Research Article

Design and optimization of multi‑stage manufacturing process

of stirling engine crankshaft
Mohsen Noorbakhsh1 · Hamid Reza Moradi1

Received: 21 October 2019 / Accepted: 28 November 2019 / Published online: 10 December 2019
© Springer Nature Switzerland AG 2019

Abstract
Crankshafts are among the most important parts in internal combustion engines, of which stirling engine is a useful
example. Manufacturing process of a crankshaft, is considered as a three-step forging process using preform, due to the
complexity in geometry. The most challenging step of the multistage forging process is to avoid stress concentration
and to create uniformity of strain by controlling metal flow. In the present study, the final part was achieved under three
manufacturing processes namely: upsetting, hot and cold forging. The models used in each manufacturing process are
designed by CATIA software. A finite element simulation on the basis of Cockcraft–Latham damage criterion was devel-
oped in DEFORM software. Using experiment design by Taguchi method, The optimization of manufacturing processes
were carried out by MINITAB software in two steps, in which the optimization objectives are considered as force, damage
and strain uniformity, and; input variables are taken as part-mold friction, pressing velocity and process temperature.
In order to find the most effective parameter of each manufacturing process, analysis of variance was conducted on
the results, in which, the most effective parameters in the upsetting, hot and cold forging processes were temperature,
friction and temperature, respectively.

Keywords  Finite element method · Optimization · Manufacturing · Design of experiment · Stirling engine

1 Introduction The process parameters involved in hot forging and

their role have been subject of investigation by [7, 8]. They
In recent years, significant efforts have been made to can be clustered under the following main groups [9]:
develop some new methods including FE method (FEM)
and intelligent control system to optimize upsetting pro- • Product geometry
cess. Quan et al. [1] Sukjantha et al. [2], Nuasri et al. [3], • Product material
Jeong et al. [4], and Quan et al. [5] analysed the influence • Tooling
of processing parameters on the upsetting process, pre- • Machine
dicted an optimum process condition, determined an • Process
optimal preform part and saved the secondary upsetting • Tool-work piece interface effects
defect, respectively by FEM. Liu et al. [6] introduced a new
computer-controlled upsetting system and solved the Junjia et al. [10] and Xing et al. [11] investigated micro-
underfill defect in next hot forging. structure distribution and mechanical properties of boron
alloy in hot forging. The complexity of global multi-objec-
tive optimization of every factor in the process is so high

*  Hamid Reza Moradi, Hamid_Reza_Moradi@yahoo.com | 1Department of Mechanical Engineering, Malek Ashtar University
of Technology, Tehran, Iran.

SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6

Vol.:(0123456789)
Research Article SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6

that several authors prefer to develop empirical expert 2 Methodology: designing the part
systems to assist in the design phase [12, 13]. and mold
Eventually, some researchers [14, 15] proposed to use
a sequential approximate optimization algorithm (SAO) Since the geometry of mold cavity is obtained based on
to optimize forging process, using the time-consuming the design, it will be considered as the first step of manu-
FEM simulation only to fit a meta-model of the process, facturing in the forging method. A precise and suitable
by Polynomial regression or Kriging interpolation. The design will result in cost savings, adequate quality of final
meta-model is used by the optimization algorithm that is part, waste reduction and mold life. In this study, firstly we
evaluated by simulating the optimum with FEM. intend to analyse and optimize the part’s geometry and
In a study by Espadafor et al. [16] failure analysis of a the fixed-end crankshaft mold. In Figs. 1 and 2, the geo-
generator crankshaft, using finite elemental simulations, metric design of one and both fixed-end part, and in Fig. 3
have been conducted and the points of the crankshaft the final part map are shown, respectively. As shown in
which have the maximum stress and are subjected to fail- the figures, the part having a crank with a diameter much
ure are identified. Chen et al. [17] in the field of crankshaft larger than the main axis of the crankshaft, which, as a
fatigue analysis, used the flexural fatigue test, the SAFL result, make the material distribution in this part to be
method, and statistical analysis to obtain fatigue limit. extremely unbalanced and, generating the final formation
Çevik et al. [18] have evaluated the performance of a diesel of this part can’t be achieved, therefore, using a preform
engine crankshaft fatigue during its forging process. Ktari is vital.
et al. [19] have conducted research on the fatigue failure The process of designing the part and mold are shown
of the crankshaft used in the train engine. in Fig. 4.
The aim of this study was to obtain the optimum values Forging part is usually designed based on machining
of process force parameters, part damage parameters and part and, various parameters of this design modification,
strain uniformity by conducting Taguchi method on the are such as adding gradients to the walls, corner radius etc.
model in which the variables were determined to be part the mechanical properties of the part based on standard
temperature, mold velocity and friction between part and conditions are given in Table 1.
mold. Meanwhile, the most effective parameters in each
manufacturing process were also determined. ` 2.1 Separation level place

