Lloyds Register EMEA

Technical Investigation Department

Failures of Medium Speed Diesel Engines and Machinery


Peter Filcek, Technical Manager, TID






































Paper 12
Failures of Medium Speed Diesel Engines
and Machinery
Peter Filcek
175

FAILURES OF MEDIUM SPEED DIESEL ENGINES AND MACHINERY

Peter Filcek

SUMMARY

The present Technical Investigations and Analysis (TIA) section of Lloyds Registers Marine Consultancy
Services grew out of the previous Technical Investigation Department (TID) formed in 1947 by the then Chief
Engineer Surveyor, Dr. S. F. Dorey. Its purpose was stated to be to give Lloyds Register a capability to
explore marine failures and to research technical problems with a view to improving the Rules. This
remains true today, over 60 years later. One feature of the section, which is perhaps unique, is the mixture of
mechanical and marine engineers, hydrodynamicists and fluid dynamicists, civil and structural engineers
and instrument specialists who combine together to provide a professional and rapid response consultancy
service to Lloyds Register, to the marine industry and to the other industries which Lloyds Register serves.
The basic tenet is to combine measurement and engineering knowledge: when you can measure what you
are speaking about, and express it in numbers, you know something about it (Lord Kelvin). The objective of
this present paper is to share the knowledge of service experience derived from the unique standpoint of a
classification society which will be of use in improving the design, materials and manufacture and
prevention of the mal-operation of machinery.


1. INTRODUCTION

Lloyds Registers Technical Investigation engineers
analyse a wide range of failures, both marine and
land based. A considerable expertise exists in
investigating problems or failures related to
propellers, sternbushes and other bearing, shafting
systems, gearing, boilers, turbines, diesel engines,
hull structures, fluid flow and vibration. At least
one paper could be written on any of these subjects
and the problem has been what to include and what
to leave out. Medium speed diesel engines are the
most common power source in the marine industry
and it is hoped that a description of some recent
investigations of these engines and associated
machinery would be of interest to the widest
possible body of people.

The cases include mechanical overload, fatigue
fractures, crack initiation by rubbing, bearing
failures and operational problems. The failure
mechanisms are well understood but it is the
combination of design features, manufacturing
problems and operating conditions which can have
unexpected results.

2. VESSEL 1 MAIN ENGINE
CONNECTING ROD BOTTOM END

This failure highlights a conflict between what was
revealed in a technical investigation of a failure and
what was reported to have been carried out during
the preceding repair.

The failure occurred to an 8-cylinder in-line main
engine with a maximum continuous service rating
of 5,280 kW at 510 rpm. The main engine had
accumulated some 12,000 operating hours since the
previous overhaul and rather more in total. The
connecting rods have the normal marine bottom
end to assist maintenance, Figure1. This design
allows the piston to be withdrawn through the
cylinder liner whilst a robust bottom end bearing
housing with a large bearing surface can be used.

The engine was being test run on completion of
work carried out by shore-side fitters to repair
fretting damage. Fretting of mating surfaces is a
characteristic problem of diesel engines. In this case
it had occurred between the lower face of the top
flange of the cylinder liners and the recessed top
land in the engine block. Following completion of
the repairs, which had necessitated removal of the
connecting rods, the engine was being prepared to
go back into service. The engine had been running
for eight minutes when firstly, valve tapping noises
were heard, followed by loud thumping noises.
The engine was immediately shut down. No.8
connecting rod and cylinder liner had failed
catastrophically.

The palm joint between the connecting rod foot and
the big end was found to be completely separated,
the lower half of the cylinder liner broken away
along with part of the cylinder block, the big end
had started to clamshell open and the balance
weight had come off completely.





177


Most of the broken parts, including the four broken
palm joint studs and nuts, were recovered from the
vessel and the jigsaw of parts from the unit was re-
assembled in the Lloyds Register Materials and
NDE Laboratory in Croydon.

The results of the initial examination were
unexpected. The two starboard/outboard studs
were found with the nuts at the top of the threads.
Both studs had damage to the threads where the
nuts should have been sitting in their normal
tightened position (Figure 1). Impact damage was
also observed on the seats of the nuts. Hammering
to the seats of the two nuts on the port/inboard side
suggested these also had not been tight before the
engine was started.



Figure 1: The failed connecting rod bottom end. The nuts
at the top of studs may be seen on the left.

