Fatigue Analysis of Stripper Bolt

Under Combined Loading for

Improvement of Stamping Die
Design
Sang Hoon Lee A common problem of fatigue failure of stamping dies was experienced during the
Principal Project Engineer, stamping operation with socket-head screws. In order to establish a design standard
MacNeai-Schwendler Corp., for the stripper bolt, a methodology for determination of the loads and the fatigue
Los Angeles, CA 90041 strength of the stripper bolt was developed. Stresses due to an impulsive load and a
Mem. ASME rectangular pulse were calculated based on a simplified spring mass system and the
appropriate corrections were made to elaborate the solution. This approximate solu-
tion was validated by a finite element analysis. The stripper bolt should have an in-
finite fatigue life to survive a half million stamping operations. The fatigue problem
involves a stress concentration with combined mean and alternating stresses. The
Gerber parabola and the residual stress method were employed to treat the combin-
ed loading and the stress concentration. In order to enhance the fatigue life of the
stripper bolt, a cushion is introduced at the surface where an impact load is applied.
The cushion is found very effective in improving the fatigue life of the stripper bolt.
An interactive computer program was developed as a tool for designing stamping
dies.

1 Introduction
Fatigue failure in stripper bolts has been a serious disrupting The effects of these load sources are as follows.
factor in the production line. Socket-head screws are used for
clamping the stripper of the stamping dies as shown in Fig. 1. Preload (Fp). The preload is created by the torque to clamp
The purpose of this study is to establish a design standard for the stripper against the sleeve. The preload is a static force in
stamping dies which consist of stripper bolts. A simplified the bolt and proportional to the applied torque. For Ml2
method for determining stresses and the fatigue strength of the screw in 8G:
stripper bolt is developed and verified. Fp = 3.7534 x l 0 4 N
For a prototype analysis of the stripper bolt under the Applying the effective stress area (As) of 84.3 mm 2 , the static
stamping operation, a 1/4 model of a typical stripper bolt axial stress in the stripper bolt is
subassembly, consisting of an M12 socket-head screw, a
washer, a sleeve and a stripper, is constructed as shown in Fig. F
2. It is assumed that the die set consists of four stripper bolts o-„=—£- = 4.452 x 10 8 N/m 2
with the stripper mass being uniformly distributed. The details Impulsive Load OF,). The impulsive load is induced from the
of the model are determined based on VW design data and the impact of the stripper momentum due to sudden change in
standard DIN screw specifications for class 8G screw [2]. velocity of the stripper, i.e.
2 Loading Conditions / = \T0F,(t)dt = m(vT - v„) = mAv (1)
The loads applied to the stripper bolt are derived from In order to determine Ay of the stripper, the ram motion is ex-
various sources as below: pressed as
X=Rrsmu.t. (2)
• torque for clamping
• sudden lift of the stripper Assuming the crank radius (Rc) to be 0.06 m with an angular
• compression of the stripper spring velocity of 40 rpm (wr = 4.189 rad/s), the ram motion with
• reciprocating motion of the ram respect to time is shown schematically in Fig. 3. Further,
• gravity assume that the stripper has a maximum stroke (C/max) of
0.019 m, then the sudden lift of the stripper occurs at
Contributed by the Technical Committee on Vibration and Sound for
publication in the JOURNAL OF VIBRATION, ACOUSTICS, STRESS, AND RELIABILITY
IN DESIGN. Manuscript received by the Committee, September 20, 1984.
B -L[ 2 ,-.m-.(i--^)] = i.32, (3)

222/Vol. 108, APRIL 1986 Transactions of the ASME

Copyright © 1986 by ASME
Downloaded From: http://vibrationacoustics.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jvacek/28969/ on 06/02/2017 Terms of Use: http://www.asme.org/a
 "sw,^W)^'iu\u'in

M12 BOLT

SLEEVE -

•18*
90
•-13*
••12*

J-
STRIPPER

30

120*

H-130- WASHER
y if* U*- 30 0 -

12 11.33
11.55*
J_L
•18*-
NOTES: Upper die 5. Stripper
Stripper bolt 6. Stripper spring
Sleeve 7. Guide pin
Washer Fig. 2 Stripper bolt model
Fig. 1 Stripper assembly