Separation level is a line which separates the two upper

and lower sides of the mold and is shown as a separation
line at the pleated canal in the forged part.
In this study, the supposed part is a crankshaft and the
separation level of cantilever crank is shown in Fig. 5.

Fig. 1  Fixed-end crankshaft final part map

Vol:.(1234567890)
SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6 Research Article

Fig. 2  The cantilever crankshaft final part map

process. Therefore, the corners of the part are considered to

be rounded in order to facilitate the flow of materials, reduce
forging force and energy, reduce friction and prolong forg-
ing mold’s life. Values of radiuses are proposed from Ref. [21],
based on maximum altitude and with the help of the follow-
ing relations.
1
r= ⋅h h < 100 mm (1)
10

1
Fig. 3  Fixed-end crankshaft final geometric model
r= ⋅h 250 mm > h > 100 mm (2)
20
In which, h is the height of the part.
2.2 Dimensional consideration for machining In the Considered part, the maximum height is 161 mm
(crank length), so:
According to DIN7523 standard [20], the maximum
1 1
machining rate depending on the type and dimension of r= ⋅h= × 161 = 8 ⋅ 05 mm
20 20
cantilever and fixed-end parts is considered to be 5.4 mm.
The part is made of heat-treating steel and has a length Due to the presence of radius in some of outer corners
of 5.465 mm. of the crankshaft, the same value as 2 mm is used to design
the mold, otherwise the radius of the outer edges is 8 mm.
2.3 Mold wall gradient In Fig. 6, in addition to applying the angle of gradient, the
radius of corners and edges is also considered.
Mold wall gradient facilitates the removal of the part from
the inside of forging mold cavity. Regarding that the part is 2.5 Burr design
brought to the center when it cools down, the inner wall of
the part needs more gradient than the outer wall. There are many relations to design the dimensions of the
According to the DIN7523 standard [20], the external canal and the burr hole. In this study, the relation used by
gradient of 5.4° and internal gradient of 6° are considered Brochanov and Rebelski [22] has been used to design the
for both parts. canal and burr cavity.
√
2.4 The radius of corners and edges t = 0 ⋅ 015 Aw (3)

The sharp corners result in stress concentration and also

hampering in the flow of materials during the forging

Vol.:(0123456789)
Research Article SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6

Fig. 4  Flowchart of Forging
mold designing steps

Table 1  Mechanical properties of 30CrNiMo -(vcn200)—for parts

with a diameter of 40–100 mm
VCN 200 E (GPa) Hardness H Ultimate Yield strength
alloy (RC) tensile ­SY (MPa)
strength ­Sut
(MPa)

212 41 1400 950

Fig. 6  Adding edges and corners radius

In the above-mentioned equation, Aw is the cross-sec-

tional area of forged part ­(mm2) in the separation line and t
is the burr thickness (mm).
Timensions of burr canal are shown in Fig. 7. The rela-
Fig. 5  Dotted lines showing the crossing point of separation line
tions described in Ref. [20] have been used to design the
burr enclosure.
Tg = 1 ⋅ 6Tf (4)

Vol:.(1234567890)
SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6 Research Article

