Engineering Failure Analysis 44 (2014) 285298

Contents lists available at ScienceDirect

Engineering Failure Analysis

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

Analysis of fatigue failure on the keyway of the reduction gear

input shaft connecting a diesel engine caused by torsional
vibration
Hyung-Suk Han
Naval System Research Team, Busan Center, Defense Agency for Technology and Quality, 525-2, Gwangan 1 dong, Busan, Republic of Korea

a r t i c l e

i n f o

Article history:
Received 26 November 2013
Received in revised form 14 May 2014
Accepted 20 May 2014
Available online 2 June 2014
Keywords:
Torsional vibration
Lateral vibration
Fatigue
Propulsion shaft

a b s t r a c t
The vibratory torque of a diesel engine caused by the reciprocating motion of the mass and
gas pressure force of a cylinder is one of the main causes of the failure of the driving shaft
of the diesel engine and the connecting shaft to the reduction gear. Because high cycle torsional fatigue can occur in the reduction gear shaft connecting the engine under vibratory
torsional stress, the US Navy restricts it under the MIL G 17859D(SH) standard and suggests a procedure for evaluating the safety of the shaft for the reduction gear. In this study,
the structural safety of the reduction gear input shaft in which fatigue failure occurs in typical naval vessels is investigated in accordance with the VDI 3822 RCA (root cause analysis)
procedure based on the MIL G 17859D(SH) standard. When evaluating the safety factor in
accordance with the MIL G 17859D(SH) standard, the alternating bending moment from
the lateral vibration and the stress concentration factor under static load are considered.
In addition, an improved design is suggested by CAE to satisfy the safety factor suggested
by the MIL G 17859D(SH) standard.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
Recently, shipbuilders have been manufacturing many types of ships. These ships usually adopt a diesel engine for their
propulsion system. Since a diesel engine is operated by the force of the cylinder from the explosion of the gas, the torsional
vibration from the uctuation torque is bigger than that of other types of engines, such as gas-turbine and electrical propulsion motors. Therefore, the propulsion shafts in ships frequently fail due to the extreme torsional vibration from diesel
engines. The engines of naval vessels that require more power and revolution speed usually adopt V type, 4 stroke diesel
engines and reduction gears in order to increase the output torque. Therefore, the healthy design of shafts is strongly
required for naval vessels.
Engine shaft failures usually occur at stress concentration areas such as the llet and chamfer under a dynamic load.
Peterson [1] suggested the stress concentration factor for various mechanical designs such as a hole, ange, and keyway,
and their study serves as the most commonly used reference for stress concentration factors. The stress concentration factor
at the end of the keyway dominantly affects the life of the shaft when torque is transmitted through the key. Okubo [2] suggested the stress concentration factor in keyways of the shaft from experiments with ne-grained carbon steel. Pedersen [3]
Address: Defense Agency of Technology and Quality, 525-2, Gwangan 1-dong, Suyeong-gu, Busan 613-808, Republic of Korea. Tel.: +82 51 750 2533;
fax: +82 51 758 3992.
E-mail address: daerihan@hanmail.net
http://dx.doi.org/10.1016/j.engfailanal.2014.05.012
1350-6307/ 2014 Elsevier Ltd. All rights reserved.