L.JL.XQ : Ran m o t l i

Hence,
vT=Rccar cosO>,T = 0. 1832m/s (4)
with v„ = 0.
For the given geometry of the stripper bolt model, the im-
pulse momentum is determined to be 0.4749 kg-m/s with the
stripper mass (Ms) of 2.592 kg.
Periodic Rectangular Pulse (Fr). The compression of strip-
per springs (Fs) and the gravity of stripper mass (Fg) create a
periodically pulsating load in rectangular shape on the stripper
bolt, as shown in Fig. 4. The contributions of the stripper
spring and the gravity are estimated to be 334.4 N and 25.45
N, respectively, with spring stiffness ks = 41.8 x 103 N/m,
and the initial deflection 60 = 0.008 m. Then, the rectangular LWER DIE

pulse has the peak magnitude of 359.85 N and the duration of Fig. 3 Motion of the ram-die assembly at 40 rpm
1.11s with a period of 1.50 s.
The portion of the external load which is balanced by each of
Inertia Force of the Stripper (Fd). The inertia force of the these reactions depends upon the relative spring constants of
stripper is created by the acceleration of the stripper, caused the bolt and the clamped members (Kb and Kc, respectively),
by the reciprocating motion of the ram, i.e., i.e.,
Fd(t) = Msxs
F
or >=FA-uir)F' (6)

Fd\vm=M3Rc<t = 2.729 N
This magnitude is negligible compared to the contribution of F =F, -C-^-V
r (7)
other loads.
where
Distribution of Loads. Let us consider a stripper bolt that
Rs=kc/kb, (8)
clamps a sleeve and a washer which is subjected to an external
separating force (Fe). The axial force in the bolt (Fb) and the which has a value of 1.077 in the M12 stripper bolt model.
clamping force between two members (Fc) are both equal to
the initial preload (Fp) before any external forces are applied. 3 Impact Analysis
As the external loads are applied, equilibrium considerations Stresses and deflections due to impact loading are greater
require: than those associated with a corresponding static loading. The
Fe=Fb-Fc (5) severity of the impact effect depends upon the relationship

Journal of Vibration, Acoustics, Stress, and Reliability in Design APRIL 1986, Vol. 108/223

Downloaded From: http://vibrationacoustics.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jvacek/28969/ on 06/02/2017 Terms of Use: http://www.asme.org/a

 Stress State predicted by residua) stress method
Stress state predicted by nominal "teari stress method
Survival l i m i t predicted by residual stress method and Goodman Line
Survival l i m i t predicted by nominal mean stress method and Goodman L1n>
Si/rviva) limit predicted by nominal mean stress at a higher preload
Survival U n i t predicted by Gerber Parabola

NOTES;
F p • PRELOAD IK STHIPPEK BOLT
f g - GRAVITY FOftCE Of STK1PPER MASS
F s - SPRING FOKCE DUE TO I N I T I A L DEFLECTION
Gerber Parabola
F, - IHPACT FORCE DUE TO STRIPPER MOMENTUM
F d - INERTIA FORCE DUE TO RAM MOTION

. lira

Fig. 4 Input forcing function

Table 1 Comparison of impact analysis results; finite element analysis

vs. simplified analytical method
Fig. 5 Fatigue strength diagram for M12 stripper bolt under axial load
Results of
Simplified
Results of Analytical
pr£[v| Analysis Method with
C j r r e c t i j n by 90
Ci and C^
80 /^xffi / /*
Max. Value (m) 1.097 x 10" 5 1.562 x 10" 5 70
/3>< / &
•f
Impact 60 F >> ^
ny \
50
Deflection Elapsed Time After ^' ^ «*
40
A
T v<*
v*%
Impact was Applied 122 u sec. 137.6 u sec. 30
20 ,*\ V
•<*
Max. Value (N) 6411.6 8560 \'/° *>
10
Equivalent >/ %«
-60 -40 -20 2d 40 60 SO !00
Static Force Elapsed Time After Minimum s t r e s s a , , X of S
Impact was Applied 144 ]i sec. 137.6 M sec.
7
Max. Value (N/m2) 2.39 x 10 3.645 x 107 Fig. 6 Fatigue-strength diagram for alloy steel, Su = 125 to 180 ksi ax-
Impact ial loading [3]
Stress Elapsed Time After
Impact was Applied 134 g sec. 137.6 u sec.
2/
S,= (12)
kh + K
between the rapidity of the load application and the natural In the given Ml2 stripper bolt problem:
frequency of the structure. The physical properties of the
material are also functions of the speed of the loading or strain a,- = 2.940 x l 0 7 N / m 2
rate. In deriving the approximate solutions for impact stresses S, = 1.260xl0- 5 m
due to the impulsive load of the stripper momentum, the strip-
per bolt subassembly will be idealized as a simple spring-mass Effect of Strain Rate. The time required to apply the impact
system. Then, appropriate corrections will be followed as load, i.e., to increase the load from zero to its maximum value
required. may be estimated by
5,-
Basic Impact Stresses. Let us begin with the assumptions: A/ = - 1.376xl0- 4 s
vJ2
• The mass of the spring is negligible. assuming a constant acceleration. The average strain rate may
• The deflection within the mass itself (stripper) is negligible. then be estimated by
• Damping is negligible. d / du \ _ 1 S,
-= 1.017s-
On this basis, the principle of conversation of energy requires dt V dx ) ~~At L
that the kinetic energy (T) in the stripper is completely The effect of the strain rate on the steel properties is not
transformed into strain energy (U) in the spring (stripper bolt significant at this strain rate [3] and will be ignored in this
and sleeve) at the instant the maximum stress occurs, i.e., analysis.

