Progress in Natural Science: Materials International 26 (2016) 650–656

HOSTED BY Contents lists available at ScienceDirect

Progress in Natural Science: Materials International

journal homepage: www.elsevier.com/locate/pnsmi

Original Research

Experimental investigation and numerical simulation of plastic flow MARK

behavior during forward-backward-radial extrusion process
⁎
A. Farhoumand , R. Ebrahimi
Department of Materials Science and Engineering, School of Engineering, Shiraz University, Shiraz, Iran

A R T I C L E I N F O A BS T RAC T

Keywords: Finite element method was employed to investigate the effect of process parameters of plastic deformation
Finite element analysis behavior in Forward-Backward-Radial Extrusion (FBRE) process. The result of an axisymmetric model shows
Extrusion that the friction between die components and the sample has a substantial effect on the material flow behavior.
Forming Although strain heterogeneity index (SHI) slightly decreases with an increase in friction, large portion of the
Plastic deformation
sample experiences significant strain heterogeneity. Increasing the friction factor also localizes the strain
Strain heterogeneity
heterogeneity effect in the backward section, and spread the effect in the forward section. Decreasing the friction
in the FBRE process can reduce the amount of the strain heterogeneity in the product while decreases the
required punch force substantially. Furthermore, an increase in gap thickness increases the deformation in the
area close to the lower punch at the expense of the area in the vicinity of the upper punch. The numerical
simulation has a good agreement with the experimental results which confirms the accuracy of the proposed
finite element model.

1. Introduction In a study by Kim et al. [10], the effect of friction on the material
deformation during equal channel angular pressing process was
Extrusion in comparison to other manufacturing methods used in investigated using a 2D finite element model. It was found that friction
industrial application has many advantages such as: minimum material intensifies the shear deformation for the surface elements. This is due
waste, high dimensional accuracy, reduction or complete elimination of to substantial effect of friction acting on the opposite direction of the
machining, good surface finish, better mechanical properties of pro- moving surfaces during the process. Thus, the friction can decrease the
ducts than those of the parent material. The basic processes involving extent of the less-deformed shared zones in the process.
cold extrusion are classified based on their forming direction as Altan et al. utilized finite element analysis of an axisymmetric
forward, backward and radial/lateral extrusion [1]. Radial extrusion model for a deep cup drawing process to investigate the effect of
process can be used to manufacture complex parts such as collar friction [11]. It was concluded that the variation in the friction
flanges, spur gears, splines with shafts and tube fittings [1–3]. coefficient over a small range, in which the coefficients are close to
Besides, combinations of extrusion processes in which a billet is the actual values, does not significantly affect the deformation, as long
extruded simultaneously in forward, backward and radial directions as the material is not within the instability limit [11]. In contrary where
can also facilitate to eliminate the need for multistep forming of material instability is likely to occur due to the large tensile stresses
relatively complex shaped parts [4]. For instance backward-forward associated with a great punch force or a large blank diameter or small
extrusion [5], radial forward extrusion [6,7], radial backward extrusion sheet thickness the sensitivity of the process to changes in the
[4] and double backward extrusion [8] are some of these processes. conditions of friction becomes significant. Farhoumand et al. employed
Since most of the components are produced on the basis of a 3D finite element analysis of a novel extrusion process for quantita-
experience and trial-and-error [5], it is imperative to eliminate the tive assessment of strain accumulation in relation to the process
unnecessary production cost by modeling the process and optimizing parameters [12,13]. More recently, in an effort to increase the
the parameters. The significance of an analysis for a forming process accumulated plastic strain and obtain a more uniform distribution, a
lies in the determination of required punch force, flow behavior as well modified backwards extrusion was proposed by Shatermashhadi et al.
as stress-strain state during the process [9]. [14] and, Finite Element (FE) analysis results confirmed the successful

Peer review under responsibility of Chinese Materials Research Society.