286

H.-S. Han / Engineering Failure Analysis 44 (2014) 285298

suggested a new design for a keyway modeled as a hyperbola, which could reduce the stress concentration factor compared
to the standard design as per DIN 6885 [4].
Fracture analysis reports of a shaft with a keyway have been published in many journals. Goksenli and Eryurek [5]
studied the fatigue fracture of the drive shaft of an elevator with a keyway. He estimated the safety factor of the driving
shaft using the Goodman method and concluded that fracture was caused by torsionalbending fatigue through the evaluation of the microstructural, mechanical, and chemical properties of the driving shaft. Bhaumik et al. [6] reported cracks
in the hollow shaft of a single-stage helical gear box. Subsequent investigations revealed that the crack initiated by fatigue
at one of the keyway edges and propagated in a helical manner. He suggested possible root causes as fault machining,
inadequate radius at the keyway run out radius, and elongated MnS inclusion. Parida et al. [7] reported the root cause
of the torsional fatigue failure of a coal pulverizer mill shaft. Through mechanical and chemical analysis, it was found that
the initiation and growth of cracks were facilitated by low-toughness brittle microstructures resulting from improper heat
treatment.
In this research, the root cause analysis (RCA) of the failure of the reduction gear input shaft connecting to a diesel engine
is described in accordance with the VDI 3822 [8] RCA procedure, as shown in Fig. 1, considering the torsional vibration of a
diesel engine as well as the bending moment from the input shaft, which comes from the weight of the shaft and the lateral
vibration of the connecting engine.

2. Failure description and record of the history

Pictures of the failure of the shaft in this research are shown in Fig. 2. As illustrated, the crack of the shaft was developed
from the keyway in a 45-degree direction with a cup and corn shape as shown in Fig. 2(a)(c). In Fig. 2(c), it can be found that
the crack initiated at the end of the keyway and propagated to the rotational direction with 45 degree helical manner. It is a
representative pattern of torsional fatigue. SEM pictures on the fracture surface at the crack initiating and propagation zone
are also shown in Fig. 2(d). Upon SEM analysis of the failure surface, the initial crack appears to have occurred at the end of
the keyway and developed toward the center of the shaft with the sequence of beach marks having spiral shape striation.
Therefore, it can be concluded that this failure occurred by torsional fatigue based on the records referred to above, and
the RCA is performed to address the hypotheses in the next chapter.

Description of the failure

Document of the failure mode
Survey of special characteristics of the design, material or
processing

Record of failure history

Failure case history
View of entire system
Survey of the nominal state
Operation and environmental conditions
Specific chronological and/or local conditions
Frequency periodicity
Change in the material

Failure hypothesis (hypotheses)

Guide of the analysis
Assessment of verification cost (time and money)

Instrumental analysis
Investigation plan
Sample collection
Test and analysis
Simulation and reconstruction trials

Investigation results
Nominal/actual comparison
Iteration loop (if needed)

Cause(s) of failure
Primary and secondary failure influences (when given)

Failure collection
Actual and preventive measures

Report
Knowledge
management

Fig. 1. Performance of a failure analysis.

287

H.-S. Han / Engineering Failure Analysis 44 (2014) 285298

45
Initiation of crack

(b) view at the keyway side

(a) crack propagation

Crack
initiation zone

45 line

Final
rupture

(c) fracture surface at the flange side

SEM on the initial crack position

SEM on the crack propagation zone

Beach Mark

Crack initiation
point

Crack initiation

Fatigue striation

(d) SEM of the initial crack and crack propagationzone

Fig. 2. Failure pattern of the reduction gear input shaft connecting to the diesel engine.

3. Failure hypotheses
As referred in the previous chapter, failure hypotheses for the shaft failure are as follows:
(a) The failure occurred from torsional fatigue due to the bad material and production error.
(b) The initial crack occurred when the shaft was manufactured before installing it in the ship and developed during the
ships operation period.

288

H.-S. Han / Engineering Failure Analysis 44 (2014) 285298

(c) The shaft does not have sufcient safety factors to endure the force induced by the torsional and lateral vibration of
the engine.
Based on these hypotheses, an RCA logical tree can be drawn, as shown in Fig. 3 and RCA can be performed. The instrumental analysis for each item in Fig. 3 is described in the next chapter.
4. Instrumental analysis and investigation
When the thermal stress and corrosion in Fig. 3 was investigated, there was no special problem. Therefore, the instrumental analysis about the mechanical stress including the mechanical and material properties is performed and described in this
chapter.
4.1. Material and mechanical properties
In this research, the material of the reduction gear input shaft connecting to the diesel engine is 826M40 (Nickel
Chromium Molybdenum Steel) hardened and tempered according to BS970.
Tables 1 and 2 depict the analysis results of the material and mechanical properties. As shown in these tables, the material and mechanical properties of the reduction gear input shaft are satised with required specication.
Therefore, hypothesis 1 is not the main root cause for the shaft failure based on the instrumental analysis as above.