T=U or —-Msv\=-—kb2. Effect of Spring Mass. Introducing an effective mass of the

(9)
2 2 spring (stripper bolt and sleeve) me, referred to the point of
impact, the conservation of momentum requires that
The equivalent static force (F,) applied to the spring may be
expressed as (Ms + me)vc=MsvT (13)
Fi = kb = -J2T(kb + k~) (10) where vc is the common velocity reached by Ms and me. Then
the kinetic energy in the system may be modified to
Then the impact stress and the maximum deflection may be
expressed as 1 , Mivl
T* = — (Ms + « X = — (14)
F 2(Ms + me)
'~ (11)
1+R* Compared to the previously defined T, the impact stress (a,-),
and deflection (5,) and the force (F,) should be reduced by a factor

224/Vol. 108, APRIL 1986 Transactions of the ASME

Downloaded From: http://vibrationacoustics.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jvacek/28969/ on 06/02/2017 Terms of Use: http://www.asme.org/a

 M, reaching the fixed end, the stress wave is reflected, which
C\- (15) doubles the stress intensity. The approximate solution given in
~T Ms + me equation (11) can be rewritten, with a substitution of equation
The effective mass (me) for the impact problem can be (20), as
determined similarly to that for the natural frequency of a
simple spring-mass system, i.e., for a long rod
(21)
me= m. (16)
where m is a combined mass of the bolt and the sleeve.
The analysis of this problem by Timoshenko and Goodier
In the given M12 stripper bolt problem,
[6] gives a more accurate approximation to the maximum im-
M, pact stress by an empirical formula
Cf = 0.9895,
M^ + m,
with (22)
2
m e = 5.534xl0- kg
Therefore, the elementary solution, equation (11), should be
Effects of Threads. The thread produces two distinct ef- further modified by a correction factor:
fects. First, a local reduction in cross-sectional area causes
stress increase at the root of the thread by a factor of m
C?=l+- (23)
C*2=Ab/As, (17)
which is 1.342 for the given Ml2 stripper bolt problem. The which becomes 1.253 in the M12 stripper bolt case.
second effect is the stress concentration near the thread root, Effect of Periodic Loading. Damping causes dissipation of
which causes local yielding. Due to local yielding, however, energy and reduces the amplitude of vibration. Although the
the stress is redistributed within the cross-section, producing damping in the stripper bolt may be very small, the effect of
residual stresses. This effect will be considered in section 5. damping is significant because of relatively slow ram motion
Effect of Stress Waves. At the instant the impulse is ap- compared with the natural frequency of the structure.
plied, the stress exists only at the bolt head. The inertia of the For viscous damping, the equation of free vibration is
mass particles prevents the instantaneous transmission of this mii(t) + cii(f) + ku(t) = 0,
stress. The stress travels along the bolt at velocity c, the sound
for which the general solution is
velocity in the material. As the wave advances, the mass
behind the wave front moves at the velocity of the stripper (v). u(t)=Ae-eal cos (wdt-6). (24)
This will result in an axial strain: Consider any two successive peaks with a period of r = 2Tr/cod.
vdt From equation (24), the ratio of these two successive peaks
(18) which are m cycles apart is given by
dx
From the conservation of momentum in the segment of length U„
= exp(-2/W7r£). (25)
dx during a time interval dt, the stress may be expressed as U„
a = pcv (19) In the given M12 stripper bolt problem, the natural frequency
for the first longitudinal mode is estimated to be 11680 Hz.
At the instant the mass block hits the rod with a velocity v„, The period of the excitation force on the stripper bolt is 1.5 s
the instantaneous stress at the free end of the bar can be ob- at the ram speed of 40 rpm. Even for extremely small damping
tained by combining equations (18) and (19), i.e., (£ = 0.0001), the response amplitude ratio in one load cycle
a0 = v0\fEp (20) (1.5 s) will be 1.657 X 10" 5 . This means that the response is
The wave front of stress a„ proceeds at velocity c, and the completely dissipated within one load cycle and that the
magnitude behind the wave front diminishes as the velocity of residual response will not be coupled with the response to the
the moving mass or the end contact pressure decreases. Upon next load cycle. Therefore the response analysis for the period
of a single load cycle is adequate, even if the impulsive load is
periodic.
Table 2 Fatigue stress-concentration factors Kt, typical values for Maximum Impact Stress. The local maximum impact stress
unified and American standard threads, steel bolts-bending or axial in the M12 stripper bolt, modified by correction factors (Cf,
loading [5]
Cf, and Cf), can be obtained as
Rolled Cut af= Cf Cf Cf o-,- = 4.892 X 10 7 N/m 2 (26)
Steel Threads Threads
The validity of this simplified impact analysis was con-
Annealed firmed by a finite element analysis using an axisymmetric
(less than 200 Bhn) 2.2 2.8 model shown in Fig. 2 [1]. The finite element analysis shows
Quenched and drawn
(over 200 Bhn) 3.0 3.8
Table 3 Properties of the cushion material (polyurethane with Shore
hardness 90)
Note 1. Nominal s t r e s s e s s h o u l d be based on
the stress area of the thread Injection molded, Elastomer type
( c o r r e s p o n d s t o a d i a m e t e r between Tensile strength: a = 300 kg/cm2
t h e p i t c h and r o o t d i a m e t e r s ) .
Note 2. The t a b u l a t e d v a l u e s were o b t a i n e d Stress at 20% elongation: a ? = 50 kg/cm
with 3/8-in. bolts, and may Stress at 300% elongation: a, 0 = 150 kg/cm
c o n s e r v a t i v e l y be a p p l i e d t o s m a l l e r
sizes. L a r g e r values of Kf s h o u l d Energy loss factor: n = 0.1
be used w i t h l a r g e r b o l t s .