⁎
Corresponding author.
E-mail address: [email protected] (A. Farhoumand).

http://dx.doi.org/10.1016/j.pnsc.2016.12.005
Received 20 September 2015; Received in revised form 5 November 2016; Accepted 13 December 2016
Available online 23 December 2016
1002-0071/ © 2017 Chinese Materials Research Society. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
A. Farhoumand, R. Ebrahimi Progress in Natural Science: Materials International 26 (2016) 650–656

implementation of the modified process. In another research by

Ebrahimi et al. FE analysis was employed to investigate the uniformity
of strain distribution in extrusion of copper with optimization of
various sample and process geometries [15]. Furthermore, numerical
simulation of copper deformation in compression-extrusion was per-
formed by Babaei et al. using commercial FE software ABAQUS [16].
Forward-backward-radial extrusion process (FBRE) can be used to
manufacture components that feature a hoop protrusion like gears and
flanges that come with hollow shafts [17]. A quantitative study of FBRE
was performed to investigate the effect of process parameters on the
final dimensions of the product. But, the effect of process parameters
on the material flow behavior, strain distribution and thus the strain
heterogeneity within the final product has not been investigated. It is
imperative to have a clear understanding of the strain distribution
within the final product which can affect the mechanical properties and
microstructure of the product.
Hence, in this study, the effect of processing parameters on strain
heterogeneity within the product processed by FBRE was investigated
by implementation of FE method. Quantitative analysis of strain
heterogeneity was performed by utilizing an axisymmetric FE model.
A number of numerical simulations were performed and the results
Fig. 2. The schematic of the die assembly used in FBRE process and the geometrical
were compared with experimental work to verify the accuracy of the FE parameters (gap and wall thickness and die corner fillet).
model.
its axi-symmetrical nature similar to that of the double cup extrusion
2. Finite element analysis process. Several parameters in FBRE can influence the material
deformation behavior. Friction as a process parameter along with
An Axisymmetric FBRE model was analyzed in ABAQUS FE soft- geometrical parameters such gap thickness and die corner fillet can
ware [18] using explicit algorithm. Punches and dies were assumed to influence the material deformation behavior in FBRE. Fig. 2 illustrates
be rigid due to negligible elastic deformation whereas the sample the aforementioned parameters schematically.
material (commercially pure aluminum) was considered as deformable Several simulations were performed to analyze the effect of friction
in the model. The kinematic relation of the sample was obtained from a and geometrical parameters on the FBRE deformation behavior and
compression test, Fig. 1, which represented by a power law equation as the corresponding punch force.
σ=133ε 0.3 (MPa).
The kinematic behavior in FE model was incorporated with the von 3. Experimental procedures
Misses yield criterion and isotropic hardening. The contact condition
between sample and surfaces of die and punches were assumed to To perform the FBRE process experimentally, a proper die set was
follow coulomb friction law. Barrel compression test [19] was utilized designed and manufactured. The punches and dies were machined
to measure the experimental friction factor (m). The experimental from cold-worked alloy steel (Grade: X210Cr12 No. 1.2080) while the
friction factor (m) was found 0.13. However, it should be noted that the other die components were manufactured from medium carbon steel
acquired value is only pertinent to the performed FBRE process, used (Grade: CK45 No. 1.1191). The heat treatment of alloy steel parts was
lubricant and the existing surface condition of the FBRE die compo- performed at 970 °C followed by an oil quenching and subsequent
nents and sample. Due to the limitations in ABAQUS software input tempering at 250 °C. The assembly of the die setup under a screw press
variables, Eq. (1) was used to convert the measured friction factor (m) is shown in Fig. 3(a). The die was designed with an adjustable gap
to coefficient of friction (μ ) for numerical simulations, thickness that eliminates the need for additional die sets. The gap
thickness has been shown by the arrow in Fig. 3(b).
m0.9
μ= Cylindrical samples with 24 mm diameter and 20 mm in length
2.72(1 − m )0.11 (1)

Although Eq. (1) has been derived for double cup extrusion process
[20], this equation is considered to be applicable to FBRE process due

Fig. 1. Before and after a compression test of a commercially pure aluminum sample, Fig. 3. FBRE setup under a screw press in (a) and the gap between upper and lower dies
utilized to obtain the kinematic of the material. in (b).