Fig. 3. RCA logical tree.

Table 1
Material analysis result (unit: wt%).
Contents

Specication

Inspected value

C
Si
Mn
Ni
Cr
S
P

0.360.44
0.100.35
0.450.70
2.302.80
0.500.80
Max. 0.025
Max. 0.025

0.39
0.26
0.59
2.38
0.58
0.002
0.015

289

H.-S. Han / Engineering Failure Analysis 44 (2014) 285298

Table 2
Mechanical analysis result.
Item

Specication

Inspected value

Yield strength
Ultimate strength
Elongation
Hardness

740 MPa min

9251075 MPa
15% min
269331 HB

847 MPa
981 MPa
18%
282296 HB

Fc(coupling weight), Fv(force from lateral vibration)

Keyway
p(weight per unit length)

Fig. 4. Assumption of the simply supported beam of the reduction gear input shaft.

Fig. 5. Test setup of the lateral vibration of the diesel engine.

When microscopic and macroscopic analyses were performed for the failure surface, the initial crack could not be found.
In addition, when investigating the inspection data about the shaft for the delivery conditions, such as liquid penetrant testing (PT) and magnetic particle testing (MT), there was no record for the initial crack.
Therefore, hypothesis 2 also is not the main root cause for the shaft failure based on the recording of the shaft at the delivery condition.

290

H.-S. Han / Engineering Failure Analysis 44 (2014) 285298

(a) Vibration displacement

0.0020

acc1-x
acc1-y
acc1-z
acc2-x
acc2-y
acc2-z
acc3-x
acc3-y
acc3-z

0.0018

Displacement [m]

0.0016
0.0014
0.0012
0.0010
0.0008
0.0006
0.0004
0.0002
0.0000

10

100

Frequency [Hz]

(b) Vibration displacement (Ship 1)

0.0020

acc1-x
acc1-y
acc1-z
acc2-x
acc2-y
acc2-z
acc3-x
acc3-y
acc3-z

0.0018

Displacement [m]

0.0016
0.0014
0.0012
0.0010
0.0008
0.0006
0.0004
0.0002
0.0000

10

100

Frequency [Hz]

(c) Vibration displacement (Ship 2)

Fig. 6. Lateral vibration level of the diesel engine.

4.2. Mechanical stress analysis

4.2.1. Applied load
The denition of the applied load is necessary to analyze the mechanical stress of the shaft. In this research, the applied
load is dened considering the mean and alternating diesel engine torque as well as the bending moment that comes from
the lateral vibration and weight of the shaft.
Assuming that the reduction gear input shaft is simply a supported beam, as shown in Fig. 4, the bending moment from
the weight of the shaft, including the coupling weight, can be determined.

H.-S. Han / Engineering Failure Analysis 44 (2014) 285298

291

The bending moment of the shaft from the engine vibration can be estimated with the measured vibration of the engine
according to ISO 10816-6 [10], as shown in Fig. 5. In Fig. 5, the engine vibration was measured with three axis accelerometers
(PCB Type 356A02) and a data acquisition system (B&K Type 3560B). Fig. 6 is the measured vibration for two ships that used
the same diesel engine and reduction gear described in this research. In Fig. 6, the vibration levels of these two ships are
quite different. When analyzing the spectra of vibrations for these ships, it can be found that the difference comes from
the vibration of the engine and mounting system whose frequency range is 210 Hz. This vibration comes from the movement of the engine supported by an elastic mount and should be dependent on the sea and ship operating conditions. In this
research, the vibration level of ship 1 is adopted for the conservative analysis.
The bending moment from the lateral vibration of the engine can be calculated based on the stiffness of the exible coupling in the lateral direction and its vibration displacement, as shown in Fig. 7. Here, the vibration displacement can be calculated based on the double integral with the measured acceleration. In Fig. 7, the bending moment considering only the
weight of the shaft and coupling is shown, and the bending moment considering the lateral vibration of the engine is much
bigger than that considering the weight of the shaft and coupling.
The torque of the diesel engine can be classied as the mean torque and alternating torque. The mean torque can be calculated with Eq. (1), and Fig. 8(a) is the calculated mean torque from Eq. (1) for the revolution speed range of the diesel
engine in this research.