Journal of Vibration, Acoustics, Stress, and Reliability in Design APRIL 1986, Vol. 108/225

Downloaded From: http://vibrationacoustics.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jvacek/28969/ on 06/02/2017 Terms of Use: http://www.asme.org/a

 Table 4 Stamping die design with M16 stripper bolt fracture involves metallographic behavior of the material,
such as crack nucleation, development of microcracks, etc.
Size of s t r i p p e r bolt: M16 For engineering practice, the prediction of quantitative fatigue
Number of s t r i p p e r b o l t s t o be used: 7 characteristics relies on the macroscopic test data. The S—N
Effective l e n g t h of s t r i p p e r b o l t , m: 0.055
curve (alternating stress versus number of cycles to failure) is
constructed on a log-log scale based on rotating bending tests,
Mass of s t r i p p e r p l a t e , kg: 102 which involve a completely reversed bending load, 0.3-in.
Preload i n s t r i p p e r b o l t , N: 73000 diameter solid round specimen without stress raisers and
Crank speed, rpm: 35 , mirror-polished surface. The endurance limit corresponding
to this special case is designated by S{, rather than by the
Crank r a d i u s of p r e s s , m: 0.225
general designation S„.
Maximum s t r o k e of s t r i p p e r , m: 0.015 The endurance limit obtained by rotating bending tests with
Stiffness of s t r i p p e r s p r i n g , N/m: 26.44 x 10 3 standard specimens (SI,) has to be modified to obtain the
Initial d e f l e c t i o n of s t r i p p e r s p r i n g , m: 0.007
general definition of the endurance limit S„. This modification
can be accomplished by multiplying S^ by the correction fac-
Number of s t r i p p e r springs: 16 tors [3] as in
S„=CLCDCSS;, (30)
25-35 percent lower values of maximum displacement, force where CL, CD, and Cs are correction factors for various load
and stress, as summarized in Table 1. However, the conser- types, size effect and the surface roughness effect,
vative solution obtained by the present analysis is appropriate respectively.
for the fatigue analysis, especially due to probable adverse ef-
fects of coupled vibration response of the stripper plate Effect of the Mean Stresses. For more general conditions of
modes. fatigue loading, consider the effect of superimposing a mean
stress (cr„,) on the alternatng stress {aa). Two restrictions will
be retained: the loading is simple, i.e., bending only or axial
4 Stresses Due to Rectangular Pulse (Fr)
load only, and the material is homogeneous and free of
Since the rectangular pulse is periodic, the response should residual stresses prior to the application of loading. Numerous
be obtained from the solution to the harmonic loading with a curves have been proposed to represent a constant fatigue life
Fourier series expression of the rectangular pulse. However, under the combined stresses. The two most common curves
because of the fast attenuation of the response as indicated are the Goodman line and the Gerber parabola as shown in
earlier, only one load cycle needs to be analyzed. Fig. 5.
Consider an equation of motion
Effect of Stress Concentration. Stress concentration
mii(t) + ku(t)=Fr(t), assumes major importance in the fatigue problem, because