Fig. 7  Burr placement enclo-

sure

geometry of the crankshaft of a Stirling engine, before

r = Tf (5)
the final forging process, two preforms have been used
to manufacture the crankshaft. The preform is made
R = Tg (6) of ALSI_H_13, the first preform of upsetting process is
designed to smoothing the metal flow in the forging part
→ Aw = 27677 ⋅ 875 and the second preform of hot forging process is designed
to provide easier formation to reduce the applied force.
The preforms design is in reverse order, which means that
According to the Brochanov and Rebelski, wt for a part’s
we first design the second preform using the final shape
cross section in the previous section is equal to 3.66. As a of the part and then, we design the first preform based on
result, using interpolation technique, the values of Wg and the second one.
Tg from Table 5-4 of [20] are 33 and 5, respectively.
t = 2 ⋅ 89 ≈ 3 3.1 Hot forging process
Wf Hot forging is performed in the range of 900–1100 °C for
= 3 ⋅ 66 ⇒ Wf = 10 ⋅ 98 ≃ 11
Tf heat-treating steels (VCN 200). In order to design a pre-
form for forging parts, various methods such as co-poten-
The value of Aw for fixed-end crankshaft is 28,175 mm2.
tial lines, response level method, neural network, mass
According to Eq. 4, the burr thickness (t) is 2/92 = 3. As a
distribution method, etc. can be used. Since the investi-
result, W_g and T_g will be same as the values obtained
gated part is crankshaft and is common in industry, the
for cantilever crankshaft.
mass distribution method is used to design the preforms
in this part.
The mass distribution method has been used to obtain
3 Designing the crankshaft preform
an approximate form of preform in this design. In Fig. 8a,
b, the chart related to the forging part is depicted.
The desired part, which is the crankshaft, has a crank arm
with a height difference of about 50 mm on one side of
the part compared to the other parts of the crankshaft, 3.2 Upsetting process
thus having a great deal of geometry complexity and a
non-balanced material distribution, which itself compli- This process is similar to the forging process, except that
cates the design of the corresponding molds. On the other the press direction is along the length of the part and
hand, in the crank arm, a flow distribution of about 4 times reduction of the length. In this process, the pressing move-
greater than the other parts is required for the complete ment is usually along the longest side of the part. Upset-
filling of the mold, so the production of this part requires ting part is used when we want to apply a diagonal change
preform and middle mold. in a section of the part.
Besides, due to its application, crankshafts should In order to avoid buckling in the process of hot upset-
have high mechanical properties. Therefore, in crankshaft ting, relations (7) are used from Ref. [23]. In this relation, S
design, the last stage is cold forging, in order to achieve represents the ratio of initial length to the initial diameter.
the final target properties with high precision. The part value of s ≤ 3/2 can be formed in one step, other-
In this study, according to the explanations of the wise, two or more steps are needed to produce a well part
crankshaft design cycles, in the previous chapter and the without buckling.

Vol.:(0123456789)
Research Article SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6

A
14000
12000
Area (mm)

10000
8000
6000
4000
2000
0
0 100 200 300 400 500
part length (mm)

B Fig. 9  Meshing condition of cantilever crankshaft for upsetting

6000
5000
(upsetting) process
Area (mm)

4000
3000
2000
Table 3  The constant values of VCN 200 alloy
1000
0 Parameter Value Parameter Value
0 100 200 300 400 500 600
part length (mm) A 673 𝛼 0
B 1151 𝛽 0
C 0.029 𝜀̄̇ 0 1
Fig. 8  Forging part mass distribution diagram, A cantilever crank-
shaft, B fixed-end crankshaft D0 1 Troom 20
E 1 Tmelt 1527
n 0.31 Tb 0
m 0.49 K 0
Table 2  Simulation parameters of the upsetting and hot forging
processes using DEFORM software

30 CrNiMo8 mechanical properties

Process Heat Modulus of Poisson Density (kg/
tempera- transfer elasticity coefficient dm3)
ture °C coefficient (GPa)
(w/m k)

900–1150 33.7 139 0.3 7.80

Fig. 10  The shape and dimensions of work piece

s = l0 ∕d0
s≤2⋅3 onestage 4 Finite element simulation
s ≤ 4 ⋅ 5 twostage
(7)

s≤8 threestage For checking preforms, numerical simulation of the forg-

ing process was performed by DEFORM 10 software. The
For cantilever crankshaft, d0 = 80 mm and l0 = 161 mm. mechanical properties of the part in the processes of
s = 161∕80 = 2 ⋅ 0125 ≤ 2 ⋅ 3
upsetting and hot forging at the respective temperature
are given in Table 2.
So the number of upsetting steps is 1. The coefficient of friction between the work piece and
For fixed-end crankshaft, d0 = 90 mm and l0 = 400 mm . mold is 0.25 for upsetting process, 0.7 for hot forging and
With respect to relations (7), this preform is created in two 0.12 for cold forging. In Table 2, the mechanical proper-
stages of upsetting from the initial billet at each stage, ties of the part in hot processes are depicted. For a pre-
30 mm diameter change in work part is applied. cise control of the formation conditions, regarding that the

s = 400∕90 = 4 ⋅ 44 ≤ 4 ⋅ 5
process is isothermal, the press velocity was considered
to be 1 mm/s. Tetrahedral element type was used to mesh
the parts considering Fig. 9 in this research, the maximum

