Ocean Engineering 107 (2015) 212221

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

Estimate of the fatigue life of the propulsion shaft from torsional

vibration measurement and the linear damage summation law in ships
HyungSuk Han n, KyungHyun Lee, SungHo Park
Naval System Research Team, Busan Center, Defense Agency for Technology and Quality, 525-2,Gwangan 1 dong, Suyeong-gu, Busan 613-808, Republic of
Korea

art ic l e i nf o

a b s t r a c t

Article history:
Received 7 December 2014
Accepted 14 July 2015
Available online 26 August 2015

Most ships use a diesel engine for the 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.
Ships that require high power and revolution speed usually have V-type, 4-stroke diesel engines and
reduction gears to increase the output torque. Therefore, a robust design of the shaft is required for this
type of vessel. In this research, the fatigue stability and life cycles of the shaft are estimated with
Soderberg's safety evaluation method and the linear damage summation law based on the torsional
vibration data. When estimating them, non-standard sailing conditions such as starting the engine and
zigzag maneuvers are included in addition to normal sailing conditions such as straight maneuvers.
& 2015 Elsevier Ltd. All rights reserved.

Keywords:
Rain ow cycle counting
Linear damage summation law
Torsional vibration

1. Introduction
Diesel engines are widely used as low-speed propulsion
engines on ships. Since diesel engines operate from the explosive
force of the cylinder connected to the crank shaft, the vibratory
torque of the diesel engine is larger than that of other types of
engines such as gas turbines and electric motors. Ships that
require high speed and torque usually use V-type, 4-stroke,
high-speed engines, and reduction gear is adopted in order to
obtain high torque. The propulsion shaft of this type of engine can
fracture due to the high torque and speed. Many studies have been
conducted on shaft fractures due to dynamic loads at high-stress
areas such as the llet, chamfer, and keyway.
Okubo et al. (1968) performed the torsional fatigue test with
2 different test keyway shaft specimens with various llet radiuses
of the key and the keyway and suggested stress concentration
factors. The test results showed that the stress concentration factor
of the end of the key was larger than that of the keyway. In
addition, the researchers suggested the stress concentration factor
of the keyway could be modied by decreasing the rigidity caused
by the addition of the keyway to the shaft. Pedersen (2010)
suggested shaping the llet of the keyway in a super ellipse and
found from nite element analysis (FEM) that the stress

Corresponding author. Tel.: 82 51 750 2533; fax: 82 51 758 3992.

E-mail address: daerihan@hanmail.net (H. Han).

http://dx.doi.org/10.1016/j.oceaneng.2015.07.023
0029-8018/& 2015 Elsevier Ltd. All rights reserved.

concentration factor was reduced about 50% compared to the

circular llet. Bhaumik et al. (2002) investigated the fracture of the
stage helical gearbox of a low-speed hollow shaft transmitting the
engine load through the key and keyway and used visual inspection and a scanning electron microscope (SEM) to nd that
fracture was caused by torsional fatigue. He reported that the
fracture of the shaft was caused by the small llet radius under
specication, improper machining, and MnS included in the
material lattice. JianPing, and Guang (2008) reported the fracture
of the gear shaft of the extruder. By measuring the torsional
vibration of the gear shaft, they found the root cause of the
fracture was the extreme static load and veried it with nite
element analysis and Goodman fatigue criteria. Parida et al. (2003)
studied the root cause of fatigue fracture for the keyway of the ball
and race-type coal pulverizer and, using SEM analysis and
mechanical tests such as the tensile and Charpy impact tests,
found that the fracture was caused by anon-standard heat treatment process that reduced the endurance and toughness of the
material. Goksenli and Eryurek (2009) investigated the fracture of
the drive shaft of the elevator with a keyway and estimated the
fracture was caused by the combined effect of repetitive torsion
and a bending load. The stability of the shaft was evaluated with
the Goodman criteria, and the life cycle of the shaft was estimated
with the SN curve reecting the mean stress, and the fatigue limit,
which includes shaft size changes and surface modication factors,
was recalculated. In the investigation, the fracture was caused by
the stress concentration at the small llet at the end of the