All four studs securing the palm joint had failed by
tensile overload with the characteristic cup and
cone feature. This again was a little unexpected, as
fatigue is often found to be the principal failure
mode of rotating components. The material was not
at fault, as the metallurgical examination confirmed
that the studs conformed to the Engine builders
34CrNiMo6 steel material specification. Further
examination of the parts indicated that there had
been no loss of lubrication, or cooling, nor a seizure
of the piston which might have contributed to the
failure.

The procedure for tightening and releasing the
palm joint studs was reviewed. The procedure was
witnessed on board the vessel and the equipment
found to be satisfactory. Checking the stretch of the
studs is an integral part of tightening the studs, and
if followed correctly, the stud will be tight and there
will be a stretch measurement to go in the overhaul
report.

A discrepancy therefore existed between the results
of the investigation carried out on board and in the
laboratory, and the overhaul reports which
indicated that the palm joint studs had been
measured for stretch. The accident might have been
averted if a final visual inspection of the running
gear before the engine was closed up for the test
run had detected the slack nuts.

3. VESSEL 2 MAIN ENGINE
CONNECTING ROD BOTTOM END

The question was recently asked: Do diesel engines
have an indefinite service life? The answer is
probably yes, provided that components are
maintained within the engine builders tolerances
and provided that fretting of mating surfaces can be
kept under control as referred to in the previous
failure study. The following short example concerns
bearing housings that had worn out of tolerance.

As the vessel was berthing one of the main engines
shut down as a result of a high oil mist being
detected in the crankcase. There were no obvious
reasons for the shut down and after rectifying
several faults with the crankcase oil mist detector,
the engine was re-started. It immediately shut
down again as the crankcase oil mist shut-down
alarm was activated. The crankcase was then
opened up for inspection. This revealed that the
number 3 bottom end crankpin bearing had failed.
The bearing shells had overheated, seized and
rotated within the marine head bottom end bearing
housing.

Technical Investigations were asked by the
managers to assist in the investigation of the failure
of the bearings.



Figure 2: No.2 bottom end lower bearing shell. No.3 was
too badly damaged to provide any evidence. Fatigue of
the overlay is evident.

178

Number 3 crankpin was found heavily scored,
despite having been already ground out to a depth
of 5.8 mm. All bottom end and numbers 3 and 4
main bearings adjacent to the failed bottom end
were opened up for inspection. The bearing shells,
journals and pins were satisfactory except for the
number 2 bottom end lower shell, as seen in Figure
2, and the failed number 3 bottom end bearings
shells. The latter were too badly wiped and
overheated to provide much information. The
number 2 bottom end lower half shell provided
more information. The running surface had
suffered loss of the white metal overlay by fatigue
(Figure 2) while the back of the shell showed
evidence of fretting caused by poor contact in the
housing (Figure 3). The condition of number 2
bottom end was considered to provide a possible
indication of the condition of the failed number 3
bottom end and crankpin.



Figure 3: Number 2 bottom end upper bearing shell
showing fretting as a result of poor support due to
distortion of the housing. The back of lower half was
similar.

All the bottom end housings and bearings were re-
assembled and the bores were checked for
roundness. This revealed that distortion of the
intended circular geometry had occurred and that
the bores were elliptical. The internal diameters, in
line with the longitudinal axis of the connecting rod,
were undersize by 0.01-0.04 mm from the design
tolerances while the transverse diameters across the
butt of the shells were oversize by between 0.01 and
0.16 mm from the design tolerances.

The failure to maintain the correct housing
geometry is considered to have adversely affected
the support to the bearing shell. This then failed by
loss of the white metal due to fatigue and
consequent disruption of the hydrodynamic oil film.
Bearing shells should be inspected both on the front
running surfaces and on the back. Markings and
deposits on the housing side will provide
significant information about the wear of the
bearing and whether or not remedial action is
required. Markings on the running side can often be
associated with damage which has started in the
interface between the shell and housing.

The damaged crankpin was below Rule minimum
diameter and the crankshaft was replaced. A
recommendation was made to renew the marine
bottom ends and to check the diameters and
alignment of the main bearing housings.

4. VESSEL 3 AUXILIARY ENGINE
CONNECTING ROD BOLTS

This failed part had accumulated many running
hours and failed in an unusual location. The likely
stress raiser or crack initiation site may have been
obscured by the difficulties of taking the
components apart.