Vol:.(1234567890)
SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6 Research Article

Fig. 11  The simulation results of the first stage of the manufacturing process and the results of the damage parameter

Table 4  Comparison of the results of the present study and article 5 Fracture theory
Danno et al. work [22] This study
The intended fracture theory, with regards to using the
The effective strain 1.87 (mm/mm) 1.77 (mm/mm) finite element software DEFORM, is Cockcraft–Latham
of the first step of
forging
theory [24]. The critical damage parameter is calculated
The effective strain of 1.77 (mm/mm) 1.68 (mm/mm)
according to the Cockcraft–Latham relation, from the fol-
the second step of lowing equation:
forging
𝜀̄
Damage parameter 0.453 0.429 𝜎∗
∫ 𝜎̄
C= 𝜕 𝜎̄ (8)

force of the process and the effective tension at one point In this relation, 𝜎 ∗ is the maximum main tensile stress
were chosen for examination of the convergence of the of the material, 𝜎̄ is the effective tension during the
simulation results (sensitivity to mesh), and according to process and 𝜀̄ is the fracture strain. Since, most of the
the examination, the degree of meshing was considered processes for manufacturing of desired crankshaft, are
to be 35,000. considered as hot processes, the Johnson–cook model
is chosen for calculation in the software. This model is
shown in the following equation.

Table 5  Taguchi experiment Experi- Parameter 1 Parameter 2 Parameter Output Output Output parameter
design and the test results ment no. friction velocity 3 Temp parameter parameter Process force
designed for the upsetting Strain uni- Damage
process formity parameter

1 0.2 1 900 0.552 0.186 104,000

2 0.2 2 950 0.318 0.20 115,000
3 0.2 4 1050 0.356 0.2104 128,300
4 0.25 1 950 0.217 0.18 100,000
5 0.25 2 1050 0.316 0.222 95,600
6 0.25 4 900 0.436 0.20 121,000
7 0.3 1 1050 0.213 0.218 84,000
8 0.3 2 900 0.315 0.205 131,000
9 0.3 4 950 0.449 0.185 113,000

Vol.:(0123456789)
Research Article SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6

1.15 Johnson–Cook model are calculated by the following

NORMALIZED SIGMAL TO NOISE

1.1 relations, respectively.

1.05 ( )
1
∗
T − Troom
0.95
T =( )
Tmelt − Troom
0.9
0.85
( )𝛽
0.8 D = D0 expk T − Tb
1 2 4
Velocity
FORCE DAMAGE STRAIN The constant values of equations for VCN200 alloy are
presented in Table 3.
Fig. 12  Normalized signal-to-noise of velocity parameter in the By applying these values to the Johnson–Cook
upsetting process equation and simplifying it, the following equation is
obtained for the alloy used in the crankshaft:
1.1
𝜎̄ = 0 ⋅ 2106 673 + 1151−0⋅31 1 + 0 ⋅ 029ln 𝜀̄̇
( )( ( ))
NORMALIZED SIGNAL TO NOISE

1.05
The parameter 𝜎 ∗ is equal to k𝜀n , where k and n for this
1
alloy are 1240 and 0.61, respectively. By placing 𝜎̄ and 𝜎 ∗
0.95
in Cockcraft–Latham, this integral is obtained:
0.9

C=∫ (
𝜀̄
1240𝜀0⋅61
0.85 ) 𝜕𝜀
673 + 1151𝜀0⋅31 (1 + 0 ⋅ 029 ln (𝜀))
̇
0.8
0.2 0.25 0.3
FRICTION By placing the failure strain of VCN 200 alloy obtained
FORCE DAMAGE STRAIN from tensile test which were equal to 0.925, in the
abovementioned integral, the damage parameter value
Fig. 13  Normalized signal to noise of friction parameter in upset- is calculated as 0.74.
ting process

