Engineering Failure Analysis 92 (2018) 466–479

Contents lists available at ScienceDirect

Engineering Failure Analysis

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

Failure analysis of crankshafts used in maritime V12 diesel engines

T
a b c,⁎ b
Joao Gomes , Narciso Gaivota , Rui F. Martins , P. Pires Silva
a
Faculty of Sciences and Technology, Universidade Nova de Lisboa, Department of Mechanical and Industrial Engineering, Campus de Caparica, 2829-
516 Caparica, Portugal
b
Escola Naval, Base Naval de Lisboa, Alfeite, 2810-001 Almada, Portugal
c
UNIDEMI, Faculty of Sciences and Technology, Universidade Nova de Lisboa, Department of Mechanical and Industrial Engineering, Campus de
Caparica, 2829-516 Caparica, Portugal

A R T IC LE I N F O ABS TRA CT

Keywords: Maintenance of equipment requires constant monitoring of the components that constitute a
Maritime V12 diesel engines mechanical system, as well as the monitoring of the conditions of service, among others.
Crankshaft's failure One first indication of failure in a crankshaft is given by the low-pressure value of the lu-
Failure analysis brication circuit. This is mainly due to the accumulation of debris in the lubrication channels,
Crankshaft redesign
which causes the oil filters to be clogged. As such, this will cause poor lubrication of the
Finite element analysis
crankshaft, which can consequently cause its catastrophic failure, and frequently originates da-
mage propagation to other components of the engine, namely crankcase, bearing shells, con-
necting rods, pistons and other mechanical parts.
In the 4-stroke internal combustion maritime V12 diesel engine herein studied, there has been
a frequent failure of crankshafts. The seven cases of failure of the crankshafts reported in the last
25 years are presented in the paper and several causes of failure were listed. From those, the
influence of initial imperfections in the material was discussed in the article, as well as the
influence of the loadings applied to the crankshaft. Hence, the theoretical dimensioning of the
crankshaft was firstly assessed assuming the conservative Soderberg criterion and the crankshaft
model was then analysed using the Finite Element Method, during a complete combustion cycle,
at several stress concentration regions. The Rainflow cycle counting method was applied to de-
termine the stress cycles induced during functioning and a stress-life equation was then used to
estimate the fatigue lifespan of the crankshaft. Additionally, a modification to the crankshaft's
geometry was suggested and a significant reduction of the induced stresses was obtained.

1. Introduction

Internal combustion engines have a great importance in the daily life of humanity, and its manufacturers constantly seek to design
new models with greater power, lower fuel consumption and lower emissions of greenhouse effect gases. It is in the transport industry
that the use of internal combustion engines is most evident due to the need to displace people and goods. If one looks closely, an
internal combustion engine works due to a perfect working relationship between its components, each of which with a great en-
gineering complexity. One of these mechanical parts is the crankshaft, whose main function is to convert the linear movement of the
pistons into rotational movement transmitted to the shaft. The crankshaft is supported by several bearing journals and crankshaft
rotation occurs due to the torque generated by the connecting rods that connect pistons to the crankshaft at crankpin journals, which
are eccentric in relation to the longitudinal axis of the crankshaft. It is also worth mentioning the existence of lubrication channels

⁎
Corresponding author.
E-mail address: rfspm@fct.unl.pt (R.F. Martins).

https://doi.org/10.1016/j.engfailanal.2018.06.020
Received 13 February 2018; Received in revised form 24 May 2018; Accepted 20 June 2018
Available online 23 June 2018
1350-6307/ © 2018 Elsevier Ltd. All rights reserved.
J. Gomes et al. Engineering Failure Analysis 92 (2018) 466–479

that run inside the crankshaft in order to lubricate and cool the crankshaft journals and the crankpin journals.
Hence, crankshaft is an important mechanical part in all types of engines employed in applications like aircraft, reciprocating
compressors, marine engines, car engines, as well as diesel generators [1], and diesel engines used in power plants and to marine
propulsion are especially sensitive to outage events [2]. Nevertheless, a study conducted for diesel engines up to 2 MW revealed that
failures per year related to engine crankshaft are low [2], although they involve high mean stoppage time to perform corrective
maintenance. In fact, in case of a crankshaft failure, the repair cost frequently includes not only that of the crankshaft itself, but also
the cost of other parts, such as connecting rod, piston, cylinders, bearings, as well as the lengthy time period required for repair,
mainly because of the crankshaft location inside the engine [3].
An extensive literature survey of crankshaft failures [1] reported the following root causes for failure:

. Wrong machining and grinding processes during manufacturing.