QT




60
W
x 2
 K  1000 N m
2p xmax xmax

Here, QT is the mean torque, W is the output power (kW), xmax is the maximum revolution speed (rpm), x is the revolution
speed, and K is the safety factor (=1.1).
The alternating torque was calculated with measured alternating angular velocity with a laser velocity meter and the torque per unit alternating angular velocity from TV calculation of the diesel engine suggested by its supplier as shown in
Fig. 8(b).
Fig. 8(c) shows the measured alternating torque at each rpm step. In Fig. 8(c), the revolution speed is 1148 rpm, at which
the alternating torque is at the maximum level (8.81 kN m).
Through Figs. 7 and 8, the load applied on the reduction gear shaft can be dened, and the stress analysis is performed
employing the theory and CAE described in the next chapter.
4.2.2. Stress calculation procedure
In this research, the stress analysis is performed in accordance with MIL G 17859D, Appendix D [9]. The resultant steady
stress (Sr) can be written as given in Eqs. (2) and (3) based on the maximum shear stress theory.

Sr

q
S2c 2Ss 2

2
4

Sc 0; Ss

QTd
pd
; J
2J
32

Here, Sc is the steady compressive stress, Ss is the steady shear stress, and d is the shaft diameter
Eq. (4) [1] is the effective stress concentration factor.

K f 1 qK t  1

Fig. 7. Bending moment of the diesel engine.

292

H.-S. Han / Engineering Failure Analysis 44 (2014) 285298

(a) Mean torque

(b) Test setup for alternating torque measurement

(c) Alternating torque

Fig. 8. Torque of the diesel engine as revolution speed.

Here, Kf is the effective stress concentration factor, Kt is the ideal stress concentration factor, and q is the notch sensitivity
index.In Eq. (4), the notch factor under the dynamic load is almost 1.0. However, the notch factor of the ductile metal under
the static load is 0 due to the hardening from the local plastic deformation. Therefore, it is 0 in MIL G 17859D when calculating the resultant steady stress. However, the notch factor of a ductile metal may be able to increase under the condition
that inhibits plastic slip. For the impact load, the notch factor is recommended to have a value between 0.4 and 0.6 for ductile
metals and 1.0 for brittle materials. The notch factor of the brittle material for which the elongation ratio is under 5% under a
static load is usually 0.150.25.
The stress concentration factor of the keyway, as shown in Fig. 9, is given in Eqs. (5) and (6) for the bending moment and
torsion, respectively.

 
 2
0:1
0:1
 0:0019
K t;B 1:426 0:1643
;
r=d
r=d

0:005 6 r=d 6 0:04; d 6 165:1 mm

H.-S. Han / Engineering Failure Analysis 44 (2014) 285298

293

Fig. 9. Shaft with keyway.

 
 2
0:1
0:1
 0:0021
1:953 0:1434
;
r=d
r=d

K t;T

0:005 6 r=d 6 0:07

Here, Kt,B is the ideal stress concentration factor for the bending moment, Kt,T is the ideal stress concentration factor for the
torsion, r is the llet radius, and d is the shaft diameter.The notch factor under the alternating load can be represented as
given in Eqs. (7) and (8) [1].