651
A. Farhoumand, R. Ebrahimi Progress in Natural Science: Materials International 26 (2016) 650–656

Fig. 4. Cross section view of FBRE processed aluminum samples for gap thicknesses of
2 mm (a) and 4 mm (b).

were machined from commercially pure aluminum (grade 1100). FBRE

process was performed in a 200 kN capacity screw press machine with
the crosshead speed of 0.2 mm/sec at ambient temperature.
The die corner fillet and the wall thickness in the experimental
setup were 1 mm and 2 mm, respectively, while samples with two
different gap thicknesses of 2 mm and 4 mm, were processed. The cross
sections of the processed samples are shown in Fig. 4.

4. Results and discussion

Variation of frictional conditions during forming processes can

significantly alter the rate of strain hardening and subsequently the
mechanical properties of different sections within the product.
Considering the application of the product, these variations of mechan-
ical properties within the product shall be thoroughly examined to
assure its compliance with the engineered design limits.
Hence, in case of FBRE product and considering the presence of
three different sections, namely forward, backward and radial, this
frictional variation needs more in depth investigation. Therefore,
analysis of plastic strain distribution in different sections of FBRE
product was performed in several frictional conditions. The quantita-
tive assessment was fulfilled by defining Strain Heterogeneity Index, Fig. 5. Meshed model with defined paths for various sections for the strain heterogeneity
SHI as below, analysis (Un-deformed and Deformed Mesh geometry on the left and right of
Axisymmetric axis, respectively.).
(εmax − εmin )
SHI =
εave (2) increases from 0 to 0.13, the location of the maximum SHI in Fig. 6(a)
where εmax , εmin and εave denote the maximum, minimum and shifts towards point B. With further increase of friction factor, m=0.8,
average effective plastic strain over a path, respectively. the maximum SHI for backward section increases significantly which
Different paths in cross section of the FBRE sample were defined shows the significant effect of friction on deformation behavior in
and mapped in the deformed mesh as shown in Fig. 5. For the backward section of FBRE sample. Furthermore, not only the SHI
backward section, paths were defined along the wall thickness of the increases with increase of frictional conditions, but also the height of
simulated deformed mesh with 1 mm spacing. As shown in Fig. 5, a the material in backward section decreases. Therefore, not only an
typical path in backward section is from point ai to bi while the range of increase in friction factor in backward section intensifies the strain
i was chosen according to the extent of this section from point A to B. heterogeneity but also it restricts the material progress into backward
For each frictional condition, SHI was calculated along each path by section. In frictionless condition, the SHI in forward section, Fig. 6(b),
using the FE simulation results as per Eq. (2). is much more intense in the vicinity of the lower punch, (point C in
Similar approach was applied for the forward section as paths were Fig. 5). However, as the friction factor increases, the SHI decreases
annotated as cj to dj from point C to D as shown in Fig. 5. slightly, but wider section of the sample experiences significant strain
For the center section, only one path was considered on the heterogeneity. It was also observed that an increase in friction factor
axisymmetric axis of the sample from the upper punch (point E) promotes the material progress into forward section which is in
towards the lower punch (point F) as indicated in Fig. 5. contrast to that of backward section, as previously discussed.
Therefore, an increase in friction in FBRE process has opposing effects
for backward and forward sections. Increasing the friction factor will
4.1. Effect of friction between sample and the surfaces of die and promote the material progress into forward section at the expense of
punches in FBRE process the backward one. Besides, increasing the friction factor narrows the
strain heterogeneity curve in backward section while increasing its
The effect of different frictional conditions on the heterogeneity of maximum which is in contrary for that of forward section.
the accumulated strain in backward and forward sections of FBRE For the center section, the equivalent plastic strain for various
sample is depicted in Fig. 6(a) and (b), respectively. frictional conditions has been depicted in Fig. 7. When the frictionless
In frictionless case SHI is the maximum in the vicinity of point A conditions prevails (m=0), the accumulated strain is higher in the
which rapidly declines towards point B. This strain heterogeneity vicinity of the upper and lower punches (point E and F) in comparison
adjacent to point A is due to intense deformation at the corner of the to that of middle section (normalized distance 0.4–0.8 in Fig. 7). As the
punch which causes localized strain heterogeneity. As the friction factor