with at-rest initial conditions. The general solution is fatigue fractures invariably originate at points of stress con-
F centration such as the bolt head fillet and the threads in the
U(t) = —L(l-cos at) for 0<t<r (27) stripper bolt. The maximum stresses at such points are
customarily determined by first calculating the normal stress
and and then multiplying this value by a stress concentration
factor.
U(f) Laboratory tests show that reductions in fatigue strength
U{t) = sin citl + U(j) cos wt\ for r > r (28)
caused by stress raisers are less than the predicted reduction
where tx = t — r and T is the duration of the loading phase. based on the geometric stress concentration factor (K,). To
Equations (27) and (28) represent responses to the first phase deal with this situation, the fatigue stress concentration factor
step loading and the second phase free vibration, respectively. or fatigue-strength reduction factor (Kf) is defined as
The maximum response occurs in the first phase with a Sn for unnotched sample
dynamic magnification factor of 2. f
S„ for notched sample
In the given M12 stripper bolt problem, Fr is found to be
359.85 N. The peak response will be reached at a very early Efforts to estimate Kf without a costly test program were
stage of the step loading phase. Since the step loading is also made by many researchers such as Neuber, Kuhn, Peterson,
an impulsive loading, more accurate results should be ob- etc. A refinement of the Neuber's approach, developed by
tained by applying the same correction factors as those Peterson, is most convenient to use and has been widely ac-
described in the preceding section. Then the maximum axial cepted. Peterson has developed a family of curves, giving a
stress in the stripper bolt due to rectangular pulse can be ob- factor called notch sensitivity (q) for various materials. The
tained from notch sensitivity is a function of the notch radius and defined,
in terms of K, and Kf, as
fff=2CfC2*C3* - = 5.098 x l 0 6 N / m 2 (29)
d+Rs)Ab */-* (32)
q-
K,-\
5 Fatigue Considerations An example can be illustrated with the Ml2 stripper bolt us-
The stamping die set is expected to survive at least one-half ing the results of the finite element analysis [1]. Subjected to
million stamping operations. However, by virtue of high the preload of 19748 N, the model in Fig. 2 produced
natural frequency and the low damping, the stripper bolt is
subjected to thousands of stress cycles in one stamping opera- • axial stress at the bolt shank;
tion. Thus the design dictates that the stripper bolt have an in- formal = 1-746 X 1 0 8 N / M 2
finite fatigue life. It is fortunate that the steel has an encurance 9
limit (S„), which allows the material to withstand an infinite maximum principal stress at the bolt head fillet;
number of loading cycles. amax = 5.892 x10 s N/m 2
Generalized S—N Curve. The mechanism of the fatigue For the fillet radius of 1 mm and the ultimate strength of 7.84

226/Vol. 108, APRIL 1986 Transactions of the ASME

Downloaded From: http://vibrationacoustics.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jvacek/28969/ on 06/02/2017 Terms of Use: http://www.asme.org/a