H. Han et al. / Ocean Engineering 107 (2015) 212221

keyway, and the researchers suggested corrective actions to

increase llet radius at the keyway, which was veried with FEM
analysis. Han et al. (2012) investigated the root cause of the
fracture of the coupling that connects the gas turbine and the
reduction gear in a ship. In the investigation, the fracture was not
caused by the misalignment of the shaft itself but by the
independent movement of the gas turbine and reduction gear
supported by the independent resilient mounting system when
the ship was sailing in rough seas. Therefore, a corrective action
was suggested to add horizontal mounts that can reduce the
movement of the gas turbine and reduction gear. In these studies,
the shaft fractured frequently in the industrial eld, and had
various root causes such as design error, manufacturing error,
extreme operating conditions, and other external causes. Ma and
Wang (2006) researched the propeller shaft for the high speed
craft whose material is SUS630. He found out that the fracture of
the propeller shaft was caused by pitting corrosion and the life
cycle was reduced about maximum 27%. Arisoy et al. (2003)
reported the fracture of the 174 ph precipitation hardening
stainless steel propeller shaft installed in a sailboat working in
marine environment. He found out that the fracture of the
propeller shaft was broken because the stress corrosion cracking
which is progressed transgranulary in the martensitic matrix was
occurred in the propeller shaft under the serious vibratory torque.
Fonte et al. (2011) performed the failure analysis of two helical
gear wheels of a ducted azimuth thruster. Through SEM (scanning
electron microscope), it can be found out that the fracture was
caused by fatigue and the root cause of this fatigue fracture is the
inappropriate lubricating of the gear shaft.
In this study, effective experimental evaluations of shaft safety
for torsional vibration are described with the input shaft of the
reduction gear connected to the diesel engine. The torsional
vibration was measured directly with a strain gauge and a
telemetry system. The life cycles under the measured stress
conditions were estimated using rain ow cycle counting and
the linear damage summation law, and the results were compared
to a standard stability evaluation method such as the Soderberg
criteria. In this investigation, transient and non-standard ship
sailing conditions such as zigzag maneuvers and engine starting
were also examined in order to evaluate the safety of the shaft.

2. Measurement of the torsional vibration

213

Fig. 1. Telemetry system measuring TV, (a) test setup, (b) schematic diagram of the
telemetry system.

Table 1
Test system.
Test system

Maker

Model

FFT Analyzer
Strain gauge
Telemetry
Tachometer

B&K
MM
Binsfeld engineering Inc.
Monarch instrument

Pulse 3053-B12/0
CEA-06-250US-350
TT 10K-LP
ROLS-P

2.1. Measurement setup

The fatigue stability is generally dependent on the static as well
as alternating stress. The static stress of the propulsion shaft is
caused by the mean torque of the engine and the alternating
torque of it is caused by the vibratory torque of the engine. Since a
diesel engine is operated by the force of the cylinder from the
explosion of the gas, the vibratory torque of the diesel engine is
bigger than that of the other type of engines such as gas turbine
and electric motor. The alternating stress as well as the vibratory
torque of the shaft system is called torsional vibration and it
should be restricts to avoid torsional fatigue failure. The torsional
vibration can be measured and evaluated by the elector torsiograph, laser torsiograph, strain gauge and so on. In this research,
the vibratory torque of the reduction gear input shaft connected to
the diesel engine was measured with a full bridge shear strain
gauge and a telemetry module as shown in Fig. 1, and the rotating
speed (rpm) of the shaft was measured simultaneously. The
telemetry system is consists of the transfer and receiver module.
The full bridge shear strain gauge is connected to the transfer
module through the Wheatstone bridge and the strain data is
transferred to the receiver module as shown in Fig. 1(b). The data

from the receiver module was collected with a data acquisition

system. The specications of the strain gauge, telemetry module,
and data acquisition system are shown in Table 1.
2.2. Measurement of torsional vibratory torque and engine vibration
Since the exhaust pipe of the diesel engine discussed in this
research is installed near the free surface of the sea as shown in
Fig. 2, the torsional vibratory torque can be increased when the
back pressure varies more than the specied limit because of
variations in the draft line, for example, in rough seas. This kind of
design of the exhaust line for the diesel engine is usually applied
to the special ship such as naval vessel.
Changes in the torsional vibration of the engine caused by
changes in the back pressure have been studied for submarines in
snorkeling conditions (Mann, 2011; Hield, 2011). Studies showed
that the torsional vibration varied according to the wave height
and period because the exhaust pipe of the diesel engine is located
under the free surface.
Figs. 3 and 4 show the torsional vibratory torque during
straight maneuver sailing conditions as well as zigzag maneuver

214

H. Han et al. / Ocean Engineering 107 (2015) 212221

bar
0.14

Outlet of the exhaust

Back Pressure[bar]

0.12
0.10
0.08
0.06
0.04

Fig. 2. Outlet of the exhaust for a diesel engine of a typical ship.

0.02
0.00
60

80

100

time[s]

kNm

Fig. 3. Vibratory torque in accordance with speed at straight maneuver.