. Cracks produced at stress concentration due to the combined effect of mechanical (bending and torsion) and thermal fatigue loads.
. Torsional vibrations.
. Surface contact which resulted into excess of pitting or spalling.
. Fatigue cracks may propagate through residual stress in the fillet regions.
. Improper assembly, or
. Improper lubrication.

Moreover, Yu and Xu [4] reported a crankshaft failure that occurred due to the initiation of a fatigue crack in a stress con-
centration region where the absence of a hardened layer due to nitriding was noticed. In fact, it is well known that surface hardening
by nitriding can raise the fatigue strength of the material [4]. A similar case was reported in [5], in an induction hardened cast
crankshaft, in which a crack propagated from a non-hardened fillet region and from the periphery of graphite nodules. The absence of
the hardened material and the presence of free graphite and nonspheroidal graphite in the fillet region made fatigue strength
decrease and led to failure in a shorter time than the normal usage life [5]. Nevertheless, for heavy-duty applications solid-forged
crankshafts predominate [6] instead of using cast iron. More recently, Cevik and Gurbuz [7] evaluated the effect of fillet rolling on
the fatigue behaviour of a crankshaft. They found that induced compressive residual stresses provide a significant improvement in the
fatigue strength of the component.
In addition, according to [8], there are three features almost universal regarding fatigue failure of engineering components and
structures, hence related with crankshafts, namely in-service failures are almost always high-cycle fatigue failures, crack invariably
initiates at stress-concentration and are essentially linear-elastic events, whereas any plasticity will be restricted to a very localised
region near the notch discontinuity. In Fonte et al. [9], a crack was found at the crankpin web-fillet and fatigue was the dominant
failure mechanism; the root cause of failure was, accordingly with the authors [9], probably a poor design and a deficient assembling
of the crankshaft. In another study [10], a failure analysis of two damaged crankshafts was presented and in both cases a crack grew
from the crankpin-web fillet, being the crankpin web-fillets the critical zones where the cracks can initiate. Ktari et al. [11] carried
out a failure investigation on three cases of failed diesel engine crankshafts made up of forged carbon steel; all failures were due to
fatigue, and crack initiation occurred due to the high stress concentration on fillet radius and the unusual friction on bearing journals.
Witek et al. [12] also studied a crankshaft's failure of a diesel engine; these authors carried out a numerical analysis of the crankshaft
using the Finite Element Method and found that the main reason of premature fatigue failure was probably the large alternating
bending stresses applied to a small crankpin fillet radius (notch effect).
Another study, [13], considered crankshaft failures, either at crankshaft journal or at crankpin journal. According with [13],
crankshaft bearing failures are either caused by mechanical or tribological failure and among the later mainly due to abrasive (60%),
adhesive (19%) or surface (11%) fatigue wear. In relation to abrasive wear, this type of failure would be due to oil contamination
with wear products or with foreign particles [13]; concerning adhesive wear, that would be caused by low oil viscosity, by in-
sufficient oil supply, by oil film breakdown (overload, vibrations) or insufficient clearance (poor design, fault assembly), while
surface fatigue wear could be due to insufficient fatigue strength of lining material, or to misalignment, or due to the application of
loads that exceeds the fatigue strength of the materials [13].

2. Materials and methods

2.1. Description of the propulsion system under study

The internal combustion maritime engines are mostly diesel powered. This is because this type of engine has a much higher
compression ratio than that of gasoline engines, thus ensuring a higher torque at low engine speeds, which is ideal for overcoming the
inertia caused by the high displacement of the ships.
In the case herein studied, the propulsion of the vessels is a Combined Diesel or Gas system (CODOG), with two lines of shafts,
each with a 4-stroke supercharged diesel engine of 3600 kW of maximum power at 1200 rpm, one General Electric LM 2500 gas
turbine, a Renk Tacke gearbox and an associated Escher Wyss variable pitch propeller. The CODOG system allows to have only one
engine type delivering power to the shaft at a time, and not simultaneously.
Each diesel engine is composed by 12 cylinders arranged in “V” in the block (Fig. 1a), with a unit capacity of 11.63 dm3, and V
cylinders of the two engine's banks, namely A and B (Fig. 1a), are linked to the same crankpin journal of the crankshaft (Fig. 1b). The
two engines in operation allow a maximum speed of 19 knots, corresponding to 60% of the maximum speed of the ship, and the firing

467
J. Gomes et al. Engineering Failure Analysis 92 (2018) 466–479

Fig. 1. a) Representation of a 12 V maritime diesel engine; b) Overall view of the crankshaft under study: nomenclature and identification of main
bearing journals and crankpin bearing journal.