 x 108 N/m 2 , the notch sensitivity (q) of the steel is estimated nating stress for survival are 0.1636 ou and 0.1944 au based on
to be 0.8 [3]. The stress concentration factors are the residual stress method and the nominal mean stress
method, respectively. These points are indicated by 3 and 4
*,= °mm =3.38 respectively, in Fig. 5. Suppose the stripper bolt is preloaded
^nominal to 75 percent rather than 71 percent of the yield point. Then
and the limit stress level for survival will be reduced according to
Kf=\+(K,-\)q^2.9. the nominal mean stress method (indicated by 5 in Fig. 5),
while the limit does not change according to the residual stress
Some fatigue stress concentration factors for the standard method. Notice that, according to the residual stress method,
steel bolts obtained experimentally are provided in Table 2 [5]. the yield to ultimate stress ratio (therefore ductility) plays a
Based on Table 2, the fatigue stress concentration factor of 3 very important role in predicting a fatigue life. For the current
will be assumed for the calculation of fatigue strength of the stripper bolt problem, the residual stress method is adopted.
M12 stripper bolt. Notice that these tabulated values already This analysis procedure was applied to several different
took into account the effect of the surface condition. cases. However, the Goodman line as a criterion for the in-
Effects of Residual Stress. The fatigue problem involving finite fatigue life of the stripoper bolt was found to be too con-
stress concentration together with combined mean and alter- servative according to the factory experience. The Gerber
nating stress become complicated because of the residual parabola, is employed based on the data presented in Fig. 6.
stresses resulting from local yielding. Theoretical studies tend The Gerber parabola can be expressed by
to be extremely complex and dependent upon simplifying y = ax2 + bx+c (33)
assumptions, which may deviate objectionably from the true
situation. To overcome the difficulties, relatively simple pro-
cedures are employed in the engineering practice. Two of the
most widely used methods are the residual stress method and
the nominal mean stress method. Three constraints are required to determine the values of the
coefficients. Two constraints are immediately apparent from
The residual stress method assumes that all the stresses
the extreme cases
(mean as well as alternating) are increased by the fatigue stress
concentration factor (Kf) and that the residual stresses caused 7(0) = —— and 7(1) = 0.
by the initial application of peak loads can be determined
from conventional stress-strain curves. Since the development The third constraint is established by examining the slopes of
of residual stresses upon the initial application of a peak load the curves at x = 0 in Fig. 6 [3], The slopes of the curves
is caused by the local yielding, it is implied by this method that representing the life of 106 and 107 cycles can be closely
such localized yielding reduces the effective mean stress. With matched by a constraint
the nominal mean stress method, a fatigue stress concentra-
tion factor is applied only to the alternating stress. y'(0)=-c + 0A5.
In the example case, the M12 stripper bolt will survive ac-
6 Fatigue Analysis of Stripper Bolt cording to the Gerber criterion, and the survival limit is found
to be 0.23 as indicated by point 6 in Fig. 5.
Based on the results of stress analysis and the collection of
fatigue data, we can predict the approximate fatigue life of the
stripper bolt. Since we are only interested in the stripper bolt 7 Discussion
with an infinite fatigue life, the analysis will be confined to The fatigue strength of the material is inherently statistical.
predicting whether the bolt would survive or fail, and the Strength data obtained from groups of presumably identical
limiting stress level for survival. fatigue tests have much more scatter than similar static test
First, the endurance limit for the Ml2 stripper bolt can be data. Fatigue fractures start at highly localized points of local
obtained by weakness. The strength of these points is sensitive to the sur-
S„ = CLCDCsSl = 0A5ou face finish and even to the crystalline structure of the material.
with The variations of these factors are difficult to control, and the
statistical considerations become inevitable. Therefore the
S^=0.5a„ (for steel) results form the fatigue analysis are approximate and by no
CL =0.9 (for axial loading) means precise. For more accurate fatigue analysis, the fatigue
CD = 1.0 (0.9 for larger screws) strength has to be accompanied by the probability of failure
Cs = 1.0 (Kf will account for this effect based on Table 2) and the confidence level, which could be estimated from a
and the yield stress (ay) is 80 percent of the ultimate stress (au). large volume of test data.
The fatigue strength diagram can be constructed using the In this analysis nominal stresses are calculated at the root of
Goodman line as shown in Fig. 5. the thread. The stress concentration at the bolt head fillet is
Second, the stresses have to be calculated. The mean and the assumed to be equal to that at the thread root. This is
alternating stresses can be conservatively established by validated by a finite element analysis. It is also noted that fac-
tory experience indicates the fatigue failure occurs in either
am =<jp = 0.568 au one of these places with approximately equal probability. The
<ra = oi* + ar*=0.069ou. solution presented in this paper is somewhat conservative.
This is well justified because of the irregular nature of the
Including the effect of the fatigue stress concentration factor fatigue behavior.
(Kf = 3.0), the peak streses are There are a number of ways to improve the fatigue strength
o,„* =Kfam = U04au fo the stripper bolt without having to use the larger size screw.
One method is to increase the stiffness ratio (Rs) defined in
oa*=Kfoa = 0.207ou. equation (8), because a larger portion of the externally applied
The data points can be plotted on the fatigue strength dynamic load will be reacted by decreasing clamping force for
diagram as indicated by 1 and 2 in Fig. 5. Both the residual a larger value of Rs and the bolt would sustain less dynamic
stress method and the mean stress method predict failure load. The value of Rs may be increased by stiffening the sleeve
based on the Goodman line. The limiting values of the alter- or simply by necking down the bolt shank.