This failure concerned an auxiliary engine. The
engine is a four-stroke, eight-cylinder, in-line,
turbocharged with a bore diameter of 250 mm and
stroke of 300 mm with a rated output power of
1,589 kW at 1000 rpm. The connecting rod has a
small end with a plain cylindrical sleeve bearing for
the gudgeon pin and an integral large end assembly
with an angled bearing cap. The large end has thin
wall, steel backed, plain bearing shells held in place
by the clamping force from an angled bearing cap
secured by two studs and hydraulically tightened
nuts.

The vessel was underway at sea and with the
engine on load at about 800 kW when the failure
occurred. The engine was shut down manually
from the switchboard.

The number 4 connecting rod was found to have
broken across the base of the shank close to the
large end (Figure 4).

The engine had operated for about 1,800 hours since
the last overhaul. This did not include replacement
of the bottom end bearing shells or the bottom end
studs. The connecting rods had been replaced some
23,500 hours previously.

Technical Investigations were asked to assist and a
metallurgical examination was carried out by
Lloyd's Registers MateriaIs and NDE Laboratory.
Material tests confirmed that the connecting rod



179

material was a 34CrNiMo6 low alloy steel in a heat
treated condition in accordance with the Engine
builders specification. Neither the large end, nor
the small end bearings, showed signs of distress,
movement or overheating. The wear of the bearing
shells was consistent with the operating hours.



Figure 4: Connecting rod failed across the shank adjacent
to bottom end.

The mode of failure was observed to be fatigue. The
fracture surface exhibited beach marks typical of
fatigue. These had propagated from one main
initiation point and three adjacent secondary
initiation sites in the same thread root of the hole
for the short (upper) stud for the large end bearing
cap (Figure 5). The site of the fracture was between
seven and eight threads down the hole. Some
damage was found on the female threads in the
hole and on the male threads on the stud.

Fatigue cracks propagate under repetitive
alternating tensile loading. For a connecting rod
bottom end joint in a four-stroke cycle diesel engine,
the loads are expected to reach a maximum once
every two revolutions towards the end of the
exhaust stroke and during the induction stoke. It
was estimated that the rod had undergone some 7 x
10
8
tensile load cycles since replacement and about
5.5 x 10
7
tensile load cycles since overhaul. The rod
ought therefore to have survived.



Figure 5: Fracture origin at thread root towards bottom
of the tapped hole, 7-8 threads below the surface.

Additional factors were therefore looked for which
would have increased the local stress in the thread
root above the infinite endurance fatigue strength.
These were considered to be particular to this stud
in this connecting rod, as no damage was found to
the threads of No.2 connecting rod which was also
examined. No other failures had been reported. No
evidence was found to suggest that there had been
an increase in tensile stresses in the connecting rod
or studs such as might be associated with a piston
seizure or an interruption in lubrication.

The investigation proceeded by examining possible
contributory factors to arrive at a most probable
cause.

Close examination revealed that the thread root at
the crack initiation sites was damaged by gouging
due to interference from the thread of the stud. This
damage could provide the additional stress
concentration to raise the stress locally in what is
already a highly stressed component, or
alternatively it may have provided a pre-formed




180

initiation site from which a crack could propagate.
This would have significantly lowered the fatigue
life as more time is spent in crack initiation than
propagation.

Unfortunately, the metallurgical examination was
unable to conclusively determine when the gouging
occurred. It could not be precluded that it had been
done during removal of the damaged stud for
examination. Thread binding or interference, which
might have caused gouging, was not found on the
studs fitted to the undamaged No.2 and No.3
connecting rods submitted for comparison. No
inclusion or other discontinuity in the parent
material was found at the initiation point.

Consideration was given to whether or not the stud
and nut assembly could have been over-tightened
by the hand operated hydraulic stud tensioning
equipment. The threads were not plastically
deformed and for a stud in a threaded hole the load
is carried predominantly by the three to four
threads immediately below the surface. As the
initiation point was close to the bottom of the hole,
it is unlikely that the failure was caused by over-
tightening, or conversely by lack of tightening, of
the large end cap nuts.

A worn bottoming tap might have been used to cut
the threads in the hole causing the crest of the stud
thread to bind in the root of the female thread, or a
load bias in the area of the initiation point. The
major diameters of all these threads were measured
and the variation was found to be 0.14 mm (across
six studs). The examination showed that the bottom
threads (towards the free end) of the studs from the
additional connecting rods were polished. All of
these studs, as well as the long leg (lower) stud of
the failed connecting rod, were easily screwed by
hand in and out of their holes. As all threads were
likely to have been machined in a CNC machine
centre, it is unlikely that worn tools were a cause.