1.2
6 Verification
NORMALIZED SIGNAL TO NOISE

1.15
1.1 Based on the aforementioned method, a FEM simulation
1.05
1
was developed in which, the results were compared with
0.95 the study of Danno et al. [22]. In their research, the geom-
0.9 etry was taken as Fig. 10, its simulation was performed
0.85
0.8 using DEFORM-2D with the coefficient of friction, the press
0.75 velocity and the cushion force as 0.12, 0.1 mm/s and 17
0.7
KN, respectively.
900 950 1050
TEMPERATURE In Fig. 11a the simulation contours of paper for the first
FORCE DAMAGE STRAIN
step of forging are demonstrated, the performed simula-
tion results are shown in Fig. 11b and the results compari-
son are demonstrated in Table 4.
Fig. 14  Normalized signal to noise of temperature parameter in
upsetting process

( ( ))( )𝛼 7 Optimization
𝜀̄̇ 𝜀̄̇
𝜎̄ = A + B𝜀̄ n 1 + C ln
( ) ( )
D − ET∗m (9)
𝜀̄̇ 0 𝜀̄̇ 0 Since, one of the effective parameters in crankshaft frac-
ture is strain concentration during formation, it’s possible
The parameters A, B, C, D0 , E, α and m are constant
to reduce the probability of fracture in the part by improv-
values of the model and are different depending on
ing the strain distribution.
the type of material. The parameters T∗ and D in the

Vol:.(1234567890)
SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6 Research Article

Fig. 15  The table of variance

analysis of upsetting process’s
strain uniformity

Table 6  Taguchi experiment Experi- Parameter 1 Parameter 2 Parameter 3 Output Output Output parameter
design and test results ment no Friction Velocity Temp parameter parameter Process Force
designed for hot forging Strain Dis- Damage
process tribution parameter

1 0.3 1 900 0.781 0.488 422,000

2 0.3 2 950 0.691 0.51 361,000
3 0.3 4 1050 0.678 0.68 328,000
4 0.5 1 950 0.763 0.49 542,000
5 0.5 2 1050 0.721 0.81 311,000
6 0.5 4 900 0.756 0.84 583,000
7 0.7 1 1050 0.785 0.78 378,000
8 0.7 2 900 0.823 0.85 645,000
9 0.7 4 950 0.897 0.62 514,000

Therefore, optimization process has been performed to strain uniformity in each process. The design of experiment
achieve the lowest strain concentration in the final forging was done using Minitab software and Taguchi method. The
part. Optimization process is done in two steps. In the first optimization has been applied on a three-stage forging
step, the goal is to find the optimal levels of input param- of fixed-end crankshaft that involves upsetting process,
eters and also determine the most effective input on the hot and cold forging. Input parameters of the software

Vol.:(0123456789)
Research Article SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6
NORMALIZED SIGNAL TO NOISE 1.5 The analysis of ANOVA’s variance and GLM method are
1.4 used to find the most effective input parameters in each
1.3 process. In this research, the strain uniformity is obtained
1.2 by calculating the standard deviation of the strain param-
1.1 eter of all the elements extracted from the software.
1 In the second step, optimization was performed to
0.9 estimate the best conditions of the processes to acquire
0.8 the lowest uniformity of strain in the final part. Accord-
1 2 4 ing to the results obtained from the initial optimization,
Velocity the most effective parameter for the strain uniformity
FORCE DAMAGE STRAIN of each process is chosen. These parameters are inputs
of the second stage of optimization and the uniformity
Fig. 16  Normalized signal to noise diagram of velocity parameter in of strain in the final part, is considered the output of
hot forging process
this optimization.

1.5
NORMALIZED SIGNAL TO NOISE

1.4
1.3
8 Results and discussion
1.2
1.1 As mentioned above, the MINITAB software and Taguchi
1 test design are used to optimize the temperature, press
0.9
velocity and friction parameters. In Table 5, the experi-
0.8
0.7 ments and output values for upsetting process are repre-
0.6 sented. To obtain the strain uniformity, standard deviation
0.3 0.5 0.7
of all the effective strains belonging to all part’s elements
FRICTION
FORCE DAMAGE STRAIN
without considering the burr, has been calculated.
In Figs. 12, 13 and 14, the normalized signal-to- noise
Fig. 17  Normalized signal to noise diagram of friction parameter in diagrams relative to the mean value of vertical axis repre-
hot forging process senting the coefficient of influence, designed for investi-
gating the effect of each of temperature, press and fric-
tion inputs, respectively, on force, damage parameter and
1.6 strain uniformity outputs are depicted.
NORMALIZED SIGNAL TO NOISE