Journal of Vibration, Acoustics, Stress, and Reliability in Design APRIL 1986, Vol. 108/227

Downloaded From: http://vibrationacoustics.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jvacek/28969/ on 06/02/2017 Terms of Use: http://www.asme.org/a

 Another method for enhancing the fatigue life is to in- In order to demonstrate the effect of the cushion, a stamp-
troduce a damping material to absorb the energy induced by ing die with Ml6 stripper bolt is analyzed with the data pro-
the impact. This method is proven to be very effective as vided in Table 4. It is noted that the constant for the size effect
presented in the following section. is CD = 0.9 in the case of the Ml6 stripper bolt. The max-
imum alternating stress in the stripper bolt is 59.4 percent of
8 Application of the Cushion the ultimate stress, while the survival limit stress is only 19.7
percent of the ultimate stress. The stripper bolt fails
In order to enhance the fatigue life of the stripper bolt, the miserably.
design of the stripper bolt subassembly was modified to in- When the cushion is introduced as shown in the Appendix,
troduce a cushion made of polyurethane. This cushion in a the maximum alternating stress in the stripper bolt is only 4.9
ring shape is inserted on the washer as shown in Fig. 1. percent of the ultimate stress, and the stripper bolt will survive
In the previous impact analysis of the stripper bolt, the by a wide margin.
kinetic energy of the stripper was converted to the strain
energy of the stripper bolt and the sleeve, treating the stripper
plate as a rigid mass. For the impact analysis with the cushion,
the stripper bolt subassembly is idealized as a one-mass-two-
spring system, consisting of a soft spring (kd) for the cushion, 9 Concluding Remarks
connected in series with a stiff spring (Ks) for the stripper bolt The stress solution for the fatigue analysis was obtained by
and sleeve. In the presence of damping, the conservation of approximate methods followed by appropriate corrections.
energy dictates: This approximate solution was found to be conservative com-
pared to the finite element analysis results. But the conser-
T=U, + Ud+R (34) vative solution is rather pertinent for the fatigue analysis
where the kinetic energy in the stripper is because of its dispersed nature. For the fatigue life prediction,
the Gerber parabola was determined to be superior to the
Goodman line as a realistic criterion. A cushion was intro-
duced in an effort to increase the fatigue life of the stripper
the strain energy in the stripper bolt and the sleeve is bolt, which was proven to be very effective.
An interactive program was developed for the stripper bolt
i-A,V=-L(*6+*c)v, analysis. Although it is not a computation-intensive analysis, a
computer program is reliable and well suited for the repetitive
the strain energy stored in the cushion is application of design and analysis procedure, as in the stamp-
ing die design. Especially with the low-cost microcomputer
Ud~kdhd\ based engineering work-station, this simple interactive pro-
gram will be an essential engineering tool for die designers.
and the energy dissipated by the damping in the cushion is
1
R=—-igkd5d2. References
T
Using the properties in Table 3, the elasticity modulus of the 1 Lee, S. H., and Burfeindt, H., "Finite Element Analysis of a Preloaded
cushion may be estimated as Stripper Bolt," Finite Element News, June, 1981.
2 Illgner, K. H., and Blume, D., "Schrauben Vademecum," Bauer &
Schaurte Karcher GmbH, 1980.
Ed=- = 2.455 x l 0 7 N / m 2 .
0.2 3 Juvinall, R., Stress, Strain, and Strength, McGraw-Hill, New York, 1967.
4 Lee, S. H., "Fatigue Analysis of Preloaded Stripper Bolt under Combined
Taking into account the energy loss factor, the complex Loading of Impact and Rectangular Pulse for Improvement of Stamping Die
modulus of elasticity can be expressed as Design," Volkswagen-werk Statusbericht, Aug. 1980.
5 Lipson, C , Noll, G. C , and Clock, L. S., Stress and Strength of
E*d=Ed(l + ir,). Manufactured Parts, McGraw-Hill, New York, 1950.
6 Timoshenko, S. P., and Goodier, J. N., Theory of Elasticity, McGraw-
However, the energy dissipated by damping in the cushion (R) Hill, New York, 1970.
is an order of magnitude lower than the strain energy (Ud)
stored in the cushion, and the dissipated energy term (R) may
be neglected for simplicity.
Introducing a constraint
, k„ A P P E N D I X
cVa-Kfii or 5,= \-^-)?>d
M,
to equation (34) results in Computer output for fatigue analysis of M16 stripper bolt
with cushion
K,
S„=- FATIGUE ANALYSIS OF PRELOADED STRIPPER BOLT
oid " kd+Ks UNDER COMBINED LOADING OF IMPACT AND RECTANGULAR PULSE
FOR IMPROVEMENT OF STAMPING DIE DESIGN
Then the equivalent static force due to impact is obtained by
YIELD STRESS OF STRIPPER BOLT IN • 0.6a7BE+09
ULTIMATE STRESS OF STRIPPER BOLT IN = 0.7840E+09
STRIPPER BOLT DIAMETER SIZE - 0.01608
FATIGUE CRITERION FOR COMBINED STRESSESiGERBER PARABOLA
The deflection of the cushion (5rf) is, however, limited to 8d as FATIGUE STRESS CONCENTRATION FACTOR = 3.0
shown in Fig. 1. When 5d reaches 5d, the equivalent static IMPACT IS ABSORBED BY CUSHION UITH E= 0.E455E+08