Alternating Torque[kNm]

110

100

90

80

70
60

80

100

time[s]
Fig. 5. Backpressure and torsional vibration at the propeller shaft at the straight
maneuver condition, (a) back pressure, (b) vibratory torque at the propeller shaft.

Fig. 4. Vibratory torque in accordance with speed at zigzag maneuver.

conditions at various sailing speeds (power control level (PCL) 4

10). In PCL 49, the torsional vibratory torques of the straight and
zigzag maneuvers are almost the same. However, the torque
increased dramatically at the PCL 10 of the zigzag maneuver and
varied with time. This situation may occur because the back
pressure variations are larger than specied. In the zigzag maneuver, the inclination of the ship varies, and the exhaust pipe can
be located under and above the free surface repeatedly. Therefore,
variation in the back pressure of the exhaust pipe at the zigzag
maneuver can be higher than that during straight maneuvers.
Figs. 5 and 6 show the back pressure in the exhaust pipe of the
diesel engine as well as the torsional vibration of the propeller
shaft at straight and zigzag maneuver conditions respectively.
When the ship sails straight, the back pressure is constant and
the torsional vibration of the propeller shaft is also constant as
shown in Fig. 5(a) and (b). However, as shown in Fig. 6(a) and (b),
it can be found that the back pressure in the exhaust pipe of the
diesel engine varied dramatically when the ship sails zigzag and

the torsional vibration of the propeller shaft also varied in coincide

with back pressure variation. In Fig. 6, it can be found that the
mean torque of the propeller shaft is reduced due to the reduction
power of the diesel engine caused by increasing back pressure. In
addition, it can be found that the alternating torque is increased
dramatically when this mean torque and back pressure are varied
instantaneously.
In addition to the backpressure, the variation of the propeller
exciting force at zigzag maneuver condition can affect the variation of the torsional vibration of the shaft. However, it is very
difcult to classify how much affect the variation of the engine
exciting force caused by the backpressure and that of propeller
exciting force at zigzag maneuver condition to the torsional
vibration respectively.
In this research, it is assumed that the variation of the torsional
vibration from the propeller exciting force variation occurs together
with that from the backpressure variation when the ship sails zigzag.
Therefore, the resultant torsional vibration when the backpressure
variation increases at zigzag maneuver condition includes the torsional vibration caused by the variation of the propeller excited force
even though it cannot be identify separately.

H. Han et al. / Ocean Engineering 107 (2015) 212221

215

bar
0.14

Back Pressure[bar]

0.12
0.10
0.08
0.06
0.04
0.02
0.00
20

40

time[s]

kNm

Alternating Torque[kNm]

110

100

90

Fig. 8. Vibration velocity of the engine at straight maneuver, (a) transverse, (b) vertical.

80

70
20

40

time[s]
Fig. 6. Backpressure and torsional vibration at the propeller shaft at the zigzag
maneuver condition, (a) back pressure, (b) vibratory torque at the propeller shaft.

Fig. 7. RPM variation according to sailing conditions

Fig. 7 is the rotating speed variation of the diesel engine during

straight and zigzag maneuvers. In Fig. 7, the variation in rotating
speed during the zigzag maneuver is larger than that during the
straight maneuver, and could be caused by the back pressure
variation. Figs. 8 and 9 are the vibration spectra of the diesel engine.
The accelerometers are attached on the diesel engine to measure the

Fig. 9. Vibration velocity of the engine at Zigzag maneuver, (a) transverse, (b) vertical.

216

H. Han et al. / Ocean Engineering 107 (2015) 212221

conditions (e.g., straight maneuver condition) as well as non-standard

conditions (e.g., zigzag maneuver and engine starting condition).

3. Investigation
3.1. Fatigue stability of the shaft
Based on the measured torsional torque, the mean stress and
the alternating stress of the shaft should be calculated to evaluate
the fatigue stability of the shaft. In this research, the mean stress
and the alternating stress are calculated following the procedure
in MIL G 17859D Appendix D (D.o.D, 1993).
The mean stress (Sr) of the shaft from static torque of the diesel
engine can be written as given in Eq. (1) (Fonte et al., 2011)
according to the maximum shear stress theory.
Sr

Fig. 10. The variation of the vibration of the diesel engine, (a) transverse, (b) vertical.

