initial one at the same place in Fig. 6 for both deformation speeds. The patterns observed at the bottom of the specimen revealed the type of constraint encountered by the deforming material. It is well known that asperities in the specimen surface produced by machining prevented free deformation by the restraint of the underlying elastically stressed material [4]. Since the rough counter punch surface with Ra = 0.63 lm used in the present investigation was left as machined, interlocking of grooves between the specimen and counter punch surfaces is likely, thereby increasing resistance to free

123

Tribol Lett (2010) 37:353359 Fig. 6 Surface topologies of a the bottom of the initial specimen and deformed specimens at two different strokes under the lubrication of VG32 for deforming speeds, of b 0.1, and c 5 mm/s, respectively, observed by a white-light scanning interferometer microscope

357

deformation. As the load increases, new surface contact with no lubricant will be generated to increase the load level and lubricant pockets can also be formulated to create a hydrodynamic pressure to partly support the load for the liquid lubricants [9]. These two mechanisms will govern the load requirement during deformation. In the case of the smooth counter punch (Ra = 0.08 lm) for the lower deformation speed of 0.1 mm/s the asperities in the original specimen were reduced continuously in this

gure as deformation increased, resulting in higher new surface contact and smaller lubrication pockets leading to a higher load requirement at the later stage of deformation for VG100 and 32 as seen in Fig. 5c, d. For the case of the higher deformation speed of 5 mm/s, however, this kind of conversion between the rough and smooth counter punch surfaces was not noticeable owing to counter balance between the new surface contact and lubrication pocket generation. By comparing the changes of surface

123

358

Tribol Lett (2010) 37:353359

Fig. 7 Plots of normalized maximum load versus tip distance of the tip test result for AL6061-O specimen

topologies, it can be seen in Fig. 6 that the new surface contact increased and lubrication pockets decreased as deformation speed increased. Figure 7 shows the plots of experimental data of normalized maximum load (L/1000) versus tip distance (d/t) obtained from the tip test under various lubrication conditions and two different surface roughness of the counter punch. In this gure, the measured load and tip distance were normalized by 1,000 kN and tip thickness t of 1.214 mm, respectively. According to this gure, linearity between the normalized load and tip distance was reconrmed and change of the slope from positive (Ra = 0.08 lm) to negative (Ra = 0.63 lm) was observed due to the relative levels of friction between the punch and counter punch interfaces depending on the surface and lubrication conditions at the interfaces. The measured tip distances clearly demonstrated different aspects of frictional behavior depending on the processing and interfacial conditions in this gure. According to the nite element simulations using an inhouse program CAMPform [1012] to match the slope of L/1000 versus d/t graphs, x values were determined to be 0.7 and 3.5 for the smooth (Ra = 0.08 lm or positive slope, 0.2) and rough counter punch (Ra = 0.63 lm or negative slope, -0.44), respectively. In other words, the slope is notably dependent on the surface and processing conditions at the interfaces during the deformation and is determined to be positive when 0 \ x \ 1 and negative for x [ 1. This experimental observation obtained from the tip test is very unique in determining the relative magnitude of the friction at the punch and counter punch interfaces depending on the given surface and processing conditions. In Fig. 7, these two lines converge to one point which is close to the frictionless case ideally. According to the previous work by Im et al. [6], the normalized tip distance

d/t for the sticking condition at the punch interface was numerically calculated to be around 0.5 although it is not possible to achieve the fully frictionless and sticking conditions in the experiment. Thus, the normalized load requirement for the frictionless or sticking condition can be estimated in this gure as well. It is clear in Fig. 7 that measurement of the slope between the normalized maximum load and tip distance can be used as a determining parameter of the friction levels between the punch and counter punch interfaces without measuring their surface conditions based on this observation. If the measured value of the normalized tip distance is larger than the converging point value in this gure, then the friction at the punch interface is larger than the one at the counter punch interface and vice versa. Interestingly, the slope of the L/1000 versus d/t graph obtained from the current investigation for the positive case (0 \ x \ 1) was almost the same to the one obtained in the earlier investigations [57] in which the bigger specimen size (diameter 9 height = 30 9 15 mm2) was used in experiments. For the negative case (x [ 1) it was also the same to the one reported in the work by Chauviere et al. [13]. This reassures reproducibility of the test results for the linearity. However, further work is necessary to calibrate the present experimental observation to derive a nondimensional equation between mf p and d/t for both positive and negative slope cases.