force in the stripper bolt may be obtained by 1. DISTRIBUTION OF DYNAMIC LOADS

STIFFNESS AREA OF THE STRIPPER BOLT = 0.2011E-03
AXIAL STIFFNESS OF THE STRIPPER BOLT = 0.7677E+89
F,- = kd5d+Fc (36) STRESS AREA OF THE CLAMPED MEMBER • 0.E073E-03
AXIAL STIFFNESS OF THE SLEEUE - 8.7917E+09
where STIFFNESS RATIO - 0.1031E+81
FRACTION OF EXTERNAL FORCE APPLIED TO BOLT 0.4933E+00
F/ = 2KS(T--Lkd5*/). 2. PRELOADED STRESS
MEAN STRESS DUE TO PRELOAD * 0.46S0E+09

228/Vol.108,APRIL1986 Transactions of the ASME

Downloaded From: http://vibrationacoustics.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jvacek/28969/ on 06/02/2017 Terms of Use: http://www.asme.org/a

 3. IMPACT LOAD Alternating Stress/
ANGULAR VELOCITY OF THE RAM IN RAD/SEC • 0.366SE+01
TIME UPON IMPACT IN SECONDS • 0.1386E+01 Ultimate Stress
VELOCITY OF THE STRIPPER UPON IMPACT • 0.a961E+00
TOTAL IMPULSE MOMENTUM IN THE STRIPPER • 0.3030E+0a 1.0
Gerber Parabola
4. APPROXIMATE ANALYSIS OF IMPACT STRESS
TOTAL KINETIC ENERGY IN THE STRIPPER • 0.4470E+01
STIFFNESS OF THE CUSHION UPON IMPACT - 0.S36SE+07 Goodman Line
DEFLECTION OF CUSHION UPON IMPACT • 0.4871E-03
EQUIVALENT STATIC FORCE OF THE IMPACT - 0.1839E+0S
Yield Criterion
NOMINAL IMPACT STRESS BEFORE CORRECTION « 0.6398E+07
MAXIMUM DEFLECTION DUE TO IMPACT • 0.a862E-04 Data Point with Cushion
5. MORE ACCURATE SOLUTION FOR IMPACT STRESS
MASS IN THE STRIPPER BOLT AND THE SLEEVE • 0.1765E+00 Data Point without Cushion
EFFECTIVE MASS IN STRIPPER BOLT AND SLEEVE • 0.5883E-01
CORRECTION FACTOR FOR ALLOWANCE OF SPRING MASS • 0.9988E+00
CORRECTION FACTOR FOR EFFECTIVE AREA OF THREAD = 0.1381E+01
CORRECTION FACTOR FOR EFFECT OF STRESS UAVES • 0.1083E+01
MAXIMUM LOCAL IMPACT STRESS AFTER CORRECTION- 0.8865E+07

6. RECTANGULAR PULSE
FORCE FROM INIT. COMPRESSION OF STRIPPER SPRING- 0.a961E+04
STRIPPER FORCE DUE TO GRAVITY - 0.1003E+04
MAGNITUDE OF THE RECTANGULAR PULSE - 0.3963E+04
MAXIMUM STRESS DUE TO RECTANGULAR PULSE • 0.3841E+07
7. FATIGUE ANALYSIS
ENDURANCE LIMIT OF THE STRIPPER BOLT = 0.3175E+89
MAXIMUM ALTERNATING STRESS » 0,ia71E+08
MEAN STRESS AS Si OF YIELD STRESS - 0.7413E+03
MODIFIED ALTERNATING STRESS U1TH KF - 0.381EE+08
0.486aE+01
MODIFIED ALTERNATING STRESS AS X OF ULTIMATE STS- 0.8058E+09
RESIDUAL STRESS AT THE POINT OF STRESS CONCENT. - 0.7S14E+03
EFFECTIVE MEAN STRESS AS % OF ULTIMATE STRESS - 0.S134E+08
LIMITING VALUE FOR THE SAFE ALTERNATING STRESS • 0.19S5E+03
LIMIT ALTERNATING STRESS AS % OF ULTIMATE STRESS-
CONCLUSION : STRIPPER BOLT UILL SURVIVE.

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Mean Stress/Ultimate Stress

Fatigue Strength Diagram for M16 Stripper Bolt

Journal of Vibration, Acoustics, Stress, and Reliability in Design APRIL 1986, Vol. 108/229

Downloaded From: http://vibrationacoustics.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jvacek/28969/ on 06/02/2017 Terms of Use: http://www.asme.org/a