The rod certainly failed by fatigue, but identifying
the precise cause was not possible. It is very
tempting to say that gouging of the thread
introduced a crack initiation site, but the material
evidence did not allow such a positive conclusion to
be made.

5. VESSEL 4 AUXILIARY ENGINE
BALANCE WEIGHT

Engine builders issue service letters to inform
owners and managers of experience and possible
problems with their engines. Lloyds Register issues
Noteworthy Survey Item (NSI) letters to Surveyors
for the same purpose. In this particular case, letters
had been issued in 1995 and 1996 concerning the
problem reported here. It was unexpected to have
an apparent reoccurrence of the problem after so
many years.

The problem occurred with a balance weight on an
auxiliary engine. This is an eight-cylinder in-line
four-stroke diesel engine rated to deliver 2,830 kW
at 720 rpm. At the time of the failure, the engine
had accumulated over 90,000 operating hours.

The arrangement of balance weights on each
crankthrow is not symmetrical on this engine.
Each crankthrow has a single balance weight per
cylinder. Using the engine builders numbering
convention, where the number 1 cylinder is at the
flywheel end of the engine, crankthrows 1 (drive
end) to 4 have a single balance weight forward of
the respective connecting rod, while throws 5 to 8
have the balance weights aft of the connecting rod.
The balance weight is secured to the crankweb by
two hydraulically tensioned bolts.



Figure 6: Broken balance weight in two pieces on engine
room floor plates.

The vessel had just left port when the auxiliary
diesel generator set came to an abrupt halt as the
balance weight from the number 8 cylinder
crankweb became detached and was discharged
through the side of the engine (Figure 6). Luckily,
there were no engine room staff in the vicinity.

The balance weight had broken into two parts. The
mode of failure was fatigue as was clearly
demonstrated by the characteristic beach-marks.



181

These had propagated radially outwards from a
fillet radius in the transition between the camshaft
side abutment face and the horn (Figure 7). The
crack originated at one of several micro-cracks due
to thermal stress found to be in a heat affected zone
resulting from the flame cutting process used in the
original balance weight manufacturing process.

A close-up examination in the Lloyds Register
Materials laboratory showed that oil was present in
the region of the crack adjacent to the initiation
zone. This suggested that the crack had existed for
some time. If this was the case, what had caused it
to propagate, especially when the engine was well
past the 40,000 hours at which the engine builders
service letter considered further failures
improbable?



Figure 7: View on fracture face in the athwartships
direction, showing depth of dirt in existing crack, about 2
mm. The origin is in the inside fillet radius of the balance
weight.

A finite element model of the balance weight was
made. The combined stresses in the fillet radii due
to centrifugal loading at the normal operating speed
of 720 rpm were predicted to be approximately 114
N/mm
2
(Figure 8). This is significantly less than the
estimated maximum fatigue strength 200 N/mm
2

for infinite endurance.

The thermal micro-cracks were in a stress
concentration zone and would have provided a
ready made site from which cracks could propagate.
The puzzle was why failure had not occurred
earlier. This problem was addressed using fracture
mechanics.

In linear elastic fracture mechanics, the fatigue
crack growth for engineering materials can be
characterised by the Paris Law. The law contains
the concept of critical crack size and the threshold
stress intensity factor,
th
. It postulates that there
is a stress intensity field and crack size in which a
crack can exist without propagating. It is used
widely to safely manage crack growth in many
industries. The critical crack size for the balance
weights was estimated to be 2 mm. This was a
similar magnitude to the penetration of the dirt and
oil found in the original crack.



Figure 8: Finite element model stress contour map of the
balance weight under normal loading at 720 rpm.

The answer to why the balance weight had failed
after such a long period in service was found after
further discussion with the ships staff and
examination of the log-books. The engine had
suffered damage two months previously (at 400
running hours). It is thought that a broken valve
head had perforated the cylinder head flame deck
and piston crown, releasing water and oil, and
suddenly stopping the engine. It is postulated that
this sudden shock caused the existing crack to
extend beyond the critical length, after which
propagation to failure was a matter of time.