order in each diesel engine is A1-B1-A4-B4-A2-B2-A6-B6-A3-B3-A5-B5 (Fig. 1a), resulting in an ignition interval angle of 60° and in a
crankshaft rotation of 720° (2 rotations) during a complete cycle from A1 to B5.
Additionally, due to the large dimensions of the engine herein studied and due to the large forces involved in its normal operation,
the engine has installed a torsional vibration damper at the free end of the crankshaft in order to ensure a better performance of the
propulsion system. In fact, during operation there are periodic changes in fuel injection, as well as the generation of inertial forces,
causing torsional, axial and transverse vibrations in the crankshaft [14]. Of the three types of vibrations mentioned, the most
dangerous for the crankshaft are the torsional vibrations; hence, the need to install a vibration damper at the free end of the shaft to
dissipate the energy produced by such type of vibrations.
Concerning the crankshaft under study, it is made of an engineering steel grade type, namely AISI 4340 (275HV5), with the
chemical composition and the mechanical properties shown in Tables 1 and 2, respectively [15]. The crankshaft was initially forged
and then machined; in addition, the main bearing journals and the crankpin journals of the crankshaft were hardened by induction,
resulting in a wear resistant surface with approximately 6 mm in depth (Fig. 2) and a hardness value of 505HV5.

Table 1
Chemical composition of AISI 4340 steel determined experimentally [15].
C Si Mn P S Cr Mo Ni

0.35 0.26 0.59 0.010 0.012 1.61 0.29 1.59

468
J. Gomes et al. Engineering Failure Analysis 92 (2018) 466–479

Table 2
Mechanical properties of AISI 4340 steel grade type determined experimentally (mean value ± standard deviation) [15].
Yield Strength, σy [MPa] Tensile Strength, σR [MPa] Area Reduction [%] Ductility [%] Impact Resistance [J]

822.75 ± 5.74 977.75 ± 5.12 62.5 ± 1.29 17 ± 1.15 49.75 ± 6.5

Fig. 2. Inductive hardening of main journal. Hardening depth: approximately 6 mm.

2.2. Fault reports

Seven failures of the crankshaft under study have been reported in the last 25 years (Table 3), some of them catastrophic (three
complete failures occurred), while the rest were related with wear damage of the journals. The first failure occurred after 1200 h of
engine service, which is a rather premature time for damage [16].
Some root causes for the failures can be enumerated:

- The very frequent use of the diesel engines, higher than the use for which they were originally designed, can cause an accelerated
degradation of the components;
- The existence of an inaccurate alignment of the crankshaft in the main bearings journals, which can induce high stresses in the
shaft and result in its fracture. These misalignments can result from the wear of main bearings or eventually from the distortion of
the engine bedplate transverse members [9];
- The lower quality of the fuel and oil may contribute to increase the formation of residues in the oil and consequently in accel-
erating the process of its degradation, resulting in abrasive and/or adhesive fatigue wear of the crankshaft;
- Insufficient performance of the antifriction material of the shells' bearings;
- The existence of initial imperfections of the material and stress concentration regions, from which a fatigue crack could nucleate
and propagate.
- The very high values of torsional and bending loads in the crankshaft may be magnified by the reduction of thickness of the
crankshaft bearing shells caused by cyclic beats against the plain bearing journals. This can cause an increase of the gap between
the bearing shells and the bearing journal, allowing the exit of the oil and the occurrence of failure due to the lack of lubrication in
the remaining bearing journals;
- Surface fatigue wear could be generated due to the high stresses induced in the component by the loadings applied.

2.3. Crankshaft design

A crankshaft is a machine element that is used to transmit power and it is supported by several bearing journals. In addition, in a
crankshaft the areas exposed to highest stresses occur at either fillet transitions between the crankpin and web, as well as between the
journal and the web, or at outlets of crankpin oil bores [17], and are due to the application of alternating torsional and/or flexural
loads. Hence, the design of crankshafts is based on an evaluation of safety against fatigue in the highly stressed areas and alternating
stress should be compared with the fatigue strength of the selected crankshaft material [17]. Combining the Soderberg criterion with
Tresca criterion, a shaft may be designed, d, using Eq. 1 [18]:
1
1 3
⎡ 32n ⎡ M 2 2 2⎤
d=⎢ ⎛ t ⎞ + ⎛ Mb ⎞ ⎤ ⎥
⎢ ⎜ ⎟ ⎥ ⎜ ⎟