1
;
1 a=r

1
;
1 0:6a=r

for bending & axial loading

for torsional loading

Here, a is the material constant; quenched and tempered steel is 0.0025, annealed and normalized steel is 0.01, and aluminum alloys is 0.025.
Considering above stress concentration factor, the resultant alternating stress can be written as given in Eqs. (9)(11).

Sar

q
K f ;B S2b 2K f ;T Sas 2

Sb

Mb
pd
;Z
Z
32

10

Sas

Ta
Sr
T max

11

Here, Sar is the resultant alternating stress, Sb is the alternating bending stress, Sas is the alternating torsional shear stress, Kf,B
is the effective stress concentration factor for the bending moment. Kf,T is the effective stress concentration factor for the
torsion, Mb is the bending moment from the weight of the shaft and the lateral vibration of the engine, and Ta, Tmax are
the alternating and maximum torque of the engine, respectively.
4.2.3. Stress calculation by MIL G 17859D
The yield strength of the shaft in this research is 847 MPa as shown in Table 2 in the previous chapter, and the fatigue
limit can be calculated to 304 MPa, which is 0.5 times the ultimate strength of applying an additional modication factor
(k = 0.62) as given in Eqs. (12)(16) [11].
0

Se kSe ka kb kc kd ke kf S0e

12

ka aSbut

13
0:107

2:79 < d < 51 mm

0:157

51 < d < 254 mm

kb 1:24d

1:51d

kc 1:0Bending; 0:85Axial; 0:59Torsional

14
15

294

H.-S. Han / Engineering Failure Analysis 44 (2014) 285298

kd 0:975 0:432103 T F  0:115105 T 2F 0:104108 T 3F  0:5951012 T 4F

16

Here, ka is the surface condition modication factor (a = 1.58, b = 0.086 for grounded steel), kb is the size modication factor, kc is the load modication factor (=1.0 when using the effective von Misses stress), kd is the temperature modication
factor (=1.0 in this research), ke is the reliability factor (=1.0 in this research), kf is miscellaneous-effects modication factor
(=1.0 in this research), Se0 is the rotary beam test specimen endurance limit and Se is the endurance limit at the critical location of a machine part in the geometry and condition of use.
Through these stress values, the Soderberg plot can be drawn, as shown in Fig. 10, for the straight and zigzag maneuver
conditions of a ship at the keyway end of the reduction gear input shaft.
In Fig. 10, the safety analysis is performed when the notch factor for calculating resultant static stress is 0 and 0.15,
respectively, related to the maneuver conditions. The stress concentration factor is considered in the static load condition
because the shaft is high-strength steel and the maximum stress is under yield stress at the keyway.
Fig. 10(a) shows that the safety factor is not satised with the limit value of MIL G 17859D(1.75) under some operating
conditions when notch factor is 0. In Fig. 10(b), the safety factor is initially reduced when the notch factor is 0.15. In

(a) q s =0

(b) q s =0.15

(c) Safety factor as RPM

Fig. 10. Soderberg plot and safety factor of the reduction gear input shaft.

H.-S. Han / Engineering Failure Analysis 44 (2014) 285298

295

Fig. 10(c), the minimum safety factor is 1.36 and 1.21 at 1130 rpm when the notch factor is 0 and 0.15, respectively, in the
zigzag maneuver condition.
The Soderberg plot indicates that the safety factor is not sufciently related to MIL G 17859D and high-cycle fatigue can
occur since the safety factor can be additionally reduced according to the ship operating conditions and sea state.
4.2.4. Stress calculations using CAE
In this chapter, stress analysis is performed with CAE. MSC.Patran/Nastran is used for CAE in this research. The CAE is performed with two cases in which the assumptions of the boundary condition are different. The boundary condition of the 1st
case is that the torque is applied at the end of the input shaft and the opposite end (coupling assembling ange) is constrained. The boundary condition of the 2nd case is that the end of the ange is constrained and the torque is input to
the keyway surface.
(1) Case 1
The FEM model is shown in Fig. 11(a). The input torque is 37,703 N m at 1148 rpm of the engine. Fig. 11(b) shows the
resultant stress from CAE, in which the maximum stress occurs at the llet edge of the keyway.
Through the maximum shear stress theory, it can be found that the resultant stress is 325 MPa and the alternating stress
is 75.9 MPa when the alternating torque is 8810 N m.
The analysis of case 1 shows that the safety factor can be calculated to 1.58 based on the Soderberg safety equation at
1148 rpm, where the torsional vibration has a maximum value.
(2) Case 2
The FEM model is shown in Fig. 12(a), in which the maximum stress occurs at the end of the keyway similar to the real
fracture. The resultant static stress is 416 MPa, and the resultant alternating stress is 97.2 MPa when the alternating torque is
8801 N m.