An inspection procedure had been carried out as
advised by the engine builder. This relied on the
cracks extending to the sides of the balance weights
where they could be seen unaided by eye using a
mirror. The crack origin, in this case, was towards
the middle of the underside of the balance weight
and could not have been seen by such a method,
especially on the rough surface, without the
assistance of a non-destructive examination
technique such as fluorescent magnetic particle
(preferred) or dye-penetrant.

Two further balance weights were found to have
cracks in them during the repair of the engine.




182

As a precaution, all the original balance weights
were replaced. The replacement balance weights
were manufactured differently to eliminate the
possibility of crack initiation in the underside fillet.

Service letters are not always acted upon. Technical
Investigations carried out an investigation into the
damage of a connecting rod which had failed by
fatigue initiated by fretting between the top end
housing and bush. It came to light that a service
letter had been issued by the engine builder, but
that nobody had acted upon it as the engines had
been in long-term storage at the time.

6. VESSEL 5 - GEARBOX

The need for vigilance and cleanliness in engine
and gearbox repairs cannot be emphasised enough.
This incident concerns an accident to a twin-input,
single-output double reduction main propulsion
gearbox.

The gearbox had been opened up for inspection
following an earlier incident in which the propeller
and shafting had been damaged. During the post
repair trials the main engines were clutched-in and
turning the gearbox at idling speed. As speed was
increased towards service 520 revolutions per
minute, a loud noise was heard from the gearbox.
The engines were quickly shut down.

Technical Investigations attended on board and
identified foreign detritus in the gearbox. This was
retrieved and examined by Lloyds Register
Materials and NDE Laboratory. The material was
identified as 31CrV3 chrome vanadium alloy steel
with a flash chromium overlay plating, as typically
used in hand tools. Perhaps the most convincing
evidence was the photograph taken on board of a
similar 13 mm combination spanner, [9].



Figure 9: Comparison of detritus found in gearbox with
an intact 13 mm combination spanner.
This is not a gearbox tool. The gearbox had been
open for some time and had been fitted with
temporary covers. In all likelihood, the spanner
could have been dropped from other work areas
adjacent to the gearbox during the repair period.

The object had initially entered the mesh at the
outboard pinion, and at least some part had
transferred via the main wheel to the inboard
pinion. Four separate areas of damage on the main
wheel indicated that some detritus had remained
in-mesh for more than one revolution of the pinions.

Local hand dressing of the damaged areas of the
gears was undertaken under the supervision of the
manufacturer after the incident. Technical
Investigations concurred with this proposed course
of action and considered that the gears could
remain in service. However, the resulting torque
restriction due to the diminution of effective face
width would have resulted in a reduction in the
vessels speed and an inability to maintain
schedules. This was unacceptable to the owners and
a replacement set of gear elements was supplied.

7. VESSELS 6 THREE SIMILAR
FAILURES OF AUXILIARY ENGINE
CONNECTING RODS

Three separate but similar failures on the same type
of engine occurred within a short time of each other
on three newbuilds. The engines are eight-cylinder
in-line types built under license. The engines have a
maximum continuous rating of 3,500 kW at 720 rpm.
The part in common to all the failures is the
connecting rod.

The running hours at the time of the engine failures
were approximately 300, 1,600 and 1,800.

The connecting rod assembly is made up of a
marine head, which consists of top and bottom half
forgings to support a split, shell type white metal
big end bearing, and bolted through a palm
arrangement to the connecting rod shank. Each of
the failed connecting rod assemblies had come from
the same forge.

Technical Investigations inspected the marine heads
from the first and last of the ships at the engine
builders works and witnessed material tests on
samples cut from the damaged components. Two
additional marine heads were subsequently
selected for metallurgical examination in Lloyds
Registers Material Laboratory.





183





Figure 10: Shoulder of upper half of marine bottom end
adjacent to stud hole. The arrow shows the location of the
forging inclusions. Rough machining can be seen inside
the hole.



Figure 11: Magnetic particle indications of forging
inclusions in the area indicated by the arrow in Figure
10.

The cause of failure was the presence of laps in the
marine heads as seen in Figures 10 and 11. The
mode of failure was fatigue, originating at one or
more forging laps on the cap half shoulder radius,
external surface. Careful examination revealed
other concerns about the quality of the machined
marine heads. These included: poor surface finish,
surface flaws, local decarburisation, excessive
grinding, and poor machining of threads and run-
out bores.

During the course of the investigation it was
discovered that the forge had been removed from
the list of approved suppliers by the licensor.