4 Conclusions The effects of surface roughness, deformation speed, and type of lubricants on the normalized maximum load and tip distance are clearly observed in the current tip test investigation. It is demonstrated that the friction level of the counter punch interface due to the variation of surface roughness compared to the one of punch interface plays a key role for the slope shift from positive (0 \ x \ 1) to negative (x [ 1) during the test regardless of deformation speeds. The variation of tip distances due to surface roughness of the counter punch increased as the viscosity of lubricant and the deformation speed decreased. For the lubricant of grease, the variation of maximum load and tip distance was measured to be minimal among the cases tested. Owing to the effect of hydrodynamic lubrication pockets for liquid lubricants, load levels for the higher deformation speed decreased at the later stage of backward extrusion for both surface conditions compared to the lower deformation speed. From the current investigation, the tip test can be used for characterizing the effects of the surface condition and lubricant viscosity on frictional and ow behavior at various processing conditions under high contact pressure at room temperature. In most bulk forming

123

Tribol Lett (2010) 37:353359

359 5. Im, Y.T., Cheon, J.S., Kang, S.H.: Determination of friction condition by geometrical measurement of backward extruded aluminum alloy specimen. Trans ASME, J Manuf Sci Eng 124, 409415 (2002) 6. Im, Y.T., Cheon, J.S., Kang, S.H.: Finite element investigation of friction condition in a backward extrusion of aluminum alloy. Trans ASME, J Manuf Sci Eng 125, 378383 (2003) 7. Kang, S.H., Lee, J.H., Cheon, J.S., Im, Y.T.: The effect of strain hardening on frictional behavior in tip test. Int J Mech Sci 46, 855869 (2004) 8. ASM International: Metals Handbook Ninth Edition: Properties and Selection: Nonferrous Alloys and Pure Metals. American Society for Metals, Ohio (1979) 9. Bhushan, B.: Principles and Applications of Tribology. Wiley, New York (1999) 10. Kim, S.Y., Im, Y.T.: Three-dimensional nite element analysis of non-isothermal shape rolling. J Mater Process Technol 127, 5763 (2002) 11. Kwak, D.Y., Cheon, J.S., Im, Y.T.: Remeshing for metal forming simulationsPart I: two-dimensional quadrilateral remeshing. Int J Numer Methods Eng 53(11), 24632500 (2002) 12. Kwak, D.Y., Im, Y.T.: Remeshing for metal forming simulationsPart II: three-dimensional hexahedral mesh generation. Int J Numer Methods Eng 53(11), 25012528 (2002) 13. Chauviere, P., Jung, K.H., Kim, D.K., Lee, H.C., Kang, S.H., Im, Y.T.: Experimental study of miniaturized tip test. J Mech Sci Technol 22, 924930 (2008)

processes, the punch and dies are used for generating deformation which mostly consists of upsetting and extrusion mode. Thus, careful friction characterization using the tip test might be more benecial for general practice according to the present observations.
Acknowledgments This work was supported by the National Research Foundation of Korea through the National Research Laboratory program funded by the Ministry of Education, Science and Technology (No. R0A-2006-000-10240-0). Prof. Ajiboye appreciates the visiting scholarship from the Ministry of Education, Science and Technology through the BK21 program.

References
1. Dowson, D.: History of Tribology. Longman, London and New York (1979) 2. Wanheim, T., Bay, N.A.: A model for friction in metal forming processes. Annal CIRP 27(1), 189194 (1978) 3. Mulc, A., Kalpakjian, S.: Analysis of friction in ring compression: a factorial experiment. Trans ASME J Eng Ind 94, 1189 1192 (1972) 4. Schey, J.A.: Tribology. In: Lange, K. (ed.) Handbook of Metal Forming, pp. 6.16.25. McGraw-Hill, New York (1985)

123