(a) FEM model

Max. Stress Position

(b) Max. stress on the keyway: 325Mpa

Fig. 11. Result of linear static analysis for case 1.

296

H.-S. Han / Engineering Failure Analysis 44 (2014) 285298

The analysis of case 2 indicates that the safety factor can be calculated to 1.23 based on the Soderberg safety equation at
1148 rpm, where the torsional vibration has a maximum value.
Upon investigating the FEM results, the safety factor from case 2 appears to be 1.28 times lower than that from case 1, and
the safety factor from all cases cannot satised with the limit of MIL-G-17859D, which should be over 1.75.
In CAE, the stress concentration factor can be estimated by comparing the maximum stress of the shaft with a keyway to
that without a keyway, which is 1.92 for case 1 and 2.31 for case 2. These stress concentration factors were applied to the
resultant static stress as well as the alternating stress together, different to the stress concentration factor in MIL G 17859D.
Therefore, CAE considers a lower stress concentration factor for an alternating load and a higher one for a static load
rather than one factor for both.

5. Investigation results and cause of the failure

Through the instrumental analysis referred to in the previous chapter, the safety factor of the reduction gear input shaft
does not satisfy the requirements of MIL G 17859D, even though the safety factor from the Soderberg safety equation is over
1.0. Since a shaft can frequently become abnormally loaded over the design level when a ship is sailing, the safety factor
should have a sufcient margin to endure under such conditions.
Therefore, it can be concluded that the root cause of the shaft failure in this research is the poor design of the shaft with
respect to the maximum operating load.

6. Corrective action
Through the instrumental analysis about failure of the reduction gear input shaft, the failure seems to be caused by insufcient safety factors at the keyway end of the shaft under mean and alternating torque.

(a) FEM model

Max. Stress Position

(b) Max. stress on the keyway: 416Mpa

Fig. 12. Result of linear static analysis for case 2.

H.-S. Han / Engineering Failure Analysis 44 (2014) 285298

r=0.59mm

(a) r=0.59

r=1.0mm

(b) r=1.0

Max. Stress Position

(c) FEM result for r=1.0 (Max. stress on the keyway: 288Mpa)
Fig. 13. Increasing the llet radius on the keyway edge.

(a) d = 130.04

(b) d = 145

(c) FEM result for r=1.0 (Max. stress on the keyway: 240Mpa)
Fig. 14. Increasing the diameter of the shaft.