652
A. Farhoumand, R. Ebrahimi Progress in Natural Science: Materials International 26 (2016) 650–656

Fig. 6. Effect of friction on strain heterogeneity, (a) in backward section and (b) in forward section of a FBRE sample.

Fig. 9. Effect of friction factor on the FBRE punch force.

Fig. 7. Effect of friction on the equivalent plastic strain at center region of a FBRE
processed sample.

Fig. 8. Equivalent plastic strain contours (PEEQ) and dead zones in a FBRE sample. Cross section views of the sample for a friction factor of (a) m=0, (b) m=0.13, (c) m=0.38 and (d)
m=0.8.

653
A. Farhoumand, R. Ebrahimi Progress in Natural Science: Materials International 26 (2016) 650–656

Fig. 10. Equivalent plastic strain (PEEQ) contours in different stages of a sample formation during frictionless (m=0) FBRE at different punch strokes (in mm) of (a) 0, (b) 1, (c) 1.5, (d)
2.5, (e) 6 and (f) 15.

3
Gap Thickness = 1 mm
Gap Thickness = 2 mm
2.5 Gap Thickness = 4 mm
Equivalent Plastic Strain, ε

1.5

0.5

0
0 0.2 0.4 0.6 0.8 1
Fig. 11. Effect of gap thickness on the FBRE punch force, simulation and experimental Nomalized Distance (From E to F)
results.
Fig. 12. Effect of gap thickness on the equivalent plastic strain in the center section of
friction factor increases, the plastic strain in the vicinity of the punches FBRE processed samples.

decreases significantly which give rise to substantial plastic strain

accumulation in the middle of the center section. This shows a shift in limits the material flow into the backward section. Consequently, the
material flow preferences as the friction factor increases. In frictionless restriction of material flow into backward section, promotes the
condition, the material plastic flow occurs in two discrete regions in the material flows into forward section and therefore, the difference
vicinity of upper and lower punches. As friction increases, these two between the height of forward and backward sections decreases as
discrete regions form separate dead-zones with no further material the friction factor increases.
flow while one central region in the material becomes responsible for The effect of friction on the required FBRE punch force is also
further plastic flow of the material during FBRE. investigated as shown in Fig. 9. The difference in punch force between
The extent of dead-zones adjacent to the upper and lower punches friction less condition and m=0.13 case is not significant. This is due to
for various frictional conditions is shown in Fig. 8(a)–(d). Regardless of similarity of the deformation behavior in the aforementioned frictional
the friction factor, the extent of the dead zone adjacent to the upper conditions. As the friction increases further to m=0.38, the material
punch is much more noticeable than that of the lower punch. But, as plastic flow localizes in central areas, which in turn increases the FBRE
the friction factor increases this difference abates. required load as can be seen in Fig. 9.
It can be established from Fig. 8 that as the friction factor increases, Further increase of the friction from m=0.38 to m=0.8 has a
the extent of the dead zone adjacent to the upper punch extends which significant effect on the increase of punch force. This increase could