q
K f ;c USc 2 2 UK f ;Ts U Ss 2 ; Sc

Tt
d =4
2

engine vibration, and the vibration spectrum is collected for 60 s. All

of the spectra for 60 s are depicted together in Figs. 8 and 9. In
Figs. 8 and 9, the vibration velocity at the zigzag maneuver is much
larger than that at the straight maneuver below 10 Hz and it can be
estimated that it is the result of the variation of the torsional vibration
of the shaft caused by the back pressure variation.
Fig. 10 shows the variation of the overall vibration level for the
diesel engine during the straight maneuver and the zigzag
maneuver. In Fig. 10, the variation in the vibration increases when
the ship maneuvers in the zigzag condition.
In these investigations, it can be found that the increasing
torsional vibration during the zigzag maneuver condition is caused
by variations in the back pressure in the exhaust pipe.
When the diesel engine starts, the torsional vibration increases
dramatically within a short period and then becomes steady state,
as shown in Fig. 11. Therefore, the engine starting condition is
dened as a non-standard operating condition in addition to the
zigzag maneuver condition.
In this research, the fatigue stability and life of the reduction gear
input shaft of the diesel engine were investigated at standard operating

where Sc is the compressive stress, Tt is total thrust force in the

axial direction, Ss is the shear stress from torsion, QT is the mean
torque of the diesel engine, Zt is the polar modulus of the shaft
section, Kf,c is the effective stress concentration factor for the
compressive force and Kf,Ts is the effective stress concentration
factor for the mean torque.
The stress concentration factor under the static load for the
ductile material is not usually applied since the strain hardening
occurs near the stress concentration point by the plastic ow.
Therefore, in MIL G 17859D, the stress concentration factor of the
compressive force and the mean torque are ignored (Kf,c, Kf,Ts 1.0).
However, the stress concentration factor cannot be ignored under
the yield stress where the plastic slip does not occur. Therefore,
the stress concentration factor for the static load under the yield
stress is partially applied in this research.
The real stress concentration can be written as in Eq. (2)
K f 1 qK t  1

Fig. 11. Vibratory torque at engine starting state.

QT
d
; Zt
Zt
16

; Ss

where Kf is effective stress concentration factor, Kt is ideal stress

concentration factor and q is notch sensitivity index.
The engine and the reduction gear studied in this research
transfer the engine torque to the reduction gear by the key and the
keyway of the input shaft as shown in Fig. 12, where d 130 mm,
r 0.59 mm, b 31.9 mm, and t1 10.85 mm. When the key and
the keyway are applied to the shaft, the stress concentration factor

Fig. 12. Shaft with keyway.

H. Han et al. / Ocean Engineering 107 (2015) 212221

should be applied as given in Eqs. (3) and (4) (Peterson, 1953).


K t;B 1:426 0:1643

0:1
0:1
 0:0019
r=d
r=d

2
; 0:005 r r=d r 0:04; d r 165:1 mm

3

K t;T 1:953 0:1434

0:1
0:1
 0:0021
r=d
r=d

4
where K t;B is the ideal stress concentration factor for the bending
moment, K t;T is the ideal stress concentration factor for the
torsion, r is the llet radius, and d is the shaft diameter.
Through Eqs. (3) and (4), the stress concentration factor of the
reduction gear input shaft can be calculated as 4.12 for bending
and 4.093 for torsion.
In Eq. (2), the real stress concentration factor can be obtained
by dening the notch factor, which is dened from 0.0 to 1.0. In a
static load, the notch factor is 0.0 if the stress concentration factor
is ignored, and the real stress concentration factor becomes
1.0 from Eq. (2).
When the material is brittle (elongation o 5%), the stress
concentration factor should be considered even though the load
is static. In this case, a notch factor of about 0.150.25 is applied
(Peterson, 1953). When the maximum stress is calculated with
FEM for the reduction gear input shaft with and without a keyway,
how much stress is increased by adding the keyway to the shaft is
found, as shown in Fig. 13. Assuming that the ratio of the stress for
the shaft with and without the keyway is the stress concentration
factor under the static load, it can be dened as 1.93 based on the
FEM results in Fig. 13 and the notch factor under a static load can
be calculated to 0.3 from Eqs. (2) and (4).
When the load alternates, the stress concentration should be
considered, and the notch factor becomes almost 1.0. The resulting
alternating stress can be written as Eq. (5)
Sar

The notch factor under alternating bending and axial loads can
be represented as given in Eq. (6), and the alternating torsional
load can be represented as given in Eq. (7) (Peterson, 1953)
q