⎢ π ⎣ ⎝ σy ⎠ ⎝ σe ⎠ ⎦ ⎥
⎣ ⎦ (1)
where, n represents the safety factor, Mt and Mb are the torsional and the bending moment, respectively, σy is the Yield Strength and
σe is the corrected fatigue limit of the material (Eq. 2).
σf 0
σe = (K S Kfb Kt KT )
Kf (2)

469
J. Gomes et al. Engineering Failure Analysis 92 (2018) 466–479

Table 3
Failures registered in the engine under study during the last 25 years.
Year Ship Failure description Observations Run time Fractographic image or picture
(hours)

1992 Ship A Crack started at stress-concentration region at Catastrophic failure 1948

–Starboard main bearing journal no. 5 (half-span) and
motor propagated to crankpin journal no. 3 due to
fatigue. Crankpin journal no. 3 become totally
fractured. The lubrication holes remained clear.

2001 Ship A – Crack occurred at main bearing journal no. 4. Catastrophic failure. - Not available.
Starboard motor
2004 Ship A – Failure occurred in bearing journal no. 6; several Abrasive, adhesive or 15,281
Portside motor cracks propagated due to wear. surface fatigue wear.

2004 Ship B – Main bearing journal no. 6 and crankpin journal Abrasive, adhesive or 18,318
Starboard motor no. 4 showed evidence of wear and the presence of surface fatigue wear.
multiple cracks.

2004 Ship B – Portside Several shell bearings showed evidence of wear. Abrasive, adhesive or 18,400
motor surface fatigue wear.

(continued on next page)

470
J. Gomes et al. Engineering Failure Analysis 92 (2018) 466–479

Table 3 (continued)

Year Ship Failure description Observations Run time Fractographic image or picture
(hours)

2006 Ship B – Crack initiation occurred at crankpin journal no. 4 Catastrophic failure. 27,000
Starboard Motor and propagated to main bearing journal no. 6 due
to fatigue.

2008 Ship A – Failure occurred in main bearing journal no. 5; Abrasive, adhesive or 30,735
Starboard Motor several cracks propagated due to wear. surface fatigue wear.

In Eq. 2, KS is the surface finish coefficient, Kfb is the reliability coefficient, Kt is the size coefficient, KT is the temperature
coefficient and Kf represents the stress concentration factor, whereas σf0 is the non-corrected fatigue limit obtained for a mean stress
value equal to zero.
Nevertheless, the torsional moment and the bending moment may have a mean value, namely Mtm and Mbm, as well as alternating
components, Mta and Mba, respectively. In this situation, Eq. 1 generalizes in Eq. (3) [18]:

1
1 3
⎡ 32n ⎡ M 2 2 2⎤
d=⎢ ⎛ ta + Mtm ⎞ + ⎛ Mba + Mbm ⎞ ⎤ ⎥
⎢ ⎜ ⎟ ⎜ ⎟ ⎥
⎢ π ⎣ ⎝ σe σy ⎠ ⎝ σe σy ⎠ ⎥
⎦ ⎦
⎣ (3)

2.4. Forces applied in the crankshaft

The forces applied on the crankshaft will depend on the instantaneous values of pressure that will exist inside the cylinders for
each instant of the diesel cycle. The following values of pressure were considered during computational simulations (Table 4).
Additionally, knowing that the piston has a bore diameter of 230 mm and a surface area equal to 41,548 mm2, this allowed to
estimate the forces applied during a crankshaft rotation and conditions were assumed to be the same for all cylinders (Table 5).
Moreover, as the engine was considered in its maximum operating condition, an angular speed of rotation of 1200 rpm and an
effective power of 3600 kW was used, having been obtained a torque of 28.65 kN.m.

Table 4
Maximum values for pressure and force for each stroke of the engine.
Admission Compression Combustion Exhaust

Pressure [bar] Force [kN] Pressure [bar] Force [kN] Pressure [bar] Force [kN] Pressure [bar] Force [kN]
2.7 11,22 60 249.3 90 373.9 1 4.15

471
J. Gomes et al. Engineering Failure Analysis 92 (2018) 466–479

Table 5
Applied forces [N] at each iteration and in each cylinder (Colour code: yellow-admission; orange-compression; red-combustion; green-exhaust).

3. Results and its discussion

3.1. Crankpin theoretical design

To estimate the theoretical diameter of the crankpin (Fig. 3), its free-body diagram was drawn (Fig. 4a), having been considered

Fig. 3. Partial view of the crankshaft under study. Dimensions are in mm.