297

298

H.-S. Han / Engineering Failure Analysis 44 (2014) 285298

Methods to increase the safety factor are reducing the torsional vibration level or changing the design of the shaft to be
more healthy design. Since reducing an engines torsional vibration is more difcult than changing a shafts design, the latter
is investigated in this research.
Design changes are suggested for two cases, in which the llet radius on the edge of the keyway is increased and the
shafts diameter is increased.
Fig. 13 shows the design change for the llet radius on the edge of the keyway. When increasing the llet radius from 0.59
to 1.0 mm, the stress on the edge of the keyway is reduced from 325 to 288 MPa, and the safety factor increases from 1.58 to
1.78 for the boundary condition of case 1.
When increasing the diameter of the shaft from 130.04 to 145.0 mm, as shown in Fig. 14, the maximum stress is reduced
from 325 to 240 MPa and the safety factor increases from 1.58 to 2.2 for the boundary condition of case 1.
Therefore, increasing the shaft diameter is a more conservative correction action than increasing the llet radius on the
keyway edge in order to protect this failure.
7. Conclusion
RCA of the failure of the reduction gear input shaft in a typical naval vessel is investigated according to VDI 3822 in this
study. RCA shows that failure appears to be caused by poor design of the shafts keyway. The following conclusions are
derived from this study.
(a) Based on RCA of the failure of the reduction gear input shaft, it is estimated that the shaft cracks owing to torsional
fatigue caused by an insufcient safety factor (<1.75), which is suggested in the MIL G 17859D(SH) standard.
(b) When the notch factor calculated based on the stress concentration factor for the static load is considered to be 0.15
for the reduction gear input shaft, the safety factor reduces from 1.36 to 1.21 at 1130 rpm. Therefore, it is recommended that the notch factor for static load be considered when the shaft material is high-strength grade steel and
that plastic slip does not occur at the maximum stress to evaluate the safety of the shaft more conservatively.
(c) Through CAE, the safety factor is evaluated to be 1.58 when the torque is input at the end of the shaft constraining the
end of the opposite side of the shaft and 1.23 when the torque is input at the keyway surface constraining both ends of
the shaft.
(d) Even though the safety factor is over 1.0 for the reduction gear input shaft, the safety factor of the reduction gear input
shaft with a keyway is clearly not sufcient to satisfy the limit suggested in the MIL G 17859D(SH) standard, and it
may reduce more further the applied load on the shaft is varied according to the sea state and sailing condition of
the ship.
(e) Through design changes from CAE, it can be estimated that the safety factor increases from 1.58 to 1.78 when the llet
radius is increased from 0.59 to 1.0 mm and from 1.58 to 2.2 when the diameter of the shaft increases from 130.04 to
145 mm for the reduction gear input shaft. Through the corrected actions estimated using CAE, increasing the shaft
diameter is a more conservative corrective action than increasing the llet radius on the keyway edge to prevent
high-cycle fatigue and satisfy the limit of the safety factor (>1.75) suggested by the MIL G 17859D(SH) standard.

Acknowledgement
This research was performed in the Defense Agency of Technology and Quality, and DTAQ veried that it did not contain
any information related to military security.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]

Peterson RE. Stress concentration design factors. New York: John Wiley & Sons Inc.; 1953.
Okubo H, Hosono K, Sakaki K. The stress concentration in keyways when torque is transmitted through keys. Exp Mech 1968;8(8):3758.
Pedersen NL. Stress concentration in keyways and optimization of keyway design. J Strain Anal 2010;45:593604.
DIN 6885, 1968, Drive type fastenings without taper action; Parallel keys, keyways, deep pattern (FOREIGN STANDARD).
Goksenli A, Eryurek IB. Failure analysis of an elevator drive shaft. Eng Fail Anal 2009;16:10119.
Bhaumik SK, Rangaraju R, Parameswara MA, Venkataswamy MA, Bhaskaran TA, Krishnan RV. Fatigue failure of a hollow power transmission shaft. Eng
Fail Anal 2002;9:45767.
Parida N, Tarafder S, Das SK, Kumar P, Das G, Ranganath VR, et al. Failure analysis of coal pulverizer mill shaft. Eng Fail Anal 2003;10(2003):73344.
VDI-Gesellschaft Material Engineering. VDI 3822 Failure analysis: fundamental and performance of failure analysis; 2011.
D.o.D. MIL-G-17859D(SH): Gear Assembly, Production (Naval Shipboard Use); 1993.
ISO. IS0 10816-6: Mechanical vibration Evaluation of machine vibration by measurements on nonrotating parts Part 6: Reciprocating; 1995.
Shigley JE. Shigleys mechanical engineering design. Tata Macgraw-Hill Education; 2011.