1.5 In Figs. 12, 13 and 14, the maximum strain uniformity

1.4
and minimum force and damage are optimization objec-
1.3
1.2 tives which are obtained by comparing the resulted opti-
1.1 mal values of response level in friction of 0.25 (second
1 level), press velocity (1 mm/s), and 1050 °C (third level).
0.9
To find the most effective parameter, as shown in
0.8
0.7
Fig. 15, analysis of ANOVA’s variance has been used and
900 950 1050 the most effective parameter in upsetting process was
TEMPERATURE obtained as temperature parameter.
FORCE DAMAGE STRAIN
Table 6 shows the experimental design and simulation
results of the hot forging process.
Fig. 18  Normalized signal to noise diagram of temp parameter in
hot forging process
In Figs. 16, 17 and 18, the normalized signal to noise
diagram for investigating the effect of input parameters
on the outputs of hot forging process is shown.
are considered for all three parameters of press velocity, Regarding the output diagrams, the optimal response
friction coefficient and the work piece temperature. Strain level of hot forging has been obtained as friction 0.3,
uniformity, damage parameter, and process force are also mold velocity 2 mm/s and the work piece temp 1050 °C.
considered as output parameters of processes. It should be noted that in the diagram of Fig. 17, con-
sidering the different behaviour of the two outputs of

Vol:.(1234567890)
SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6 Research Article

Table 7  Results of experiments Experi- Parameter 1 Parameter 2 Parameter 3 Output Output Output Param-
designed in the cold forging ment No Friction Velocity Temp parameter parameter eter process
process Strain Dis- Damage force
tribution Parameter

1 0.08 1 350 0.8072 0.89 60,200

2 0.08 2 400 0.81 0.33 62,400
3 0.08 4 450 0.83 0.22 67,500
4 0.12 1 400 0.88 0.78 69,500
5 0.12 2 450 0.77 0.54 72,900
6 0.12 4 350 0.8017 0.3 78,700
7 0.16 1 450 0.7571 0.43 69,200
8 0.16 2 350 0.7356 0.34 90,100
9 0.16 4 400 0.8718 0.17 99,000

1.8

NORMALIZED SIGNAL TO NOISE

1.4
NORMALIZED SIGNAL TO NOISE

1.6
1.3
1.4
1.2
1.2
1.1
1
1
0.8
0.9
0.6
0.4
0.8

0.2 0.7
1 2 4 0.08 0.12 0.16
Velocity FRICTION
FORCE DAMAGE STRAIN FORCE DAMAGE STRAIN

Fig. 19  Normalized signal to noise diagram of velocity parameter in Fig. 20  Normalized signal to noise diagram of friction parameter in
cold forging process cold forging process

strain uniformity and damage, in order to optimize the process, which is obtained as friction and temperature
process while the whole part remains unaffected, the parameters in hot and cold forging process, respectively.
strain uniformity parameter is more important in terms In the second step of optimization, according to the
of design and has a greater effect on choosing the rela- results obtained from the initial optimization, the most
tive optimal response level. effective parameter in each process is considered as the
In Table 7, the design of the test related to the cold variable parameter and the other two parameters are
forging process is shown, which has been designed in constantly equal to the related optimal response level
accordance with the 9 previous tests for optimization. values. In this step, all three processes are simulated
The investigation results of the parameters effects on sequentially and the strain uniformity of the final part is
the simulation outputs are shown in Figs. 19, 20 and 21. calculated after the three stages of manufacturing and
As mentioned in the previous section, due to the based on this optimal parameter, the overall response
greater importance of strain uniformity over damage in level, which represents the overall optimal response con-
the two parameters of friction and velocity, this factor sidering all three manufacturing processes, is obtained.
is taken into account for setting the optimal parameter According to Table  8, the upsetting temperature
therefore, the optimal response level for cold forging of 900 °C, the friction of 0.3 between mold and work
process is obtain as friction 0.16, press velocity 2 mm/s piece in hot forging process and cold forging process
and work piece temperature 450 °C. temperature of 350 °C are all obtained as overall opti-
The variance analysis is used to determine the most mal response level considering all three manufacturing
effective parameter in accordance with the upsetting processes.