472
J. Gomes et al. Engineering Failure Analysis 92 (2018) 466–479

Fig. 4. a) Free-body diagram of crankpin; b) Theoretical stress concentration factor of a crankshaft under bending (adapted from [19]).

the force applied during the combustion phase (Table 5) and the unloaded area corresponding to the crank web (Fig. 3). Other values
needed in Eqs. (1–3), namely KS, Kt, KT, Kfb, n and Kf, were taken from references [18, 19] and are given in the text.

• K ≈ 0.73 (machined, HB = 262)

S

• K = 0.689 (reliability equal to 99.99%)

fb

• K = 0.75 (diameter of ≈ 160 mm)

t

• K = 1 (temperature below 100 °C)

T

• Kf ≈ 5 (notch sensitivity, q, considered equal to 1) (Fig. 4b)

• n = 2 (safety factor)
• And the fatigue limit of steel under bending (σ ) is: f0

⎧ σf 0 = 0.5σR; σR < 1400 MPa

⎨ σ = 700 ; σR > 1400 MPa
⎩ f0

Hence,
σf 0 0.5 × 978
σe = (K S Kfb Kt KT ) = 0.73 × 0.689 × 0.75 × 1 × = 36.89 MPa
Kf 5

By using the values referred above in Eq. (1), d equals:

1
1 3
⎡ 32 × 2 ⎡ 28650 2 18140 2⎤2 ⎤
d=⎢ ⎛ ⎞ +⎛ ⎞ ⎥ = 215.7 mm
π ⎢
⎣⎝ 823∙10 6
⎠ ⎝ 36.89∙106 ⎠ ⎥
⎦ ⎥
⎢
⎣ ⎦

Additionally, when dimensioning the crankpin with equation no. 3:

•M bmax = 18.14 kN∙m; Mbmin = 0.285 kN∙m (exhaust). Hence,

Mbmax − Mbmin M + Mbmin

Mba = = 8.93 kN ∙m ; Mbm = bmax = 9.21 kN ∙m
2 2

Moreover, during the operation of the motor, the torque will vary from rest to maximum value:

•M tmax = 28.7 kN∙m; Mtmin = 0 kN∙m (engine stopped)

Mtmax − Mtmin M + Mtmin

Mta = = 14.35 kN ∙m ; Mtm = tmax = 14.35 kN ∙m
2 2

Hence, from Eq. (3),

473
J. Gomes et al. Engineering Failure Analysis 92 (2018) 466–479

Fig. 5. a) Crankshaft model with loads and boundary conditions defined; b) 48 points of interest along the crankshaft where the von Mises stresses
were determined for each phase studied; c) Overall view of the finite element mesh of the crankshaft.

1
1 3
⎡ 32 × 2 ⎡ 14350 14350 2 8930 9210 2 2⎤
d=⎢ ⎛ + ⎞ +⎛ + ⎞ ⎤ ⎥ = 213.7 mm
⎢ π ⎢ ⎣ ⎝ 36.89∙10
6 6
822.7∙10 ⎠ ⎝ 36.89∙10 6 822.7∙10 ⎠ ⎥
6
⎦ ⎥
⎣ ⎦
As can be seen, if the Soderberg criterion is considered, a minimum diameter of approximately 214 mm is required for the
crankpin, which is higher than its real diameter (Figs. 3, 4). Nevertheless, despite the differences between the diameters calculated,
the Soderberg criterion is frequently considered conservative.

3.2. Finite element analyses

To obtain the variation of stresses during two rotations of the crankshaft (720°), twenty-four numerical analysis were simulated
with intervals of 30° (Table 5). Therefore, the forces listed in Table 5 were applied to the crankshaft, as seen in Fig. 5a (pink arrows),
and the boundary conditions, shown by green arrows (Fig. 5a), were defined as hinges, allowing the rotation of the crankshaft.
Moreover, since the engine has two cylinder's banks, namely A and B, there were 2 regions for each journal, either main bearing
journal or crankpin journal, where it was important to determine the induced stresses (Fig. 5b). So, in each analysis, 48 points of
interest were surveyed (Fig. 5b) and the Von Mises stresses determined; this allowed to determine the stress distribution in each point
(out of 48) during the rotation of the crankshaft. The finite element mesh was composed by tetrahedral elements, with an edge's
length equal to 10 mm, as shown in Fig. 5c.
Due to the very high number of results generated, only the worst case is shown in Fig. 6a and b, which was obtained for
crankshafts' phase angle equal to 60° and at main bearing journal no. 5. As it can be seen, a maximum stress value equal to 457 MPa
(Fig. 6a, b) was reached for critical point A4.