1
; for bending & axial loading
1 =r

1
; for torsional loading
1 0:6=r

2
; 0:005 r r=d r 0:07

q
M
d3
Ta
; Sas
K f ;B Sb 2 2 U K f ;T Sas 2 ; Sb b ; Z
Ss
Z
32
T max

217

where is the material constant; quenched and tempered steel

is 0.0025, annealed and normalized steel is 0.01, and aluminum
alloys is 0.025.The notch factor of the reduction gear input shaft in
this research can be estimated to 0.903 for bending and 0.939 for
torsion respectively from Eqs. (6) and (7).
When the steady stress and the alternating stress are calculated, the fatigue stability of the shaft under repetitive load can be
dened using the Goodman, Soderberg, and Geber criteria, which
shown in Fig. 14. In this research, the Soderberg method was
adopted since MIL G 17859D suggested the method for evaluating
shaft stability under repetitive load as given in Eq. (8). In Fig. 14, it
can be known that the Soderberg criterion is the most conservative one among those 3 criteria
S:F S

Sar
SY SF
r

where S.F is the safety factor, SY is yield stress and SF is

fatigue limit.
MIL-G-17859D requires a safety factor in Eq. (8) of 1.75 for
surface ship and 2.0 for submarine. Fig. 15 shows the safety factor
of the input shaft of the reduction gear for the diesel engine
according to the sailing condition and speed. Fig. 15(a) shows the
result when the stress concentration factor for the static load is
ignored, and Fig. 15(b) shows the result when the stress

where Sar is the resultant alternating stress, Sb is the alternating

bending stress caused by weight of the shaft and lateral vibration
of it, Sas is the shear stress from alternating torsional vibratory
torque, Kf,B and Kf,T is the real stress concentration factor from
bending and torsional stress respectively, Mb is the bending
moment caused by lateral vibration and shaft weight, Z is the
modulus of the shaft section and Ta and Tmax are the alternating
and maximum steady torque respectively.

Fig. 14. Goodman, Soderberg and Gerber Criteria for the Fatigue Stability (SF :
Fatigue limit, SY :Yield stress,SU : Ultimate stress, Sar : Alternating stress,Sr : Mean
stress, S.F: Safety factor).

Max. stress

Fig. 13. Maximum stress for the shaft with keyway estimated by FEM.

218

H. Han et al. / Ocean Engineering 107 (2015) 212221

shaft of the navy vessel is MIL 167-2, Type III (D.o.D, 1976). In MIL
167-2, Type III, the limit of the stress from the vibratory torque is
1/25 of the ultimate tensile stress for steel and 1/6 of the fatigue
limit for nonferrous metal. However, the denition of the maximum stress does not describe whether it is pure torsional or vonMises stress. In addition, there is no denition of the maximum
stress whether stress concentration factor is applied or not.
If the limit of the stress is von-Mises stress and includes the
stress concentration factor, it is a very severe specication. However, it is the opposite if the limit is pure torsional stress and does
not include the stress concentration factor.
IACS M68 (2012) suggested vibratory torsional stress in the
propulsion shafts such as intermediate and propeller shafts as
given in Eqs. (9) and (10), and global shipping registries such as
ABS, DNVGL, and Lloyd also restrict vibratory torsional stress same
as Eqs. (9) and (10)(ABS, 2014; DNVGL, 2014; LLOYD, 2014).

c
c

Fig. 15. Safety factor for straight and zigzag maneuver for a typical ship (a) qs 0
and (b) qs 0.3.

B 160
18

B 160
18

C k C D 3  2 for o 0:9
2

C k C D 1:38 for 0:9 r o 1:05

9
10

In Eqs. (9) and (10), c is the allowable limit of the shear stress
in continuously operating condition (Mpa), B is specied minimum tensile strength in Mpa of the shaft material, Ck is factor for
the particular shaft design features ( 1.45/scf), CD is size factor
 0:2
( 0:35 0:93 d0 ), do is the shaft outside diameter in mm, scf is
stress concentration factor, is the speed ratio, n is the speed in
revolutions per minute under consideration and n0is the speed in
revolutions per minute of shaft at rated power.
Fig. 16 is the pure shear stress of the reduction gear input shaft
shown in Fig. 1 caused by the torsional vibration and depicted with
the allowance limits of MIL 167-2 Type III and IACS M68 together.
The stress level of the reduction gear input shaft is lower than the
limit of the MIL 167-2 Type III assumed that the specied stress is
the pure shear stress of the shaft and ignored the stress concentration factor. If the stress is dened to include the stress
concentration factor as shown in Fig. 16, the stress level under
almost all conditions(within the 9001350 rpm range) cannot be
satised with the suggested level according to MIL 167-2 A Type
III. When the limit of IACS M68 is applied, the shear stress level at
the maximum speed under the zigzag operating condition is not
satised by the suggested limit. In IACS M68, the maximum
ultimate stress is used until 800 MPa even though the shaft used
the material with ultimate tensile strength over than 800 MPa.
Therefore, this limit is also shown in Fig. 16.
Comparing Figs. 15 and 16, the result for the Soderberg method
agrees with that from IACS M 68 rather than MIL 167-2, Type III.