654
A. Farhoumand, R. Ebrahimi Progress in Natural Science: Materials International 26 (2016) 650–656

2 r = 1 mm large number of infinitesimal time steps each corresponding to a small

r = 3 mm stroke increment. An incremental displacement field solution was
1.8 r = 5 mm obtained for each time step while the total strain corresponding at a
given time was calculated by integrating the strain rate along the
1.6
Equivalent Plastic Strain , ε

deformation path. Hence, the error/noise, propagated incrementally

1.4 when the effective strain was recovered via integration during the
FBRE process. This could also be correlated to the noise in the
1.2 calculation of the punch force for FBRE as shown in Fig. 11.
The effect of gap thickness is also investigated on the equivalent
1 plastic strain in center section of the sample as illustrated in Fig. 12.
When the gap thickness is 1 mm, the equivalent plastic stain is
0.8
homogenously disturbed in the sample and only areas in the vicinity of
0.6 the lower and upper punches, Point E and F which corresponds to
normalized distance 0 and 1, respectively, experiences less deforma-
0.4 tion. As the gap thickness increases to 2 mm, an obvious change in
material flow behavior can be seen from Fig. 12. A narrow gap
0.2 thickness, for instance 1 mm, restricts the material flow into the radial
section. Hence, the material tends to flow more into backward and
0 forward sections. As the gap thickness increases, the material tends to
0 0.2 0.4 0.6 0.8 1
Normalized Distance (From E to F) progress more into radial section due to an easier flow. Hence, the area
closer to the lower punch, flows in both forward and radial directions
Fig. 13. Effect of die corner fillet on the equivalent plastic strain in the center section of and therefore accumulates more plastic strain inside the material. This
FBRE processed samples.
on the other hand, causes the restriction of the material flow in the
vicinity of the upper punch hence causing lower strain accumulation in
be solely due to redundant frictional losses along die/sample. backward section. Therefore, as the gap thickness increases, the area in
interface since there is no significant change in material flow the vicinity of lower punch experiences more deformation at the
behavior between m=0.38 and 0.8. expense of the area in the vicinity of the upper punch which
Hence, increase of FBRE punch force with an increase in friction corresponds to point F and E in Fig. 11, respectively.
factor can be the contribution of two factors. Firstly, increasing the
friction is increasing the required redundant work to overcome the 4.3. Effect of die corner fillet in equivalent plastic strain in FBRE
frictional force between the die surfaces and the material. Secondly, process
increasing the friction in FBRE causes a shift in material flow to be
restricted in the central regions which also limits the deformation The effect of die corner fillet on the equivalent plastic strain in the
hence giving rise to an increase of required punch force. It can be
center section is not as significant as that of the gap thickness. As
deducted that the effect of the former is much more pronounced than illustrated in Fig. 13, a change in die corner fillet from 1 to 5 mm does
that of the latter on FBRE punch force requirements. not change the material flow behavior significantly.
Equivalent plastic strain contours of the sample in frictionless
conditions are shown in Fig. 10 in six different stages of FBRE. It could 5. Conclusion
be seen that the highest values of strains occur where the maximum
deformation exists. The plastic equivalent effective strain is higher in A reliable axisymmetric FE model was developed to investigate the
the areas where the sample is in contact with the punch corners than effect of process parameters on the material flow behavior in FBRE,
anywhere else. In addition, at the initial stage of the process, contours and the results of which were verified experimentally. It is found that
of strains are distributed in extensive areas which gradually decrease the friction between the die components namely punches and dies, and
towards the end of the process. sample has a significant effect on the material flow in FBRE. The strain
heterogeneity index, SHI, was defined for quantitative analysis of
4.2. Effect of gap thickness in punch force and equivalent plastic frictional effect on plastic deformation. An increase in friction factor
strain in FBRE process increases the SHI for both backward and forward sections. But, intense
friction confines the extent of strain heterogeneity in backward section,
Another geometrical parameter that was considered to influence the while it has an opposite effect for that of forward section. Besides, the
material deformation behavior in FBRE is gap thickness. The effect of minimization of friction in the FBRE process reduces the required
gap thickness on the accumulated effective plastic strain in center punch force significantly. As the gap thickness increases, the area in the
section is shown in Fig. 11. vicinity of lower punch accumulates more plastic strain in comparison
During the FBRE, similar to conventional extrusion, the punch to that of the upper punch. Hence, more uniform deformation is
force increases significantly with punch stroke at the start of the achieved with smaller gap thickness. The presented FE model can be
process while after certain punch stroke, the required extrusion force used to predict the plastic deformation behavior within the process
reaches a steady state. As can be seen from Fig. 11, regardless of the parameters which in turn can dictate the mechanical properties of the
gap thickness, the location of this transition in punch force is somehow FBRE process products.
constant which occurs at point of about 8 mm of punch stroke. Besides,
as the gap thickness decreases the punch force increase, which is due to References
the restriction of the material flow into the gap between the upper and
lower dies. Reasonable agreement between the simulations results and [1] B.D. Ko, et al., J. Mater. Process. Technol. 113 (1–3) (2001) 109–114.
the experimental data confirms the validity of the finite element model [2] Y. Çan, T. Altınbalık, H.E. Akata, J. Mater. Process. Technol. 166 (1) (2005)
128–134.
and the accuracy of the simulation results. [3] Y. Can, C. Misirli, Mater. Des. 29 (4) (2008) 829–838.
The proposed grid size in the axisymmetric model gave a con- [4] H.-J. Choi, J.-H. Choi, B.-B. Hwang, J. Mater. Process. Technol. 113 (1–3) (2001)
vergent and stable solution while minimizing the computational cost. 141–147.
[5] H.Y. Cho, et al., J. Mater. Process. Technol. 135 (2–3) (2003) 375–381.
To solve the problem, the punch displacement is typically divided into a