474
J. Gomes et al. Engineering Failure Analysis 92 (2018) 466–479

Fig. 6. a) Instantaneous Von Mises stress distribution [MPa] induced in the crankshaft at several critical points; b) Von-Mises stresses [MPa] at
points A3, B3, A4 and B4, at bearing journal no.5, during a crankshaft rotation of 720°.

475
J. Gomes et al. Engineering Failure Analysis 92 (2018) 466–479

Table 6
Rainflow cycle counting, fatigue life and damage accumulated during 2 rotations of the crankshaft (720°).
n (cycle) σm [MPa] σa [MPa] Nf [cycles] n/Nf

1 115,21 33,08 1,14E + 17 8,74E-18

0,5 248,325 194,385 6,46E + 08 7,73E-10
1 180,01 136,24 3,86E + 10 2,58E-11
1 92,265 48,595 2,57E + 15 3,88E-16
0,5 229,09 213,62 2,80E + 08 1,78E-09
0,5 125,9 110,43 4,68E + 11 1,06E-12
0,5 161,375 74,955 1,97E + 13 2,53E-14
0,5 109,755 23,335 4,21E + 18 1,18E-19
0,5 133,17 0,08 6,16E + 43 8,10E-45
Total 2,58E-09

To assess the number of cycles in one complete cycle of functioning of the 4-stroke engine (720°), the Rainflow cycle counting
method [20] was applied to the stress distribution given in Fig. 6b. The results are shown in Table 6; then, a stress-life Eq. (4) that
applies for nonzero mean stress was used in combination with the Palmgren-Miner Rule to estimate the fatigue lifespan of the
crankshaft if considering a total damage, D, equal to 0.1, 0.5 or 1 (Table 7).

1
b
1⎛ σa ⎞
Nf = ⎜ ⎟
2 ⎝ σ ′ f − σm ⎠ (4)

where, Nf is the number of cycles to failure, σa is the stress amplitude, σm is the mean stress and σ'f and b are material constants equal
to 1758 MPa and − 0.0977, respectively [21].
The time results presented in Table 7 were calculated assuming the engine functioning at the maximum speed of rotation, namely
1200 rpm, and the minimum speed of rotation (700 rpm); from results, it takes about 1000 h to accomplish an accumulated damage
equal to 0.1 or about 11,000 h to have an accumulated damage equal to 1, if the speed of rotation of the crankshaft is 1200 rpm.
Nevertheless, depending on the service regimes of the diesel engines, if one considers the minimum speed of rotation of the engine,
namely 700 rpm (Table 7), then it will be necessary about 9000 h to achieve an accumulated damage equal to 0.5, or about 18,000 h
to accomplish an accumulated damage equal to 1.
Considering the catastrophic failures of the crankshaft under study (Table 3), it was possible to verify that they were registered in
between 1948 h and 30,000 h, which are in the same order of magnitude of the time to failure given in Table 7. Nevertheless, for the
shorter time to failure registered (1948 h in 1992 according with Table 3), the stresses obtained in Fig. 6b seems to be lower than it
should; hence, for this case, other possibilities could be considered, such as the misalignment of the crankshaft, or the presence of
initial flaws that originated a fast crack propagation, or even higher stresses applied, or lubrication problems.

3.3. The influence of initial imperfections of the material

After the failure occurred in 1992 (Table 3), which happened after only 1948 h of engine run time, a fractographic analysis of the
crankshaft's fracture surface was carried out. The fracture could be seen as a fatigue crack propagation with origin at the fillet radius
of main journal number 5 (Fig. 7a) that propagated towards the crank web and crank journal no. 3 due to bending stresses; in
addition, a secondary fatigue crack nucleated and propagated at the transition between the fillet radius and the journal surface
(Fig. 7a). By using a scanning electron microscope (SEM), it was possible to infer that the initial crack begun 300 μm beneath fillet
surface at a non-metallic inclusion with a maximum length of 180 μm (Fig. 7b, c). An energy dispersive X-ray analysis revealed that
the inclusion was made of silicates and aluminates. Moreover, semi-elliptical beach marks are present in fracture surface and crack
growth was due to fatigue.
Hence, the combination of high bending and torsional stresses with the presence of an initial imperfection in the material – stress
raiser - led to premature failure of crankshaft; this occurrence shown the importance of guarantee a free-defect condition of the
component.