Fig. 16. Alternating stress for straight and zigzag maneuver for a typical ship.

3.2. Rain ow cycle counting

concentration factor for the static load is considered (notch
factor0.3). In Fig. 15(a), the shaft is safe in all of the operating
conditions even though it does not have a sufcient margin for the
safety factor (S.F 1.02 at maximum speed, zigzag maneuver).
However, the shaft is not safe at the maximum speed, zigzag
maneuver condition as shown in Fig. 15(b) (S.F0.83) when the
notch factor applied 0.3 is for the static load.
When the 1.75 safety factor limit of MIL G 17859D is included
for the surface ship, the input shaft of the reduction gear for the
diesel engine cannot be satised with the limit at 940 rpm and the
maximum speed ignoring the notch factor in the static load, and it
cannot be satised with a limit from 940 rpm to maximum speed
when the notch factor applied is 0.3.
MIL G 17859D is the specication of the reduction gear for the
propulsion system. The general specication of the propulsion

Since the fatigue stability of the shaft using Soderberg method

is calculated with mean static and alternating stress, the force that
occurs instantaneously cannot be considered. Therefore, the fatigue stability for the time recorded stress should be evaluated by
calculating the fatigue damage with each stress data according to
its amplitude and counted cycles.
The rain ow cycle counting method developed by Matsuishi
and Endo (1968) is widely used for counting the time recording
stress signal. The procedure is shown in Fig. 17, which counts the
mean and alternating stress according to their amplitudes. The
peak and valley of the recorded stress are identied and rearranged according to their amplitude as shown in Fig. 17. Then, the
amplitude of the mean and alternating amplitude as well as
repetitive numbers are counted as shown in Fig. 17.

H. Han et al. / Ocean Engineering 107 (2015) 212221

219

Fig. 19. SN curve of the material,

3.4. SN curve

Fig. 17. Rain-ow cycle counting.

Stress

Stress

In Fig. 19, stress amplitude with 103 cycles and 106 cycles is
necessary to draw the SN curve of the material. The 103cycle stress
can be represented as given in Eq. (16):
b
0 
log 0F =S0e
103 f USu ; f F 2  103 ; 0F Su 345 Mpa; b 
Su
log 2N e
16
N1

N2
N3

Time

Time

Fig. 18. Schematic diagram for calculating the damage with PalmgrenMiner's
linear damage summation law.

3.3. Linear damage summation law

The life cycles under fatigue can be broadly estimated using the
linear damage summation law developed by PalmgrenMiner (Miner,
1945). Following this law, the random stress signals are grouped
according to their amplitudes as shown in Fig. 18, and the damage can
be calculated with Eq. (11). In Eq. (11), the damage at a typical stress
can be dened as the ratio of the number of the typical stress and
fatigue failure occurrences. In accordance with this law, the total
damage can be dened as the sum of each type of damage, and if it is
equal to 1.0, fracture occurs.
DT

N
X
N1 N2 N3
Ni

Nf 1 Nf 2 Nf 3
N
i 1 fi

11


1
log 10 SY  log 10 S0F
3

Se kS0e ka kb kc kd ke kf S0e

17

ka aSbut

18

kb 1:24d

 0:107

1:51d

 0:157

2:79 od o51 mm
51 o d o 254 mm

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

12
13

log 10 B 3a log 10 SY

14



m
USF
S0F 1 
SY

15

where SY is the yield stress, SF is the fatigue limit under fully reversed
force and SF' is the fatigue limit under the typical mean stress ( m ).

19
20

kd 0:975 0:43210  3 T F 0:11510  5 T 2F

0:10410  8 T 3F  0:59510  12 T 4F

where DT is total damage, Ni is the measured counting number in

accordance with stress i and N is the counting number that fatigue
occurs under the stress i, which the relationships between them are
represented as shown in Eqs. (12)(15)

 
1
log 10 N f i  log 10 i  log 10 B
a

where 103 is the stress having 103 fatigue life cycles, 0F is the stress
fractured under 1 cycle fully reversal load, Su is the ultimate tensile
strength,S0e is the fatigue limit( 0.5Su) and Ne is the innite life
cycle( 106 Cycles).
Since the material of the shaft is 826M40(nickel chromium
molybdenum steel) and the ultimate tensile strength measured
with the shaft material is 981 MPa, the 103 fatigue life cycles stress
can be calculated as 795 MPa from Eq. (16).
Fatigue limits with 106 cycle stress can be represented as given
in Eqs. (17)(22)

ke 1:0 50%; 0:89 90%; 0:87 95%; 0:84 98%;