Vol.:(0123456789)
Research Article SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6

Fig. 21  Normalized signal to 1.3

noise diagram of temperature

NORMALIZED SIGNAL TO NOISE

parameter in cold forging 1.2
process
1.1

0.9

0.8

0.7

0.6
350 400 450
TEMPERATURE
FORCE DAMAGE STRAIN

Table 8  Designed tests and Experiment Parameter 1 Parameter 2 Parameter 3 Output parameter

results obtained in strain no Part’s temp in upset- Friction in hot forg- Part temp in cold Strain uniformity
uniformity optimization ting process ing process forging process

1 900 0.3 350 1.13,343

2 900 0.5 400 2.14457
3 900 0.7 450 3.17794
4 950 0.3 400 3.19653
5 950 0.5 450 3.19664
6 950 0.7 350 3.19609
7 1050 0.3 450 3.19697
8 1050 0.5 350 4.20845
9 1050 0.7 400 3.19966

Fig. 22  A Hot forging process of a fixed-end crankshaft, B Cold forging process of a cantilever crankshaft

Vol:.(1234567890)
SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6 Research Article

Fig. 23  The cantilever crank force variations A Hot forging. B Cold forging

Fig. 24  Fixed-end crankshaft force variations A: Hot forging process B: Cold forging process

The effective strain obtained from hot forging pro- smoothing the strain distribution in part, the second pre-
cess of a fixed-end crankshaft and cold forging process form of hot forging process is intended to cause defor-
of cantilever crankshaft, based on the optimum final val- mation, and the final cold forging process is applied to
ues of temp, velocity and friction input parameters, are achieve the desired mechanical properties. The critical
shown in Fig. 22a, b respectively. value of the damage parameter for 30CrNiMo8 alloy was
Considering the simulation of all three manufactur- calculated to be 0.74, using Cockcraft-Latham relation.
ing stages and the two models, it can be concluded that The maximum damage values at each stages of forg-
the strain is the same in most parts of the crankshaft, ing, for designing processes of cantilever crank was
except the fillet region (the joint zone of axis and crank 0.54, 0.59 and 0.6 and for fixed-end crank was 0.22, 0.55,
arm). 0.58, respectively; which indicates that the desired part
The force variations diagram for hot and cold forging remains undamaged and integrated.
process of cantilever and fixed end cranks are depicted in Variation trends of input parameters versus objectives
the Figs. 23 and 24. represent the following relations: in upsetting process,
there is a reversal relation for velocity- strain uniformity
and direct relation for both friction- strain uniformity and
9 Conclusion temperature- strain uniformity. In hot forging process,
there can be seen a reversal relation for both velocity-
Due to the geometry of the model, the forging of Stirling damage and friction- strain uniformity and a direct rela-
engine crankshaft is designed in three stages, includ- tion for temperature- strain uniformity. In cold forging
ing two preforms and one final forging stage. The first process, there is a direct relation for velocity- damage,
preform of upsetting process is designed in order to friction- damage and temperature- damage.

Vol.:(0123456789)
Research Article SN Applied Sciences (2020) 2:65 | https://doi.org/10.1007/s42452-019-1820-6