Table 7
Fatigue life time of crankshaft for accumulated damage equal to 0.1, 0.5 and 1. Number of hours calculated assuming an angular speed equal to
1200 rpm or 700 rpm.
Damage, D Repetitions Rotations Time [hours] Speed of rotation: 1200 rpm Time [hours] Speed of rotation: 700 rpm

0.1 3.87E + 07 7.74E + 07 1075 1842.8

0.5 1.93E + 08 3.87E + 08 5375 9214.3
1 3.87E + 08 7.74E + 08 10,750 18,428.6

476
J. Gomes et al. Engineering Failure Analysis 92 (2018) 466–479

Fig. 7. a) – Overall view of fracture surface (←Initial crack; < − − Secondary crack) B-bending, BS – Secondary bending; b) SEM image of crack
initiation site (BB > > > > > > Origin of fracture); c) Magnification of non-metallic inclusion.

3.4. Redesigning the crankshaft

Having in mind the catastrophic failures that have been registered in the crankshaft under study, a design modification was tested
in order to try to diminish the induced stresses in this mechanical component, and, hence, to increase its fatigue life time.
As was seen in section 3.1, the stress concentration factor associated with the geometry of the crankshaft, Kf, was approximately 5;
by changing the crank web width, t, from 60.5 mm to 80 mm (Fig. 8), a lower stress concentration factor can be calculated (Fig. 8).
That change was also noticed in the stresses induced in the crankshaft, which maximum value reduced from 457 MPa to 283 MPa
(−38%) (Fig. 9).
Nevertheless, increasing the crankweb width will increase the length of the crankshaft, the overall dimensions of the motor and
the mass of the component; hence, an increase in the fuel consumption will be expected, as well as a lower maximum velocity
attained by the ship. In addition, inertial forces, crankshaft's balance and frequencies of vibration of the crankshaft will change;
hence, special attention should also be addressed to the performance of the torsional vibration damper in order to avoid crack
nucleation and propagation induced by vibration. However, lower stress concentration will be expected if crank web width is in-
creased to 80 mm as well as improved fatigue performance.

4. Conclusion

Seven crankshaft's failures in maritime diesel engines under study have been reported during the last 25 years, three of them of

Fig. 8. Modification of the crank web width, t, from 60.5 mm to 80 mm: the stress concentration factor, Kf, reduced from 5 to approximately 3.5.

477
J. Gomes et al. Engineering Failure Analysis 92 (2018) 466–479

Fig. 9. –Stress reduction due to the modification of the crank web width (crankshaft at 60°).

catastrophic type, and several causes of failure were listed, namely:

- The very frequent use of the diesel engines, higher than the use for which they were originally designed, can cause an accelerated
degradation of the components;
- The existence of an inaccurate alignment of the crankshaft in the main bearings journals, which can induce high stresses in the
shaft and result in its fracture. These misalignments can result from the wear of main bearings or eventually from the distortion of
the engine bedplate transverse members [9];
- The lower quality of the fuel and oil may contribute to increase the formation of residues in the oil and consequently in accel-
erating the process of its degradation, resulting in abrasive and/or adhesive fatigue wear of the crankshaft;
- Insufficient performance of the antifriction material of the shells' bearings;
- The existence of initial imperfections of the material and stress concentration regions, from which a fatigue crack could nucleate
and propagate.
- Surface fatigue wear could be generated due to the high stresses induced in the component by the loadings applied.

From those causes, the article focused on the calculation of stresses induced in the crankshaft and on its fatigue life prediction.
Moreover, the influence of initial imperfections of the material was discussed, and a new design for the crankshaft was proposed.
Therefore, in order to carry out the finite element analysis of the crankshaft, it was necessary to estimate the forces actuating in
the component during its normal operation along one cycle of functioning. The finite element model of the crankshaft has predicted
that the most stressed regions (457 MPa) matched the fractured zone, namely main journal fillets no. 4, 5 and 6. In addition,
considering the catastrophic failures of the crankshaft under study, it was possible to verify that they were registered in between
1948 h and 30,000 h, which were in the same order of magnitude of the life time predicted during the fatigue life prediction analysis
carried out during present research.
Additionally, a design modification of the crank web width was assessed in order to lower critical stresses induced on the
crankshaft and maximum equivalent stress changed from 457 MPa to 283 MPa. Nevertheless, increasing the crankweb width will
increase the length of the crankshaft, the overall dimensions of the motor and the mass of the component; hence, an increase in the
fuel consumption will be expected, as well as a lower maximum velocity attained by the ship. In addition, inertial forces, crankshaft's
balance and frequencies of vibration of the crankshaft will change; hence, special attention should also be addressed to the per-
formance of the torsional vibration damper in order to avoid crack nucleation and propagation induced by vibration. However, lower
stress concentration will be expected if crank web width is increased to 80 mm as well as improved fatigue performance.
Finally, special attention must be given during the fabrication of the crankshaft in order to avoid initial imperfections of the
material and premature failure of the component. Moreover, the application of compressive residual stresses to crankpin fillets and
main journal fillets by rolling or shot peening should also be considered in order to lower the net stresses applied at critical regions of
the crankshaft and to increase its fatigue strength.