0:81 99%; 0:75 99:9%; 0:7 99:99%

21

22

where ka is the surface condition modication factor(a1.58,

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
Mises stress), kd is the temperature modication factor(1.0 here), ke
is the reliability factor(1.0 here), kf is the miscellaneous-effects
modication factor(1.0 here), Se 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.
In this research, the surface and size modication factors are
included, and the reliability and temperature modication factors
are ignored. When the reliability modication factor is considered,
the fatigue limit decreases additionally as increasing the probability of the fatigue fracture. Since the fatigue limit becomes so

220

H. Han et al. / Ocean Engineering 107 (2015) 212221

Fig. 20. Soderberg plot for straight and zigzag maneuver at the max. speed for a
typical ship (qs 0), (a) straight maneuver(qs 0), (b) zigzag maneuver(qs 0).
Fig. 21. Soderberg plot for straight and zigzag maneuver at the max. speed for a typical
ship (qs 0.3), (a) Straight maneuver (qs0.3), (b) Zigzag maneuver (qs 0.3).

conservative considering this reliability modication factor, it is

not applied in this research and ke is set to 1.0.
Since the von-Mises stress already considered the load modication factor for torsion, the load modication factor is not
applied to calculate the fatigue limit. Therefore, the fatigue limit of
the reduction gear input shaft in this research can be calculated as
304 MPa from Eq. (17).

3.5. Calculation of the life cycles

When the measured torque is transferred to the stress as given
in Eqs. (1) and (3), the stress amplitudes can be counted by
counting the rain ow cycle, and the total damage can be
calculated with the linear damage summation law as described
in Sections 3.1 and 3.2. The SN curve dened in Section 3.4 is used
to calculate the damage from the linear damage summation law.
In this section, the fatigue stability is investigated with the
Soderberg plot and then, the total damage of the reduction gear
input shaft is calculated with the stress data measured in straight
and zigzag maneuver conditions under various speeds including
the engine starting condition.
Figs. 20 and 21 are the Soderberg plots for the stress data
recorded in 60 s under straight and zigzag maneuver conditions at
maximum speed that include the stress concentration factor of the
mean stress or not, respectively.
In Fig. 20, the shaft is safe under all operating speeds in the
straight maneuver condition even though the shaft does not
satisfy the required safety factor in MIL G 17859D (S.F 41.75).
However, the shaft is not safe when sailing at maximum speed
under zigzag maneuver condition.

Table 2
Estimation of the life cycle of the shaft.
PCL

Starting
4
5
6
7
8
9
10

Life time[H]
Straight maneuver

Zigzag maneuver

qs 0

qs 0.3

qs 0

qs 0.3

Innite
Innite
Innite
Innite
Innite
Innite
Innite
Innite

Innite
Innite
Innite
Innite
Innite
Innite
Innite
Innite

Innite
Innite
Innite
Innite
Innite
Innite
Innite
858.3

Innite
Innite
Innite
Innite
Innite
Innite
Innite
16.6

In Fig. 21, the plot moves to the right side when a notch factor
of 0.3 is applied, and the unsafe region in the zigzag maneuver
condition increases.
Table 2 shows the life cycles calculated with the linear damage
summation law for the engine starting, straight, and zigzag
maneuver conditions that include the stress concentration factor
for the mean stress or not. In Table 2, the shaft does not have
innite life cycles under the zigzag maneuver condition at the
maximum sailing speed. In addition, the life cycle that includes the
stress concentration factor for mean stress is decreased about 11
times lower than the life cycle that ignores the stress

H. Han et al. / Ocean Engineering 107 (2015) 212221

concentration factor. Therefore, the safety evaluation of the shaft is

very conservative when the stress concentration factor for static
mean stress is included.
These investigations showed that evaluating the safety of the
shaft with the Soderberg method is very difcult when the mean
stress and the alternating stress vary randomly. Therefore, in this
case, the rain ow cycle counting method is very useful for
classifying these randomly varied stress signals. In addition, the
life cycle can be estimated with the linear damage summation law
when the safety factor is less than 1.0.