Two optimization steps have been performed in which 4. Jeong H, Cho J, Lee NK, Park H (2006) Simulation of electric
for the first step, the variable parameters were considered upsetting and forging process for large marine diesel engine
exhaust valves. Mater Sci Forum 510–511:142–145
as: part temperature, mold velocity and friction between 5. Quan G-Z, Zou Z-Y, Zhang Z-H, Pan J (2016) A study on forma-
part and mold, and the objectives were taken as optimal tion process of secondary upsetting defect in electric upsetting
value of process force parameters, part damage param- and optimization of processing parameters based on multi-field
eters and strain uniformity belonging to all part’s ele- coupling FEM. Mater Res 19(4):856–864
6. Sun Y, Liu T, Zhang Z, Zhang T, Luo T (2003) Optimum control of
ments, using Taguchi method. The optimization results process parameters in electrical upsetting. Proc Inst Mech Eng
were obtained as follows: Part B J Eng Manuf 217(9):1259–1263
7. Grobaski TC, Mehta B, Gunasekera J (2004) Preliminary investi-
1. The values of friction between part and mold, mold gation into the effects of friction, work-piece temperature, die
temperature, and stroke velocity on hot forging die life. Depart-
velocity and part temperature for upsetting process ment of Mechanical Engineering, Ohio University, Athens
was equal to 0.25, 1 mm/s and 1050 °C, respectively. 8. Altan T, Ngaile G, Shen G (2004) Cold and hot forging: funda-
2. The values of friction, mold velocity and part tempera- mental and application, 1st edn. ASM International, Almere
ture for hot forging process was equal to 0.3, 2 mm/s 9. Samal CHANDAN (2014) Study of process parameters towards
improving efficiency of closed die hot forging process (Doctoral
and 1050 °C, respectively. dissertation)
3. The values of friction, mold velocity and part tempera- 10. Cui J, Lei C, Xing Z, Li C (2012) Microstructure distribution and
ture for cold forging process was equal to 0.16, 2 mm/s mechanical properties prediction of boron alloy during hot
and 450 °C, respectively. forming using fe simulation. Mater Sci Eng A 535:241–251
11. Xing Z, Bao J, Yang Y (2009) Numerical simulation of hot stamp-
ing of quenchable boron steel. Mater Sci Eng A 499(1–2):28–31
From F distribution and variance analysis, the most 12. Kulon J, Mynors DJ, Broomhead P (2006) A knowledge-based
important factor and the most effective parameter for engineering design tool for metal forging. J Mater Process Tech-
upsetting, hot and cold forging processes are obtained as: nol 177(1):331–335
13. Caporalli Â, Gileno LA, Button ST (1998) Expert system for hot
temperature, friction and part temperature, respectively. forging design. J Mater Process Technol 80:131–135
Considering all three manufacturing processes simul- 14. Bonte MHA, Do TT, Fourment L, van den Boogaard AH, Hué-
taneously, and performing sequential simulation with the tink J, Habbal A (2006) A comparison between optimisation
aim of maximizing the strain uniformity of the final part, algorithms for metal forming processes. In: ESAFORM 2006:
9th International ESAFORM Conference on Material Forming.
the overall optimal response level of the part was obtained Publishing House AKAPIT, pp 883–886
as the part temperature of 900 °C in upsetting process, 15. Bonte MH, Fourment L, Do TT, Van den Boogaard AH, Huetink J
friction of 0.3 in hot forging process and work piece tem- (2010) Optimization of forging processes using finite element
perature of 350 °C in cold forging process with final strain simulations. Struct Multidiscip Optim 42(5):797–810
16. Espadafor FJ, Villanueva JB, García MT (2009) Analysis of a diesel
uniformity of 1.1334 (mm/mm). generator crankshaft failure. Eng Fail Anal 16:2333–2341
17. Chen X, Yu X, Hu R, Li J (2014) Statistical distribution of crank-
Availability of data and materials  Data sharing not applicable to this shaft fatigue: experiment and modling. Eng Fail Anal 42:210–220
article as no datasets were generated or analysed during the cur- 18. Çevik R, Gül G (2013) Evaluation of fatigue performance of a fillet
rent study. rolled diesel engine crankshaft. Eng Fail Anal 27:250–261
19. Ktari A, Haddar N, Ayedi HF (2011) Fatigue fracture expertise of
Compliance with ethical standards  train engine crankshafts. Eng Fail Anal 18:1085–1093
20. DIN 2-752 (1986) Steel forging, design of drop and press forging,
machining allowances, drafts, edge radii, fillet radii, base think-
Conflict of interest  The authors declared that they have no conflict nesses, rib widths and rib radii, Deutsches Institut Fur Normung
of interest. 21. Byrer TG, Semiatin SL, Vollmer DC (1985) Forging handbook. Amer-
ican Society for Metals, Metals Park, OH, p 296
22. Tschätsch H (2006) Metal forming practise: processes-machines-
tools, vol 406. Springer, German
References 23. Lange K (1985) Handbook of metal forming. McGraw-Hill Book
Company 1985:1216
1. Quan GZ, Zhang L, An C, Zou ZY (2018) Multi-variable and bi- 24. Cockcraft MG, Latham DJ (1968) Ductility and workability of mate-
objective optimization of electric upsetting process for grain rials Inst Met 96:33–39
refinement and its uniform distribution. Int J Precis Eng Manuf
19(6):859–872 Publisher’s Note Springer Nature remains neutral with regard to
2. Sukjantha V (2013) Determination of optimal preform part for jurisdictional claims in published maps and institutional affiliations.
hot forging process of the manufacture axle shaft by finite ele-
ment method. King Mongkut’s Univ Technol North Bangkok
Int J Appl Sci Technol 6(1):35–42
3. Nuasri P, Aue-u-Lan Y (2017) Influence of process parameters
on electric upsetting process by using finite element model-
ling. Key Eng Mater 728:42–47

Vol:.(1234567890)