Acknowledgements

The authors would like to thank the Portuguese Foundation for Science and Technology through project ref. UID/EMS/00667/
2013.

478
J. Gomes et al. Engineering Failure Analysis 92 (2018) 466–479

References

[1] A. Patil, G. Datar, A. Kolhe, Crankshaft failure due to fatigue — a review, Int. J. Mech. Eng. & Rob. Res. 3 (1) (2014) 166–172.
[2] J.A. Becerra Villanueva, F. Jimenez Espadafor, F. Cruz-Peragon, M. Torres García, A methodology for cracks identification in large crankshafts, Mech. Syst.
Signal Process. 25 (2011) 3168–3185.
[3] F. Jiménez Espadafor, J. Becerra Villanueva, M. Torres García, Analysis of a diesel generator crankshaft failure, Eng. Fail. Anal. 16 (2009) 2333–2341.
[4] Zhiwei Yu, Xiaolei Xu, Failure analysis of a diesel engine crankshaft, Eng. Fail. Anal. 12 (2005) 487–495.
[5] Osman Asi, Failure analysis of a crankshaft made from ductile cast iron, Eng. Fail. Anal. 13 (2006) 1260–1267.
[6] H. Bayrakceken, S. Tasgetiren, F. Aksoy, Failures of single cylinder diesel engines crank shafts, Eng. Fail. Anal. 14 (2007) 725–730.
[7] G. Cevik, R. Gurbuz, Evaluation of fatigue performance of a fillet rolled diesel engine crankshaft, Eng. Fail. Anal. 27 (2013) 250–261.
[8] D. Taylor, A.J. Ciepalowicz, P. Rogers, J. Devlukia, Prediction of fatigue failure in a crankshaft using the technique of crack modelling, fatigue Fract, Eng. Mater.
Struct. 20 (1) (1997) 13–21.
[9] M. Fonte, V. Infante, L. Reis, M. Freitas, Failure mode analysis of a diesel motor crankshaft, Eng. Fail. Anal. 82 (2017) 681–686.
[10] M. Fonte, V. Infante, M. Freitas, L. Reis, Failure mode analysis of two diesel engine crankshafts, Procedia Struct. Integ. 1 (2016) 313–318.
[11] A. Ktari, N. Haddar, H.F. Ayedi, Fatigue fracture expertise of train engine crankshafts, Eng. Fail. Anal. 18 (2011) 1085–1093.
[12] Lucjan Witek, Michał Sikora, Feliks Stachowicz, Tomasz Trzepiecinski, Stress and failure analysis of the crankshaft of diesel engine, Eng. Fail. Anal. 82 (2017)
703–712.
[13] Aleksandar Vencl, Aleksandar Rac, Diesel engine crankshaft journal bearings failures: case study, Eng. Fail. Anal. 44 (2014) 217–228.
[14] Wojciech Homik, Diagnostics, maintenance and regeneration of torsional vibration dampers for crankshafts of ship diesel engines, Pol. Marit. Res. vol 17, (1(64))
(2010) 62–68.
[15] N. Gaivota, Structural Integrity Assessment of a Maritime Engine's Crankshaft and Prevention of Failure, Master Thesis (in Portuguese), Escola Naval, Base Naval
de Lisboa, Lisbon, 2015.
[16] C.M. Branco, V. Infante, A. Sousae Brito, R.F. Martins, A failure analysis study of wet liners in maritime diesel engines, Eng. Fail. Anal. 9 (2002) 403–421.
[17] The International Association of Classification Societies - IACS, Calculation of Crankshafts for i.c Engines, Document UR M53 Rev2 CLN, pp. 3, London, (2011).
[18] C.M. Branco, J.M. Ferreira, J.D. Costa, A.S. Ribeiro, Machine Elements Design (in Portuguese), Fundação Calouste Gulbenkian, Lisbon, 2005.
[19] W. Pilkey, Peterson's Stress Concentration Factors, 2ª ed., John Wiley & Sons, Inc., New York, 1997.
[20] ASTM E 1049–85, Standard Practices for Cycle Counting in Fatigue Analysis, American Society for Testing and Materials, West Conshohocken, 1997.
[21] N.E. Dowling, Mechanical Behavior of Materials, 4th edition, Ed. Pearson, pp. 424, Virginia Polytechnic Institute & State University, (2013).

479