221

method uses the mean stress for randomly varied stress, a

conservative evaluation of the safety of the shaft can be made.
Acknowledgment
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.
Reference

4. Conclusion
The safety and life cycles of the reduction gear input shaft for
diesel engines were investigated with the Soderberg plot, rain ow
cycle counting, and linear damage summation law for ships. The
following conclusions are derived from this study.
1) When the backpressure in the exhaust pipe of the diesel engine
is dramatically increased depending on the sailing condition, it
can be found that the variations in the torsional vibration,
revolution speed of the shaft and acceleration of the diesel
engine are also increased. Therefore, sailing conditions in
which the back pressure varies dramatically should be minimized and avoided if possible by reducing the ships speed.
2) When the stress concentration factor (qs 0.3) was considered
for static mean stress, the safety factor from the Soderberg
evaluation method decreased 1.1 times, which varied from 1.02
to 0.83, and the life cycles decreased 52 times, which varied
from 858.3 h to 16.6 h.
3) In the zigzag maneuver condition at maximum speed, the
Soderberg safety factor could be less than 1.0 depending on
the sea conditions. When the stress was evaluated with IACS
M68, this condition was unsafe.
4) The Soderberg evaluation method is difcult to apply when the
amplitude of the mean stress and the alternating stress varies
randomly. Evaluating the safety of the shaft with IACS M68 may
be very effective.
5) Estimates of the life cycle with the linear damage summation
law with rain ow cycle counting of the stress data can be used
to evaluate the safety of the shaft and nd how long it can
endure the condition under torsional vibration. Since this

ABS, 2014. Rule for building and classing: Steel vessel.

Arisoy, C.F., Basman, G., Sessen, M.K., 2003. Failure of 17-4 Stainless steel sailboat
propeller shaft. Eng. Fail. Anal. 10 (6), 711717.
Bhaumik, S.K., Rangaraju, R., Parameswara, M.A., Venkataswamy, M.A., Bhaskaran,
T.A., Krishnan, R.V., 2002. Fatigue failure of a hollow power transmission shaft.
Eng. Fail. Anal. 9, 457467.
DNVGL, 2014. Rules for classication of ships.
D.o.D, 1993. MIL-G-17859D(SH): Gear Assembly, Production (Naval Shipboard Use).
D.o.D, 1976. MIL 167-2: Mechanical Vibration of Shipboard Equipment (Reciprocating Machinery and Propulsion System and Shafting).
Fonte, M., Reis, L., Freitas, M., 2011. Failure analysis of a gear wheel of a marine
azimuth thruster. Eng. Fail. Anal. 18 (7), 18841888.
Goksenli, A., Eryurek, I.B., 2009. Failure analysis of an elevator drive shaft. Eng. Fail.
Anal. 16, 10111019.
Han, H.S., Lee, K.H, Park, S.H., 2012. Analysis for the effect of the misalignment of
the power line from the displacement caused by the ship motion of the main
propulsion system. Trans. Korean Soc. Noise Vib. Eng. 22 (1), 4652.
Hield P., 2011. DSTO-TR-2531: the effect of back pressure on the operation of a
diesel engine, Defense Science and Technology Organization.
IACS, 2012. M68: Dimension of propulsion shafts and their permissible torsional
vibration stresses.
JianPing, J., Guang, M., 2008. Investigation on the failure of the gear shaft connected
to extruder. Eng. Fail. Anal. 15, 420429.
LLOYD, 2014. Rules and regulations for classication of naval ships.
Ma, F.Y., Wang, W.H., 2006. Fatigue crack propagation estimation of SUS 630 shaft
based on fracture surface analysis under pitting corrosion condition. Mater. Sci.
Eng.: A 430 (12), 18.
Mann, J.W., 2011. Twin-turbocharged diesel performance under snorkeling conditions, UDT 2011 in London.
Matsuishi, M., Endo, T., 1968. Fatigue of Metals Subjected to Varying Stress.
Japanese Society of Mechanical Engineers, Jukvoka.
Miner, M.A, 1945. Cumulative damage in fatigue. J. Appl. Mech. 67, A159A164.
Okubo, H., Hosono, K., Sakaki, K., 1968. The stress concentration in keyways when
torque is transmitted through keys. Exp. Mech. 8 (8), 375378.
Parida, N., Tarafder, S., Das, S.K., Kumar, P., Das, G., Ranganath, V.R., Bhattacharya, D.
K., 2003. Failure analysis of coal pulverizer mill shaft. Eng. Fail. Anal. 10,
733744.
Pedersen, N.L., 2010. Stress Concentration in Keyways and Optimization of Keyway
Design. J. Strain Anal. 45, 593604.
Peterson, R.E., 1953. Stress Concentration Design Factors. John Wiley & Sons, Inc.,
New York.