655
A. Farhoumand, R. Ebrahimi Progress in Natural Science: Materials International 26 (2016) 650–656

[6] Y.S. Lee, et al., J. Mater. Process. Technol. 113 (1–3) (2001) 136–140. [13] A. Farhoumand, P.D. Hodgson, S. Khoddam, 48 (6) (2013) 2454–2461.
[7] R. Ebrahimi, M. Reihanian, M.M. Moshksar, Mater. Des. 29 (9) (2008) 1694–1700. [14] V. Shatermashhadi, et al., Mater. Des. 62 (0) (2014) 361–366.
[8] A. Buschhausen, et al., J. Mater. Process. Technol. 33 (1–2) (1992) 95–108. [15] M. Ebrahimi, F. Djavanroodi, Prog. Nat. Sci.: Mater. Int. 24 (1) (2014) 68–74.
[9] D.Y. Yang, K.J. Kim, J. Mater. Process. Technol. 191 (1–3) (2007) 2–6. [16] A. Babaei, M.M. Mashhadi, Prog. Nat. Sci.: Mater. Int. 24 (6) (2014) 623–630.
[10] H.S. Kim, M.H. Seo, S.I. Hong, J. Mater. Process. Technol. 113 (1–3) (2001) [17] A. Farhoumand, R. Ebrahimi, Mater. Des. 30 (6) (2009) 2152–2157.
622–626. [18] V. Abaqus/Explicit, Dassault Systèmes Inc., USA, 2009.
[11] H. Darendeliler, T. Altan, J. Mater. Process. Technol. 58 (2–3) (1996) 293–301. [19] R. Ebrahimi, A. Najafizadeh, J. Mater. Process. Technol. 152 (2) (2004) 136–143.
[12] A. Farhoumand, S. Khoddam, P. Hodgson, Model. Simul. Mater. Sci. Eng. 20 (8) [20] S.H. Molaei, M. Shahbaz, R. Ebrahimi, Iran. J. Mater. Form. 1 (2) (2014) 14–22.
(2012) 085005.

656