Recommendations were made to replace all marine
heads from that forge as soon as reasonably
practicable and, in the meantime, to restrict access
to the areas around running auxiliary engines.




184

An unusual aspect of the investigation was the
request to examine the certification for the
components as it was perceived that there may
have been problems. Class societies had accepted
certificates from other societies in order to facilitate
the delivery of the engines with components from
different countries. This is normal practice within
defined acceptance criteria; for example, material
specification requirements and approval of the
manufacturer. This added to the complexity of
tracing responsibility for components certificates
and for their manufacture. Some certificate
anomalies and shortcomings were identified and
made known.

8. VESSEL 7 MAIN/AUXILIARY ENGINE
CYLINDER HEAD DAMAGE

Two problems were identified on an engine which
formed part of the ships diesel-electric propulsion
system. The engine has 12 cylinders of 380 mm bore
by 475 mm stroke arranged in a vee-shape
developing 8700 kW at 600 rpm. The engine had
operated for about 10,200 hours at the time of
failure.

The first incident, which had initially been
investigated by the engine builder, concerned
damage to number 4 cylinder head in A-bank
(Figure 12) and damage to the A-bank turbocharger.
During the subsequent repair, cracks were found in
the underside flame deck of cylinder head number
1 on B-bank. The ships managers were concerned
that the two incidents were related and asked
Technical Investigations to investigate the matter.



Figure 12: Fluorescent magnetic particle indications of
cracks in fame face of cylinder head A4 damaged by the
broken valve.



Textile debris had been found in the lubricating oil
supply to the inlet valve bridge. It is likely that this
had been accidentally left there during an overhaul
1,200 hours before the failure. The spindle of valve
B had broken due to a fatigue crack which
propagated from a site on the inside of a very slight
bend. The maximum total indicated runout of 0.29
mm was measured adjacent to the fracture. This
compares with an allowable out-of-straightness of
0.03 mm. It was concluded that the reduced
lubrication had allowed the valve to stick slightly
and that it had been hit by the piston. Consequent
operation under the valve closing springs had then
propagated a crack across the stem until fracture
occurred.



Figure 13: Heavy fuel oil contamination of cooling water
passages under valve insert in cylinder head B1.

Cylinder head B1 was examined carefully. All the
valve seat inserts were in place and satisfactory.
Two cracks were observed in the ligament between
the injector sleeve bore and the exhaust valve seat
bore. The water transfer passages between the
cylinder head and liner were contaminated with
black oily deposits. The valve seat inserts were
removed to ascertain the extent of the cracking
damage. Further contamination of the cylinder
head cooling water passages with viscous black oily
deposits was found (Figure 13). Analysis showed
this to be heavy fuel oil which had reduced the heat
transfer characteristics across the flame deck. It is
apparent that the increased temperature reduced
the thermal fatigue properties of the thin ligament
between injector ports and exhaust ports to such an
extent that thermal fatigue cracking took place.
Some components were no longer available for
inspection and the source of the oil was not
identified.
The ships managers were advised that it was
highly improbable that the two damages were
linked.
9. VESSEL 8 MAIN/AUXILIARY ENGINE
CRANKSHAFT
The vessel is a passenger ship with diesel-electric
propulsion machinery. The generators are driven
by nine-cylinder in-line, four-stroke diesel engines
with a bore of 500 mm, a stroke of 560 mm and a
maximum continuous rating of 6,055 kW at 514 rpm.
One of the engines suffered crankshaft damage
after approximately 50,000 running hours. The
engine had been on load in mid-ocean passage
when it was shut down on a high crankcase oil mist
alarm. It was immediately apparent to the ships
engineers that the crankshaft was badly damaged.
Technical Investigations attended on board at the
next port to investigate the possible reasons for the
failure.
The crankshaft was examined. Number 7 crankpin
(towards the free-end of the crankshaft) was found
to have a single, branched crack on its surface
(Figure 14). The crack extended from the starboard
side of the forward crankweb, into the fillet radius
and around the circumference of the journal close to
the forward fillet radius through an arc of some
180. In the other direction it extended at an angle
of 45 across the width of the port side of the
journal.
Cracks extending at 45 to the axis immediately
suggested a torsional vibration problem. The
planes of principal shear strains due to torsion lay
at this angle and torsional cracks will propagate
along these axes. The successful number of
operating hours before the failure suggested that
there was little wrong with the fundamental
torsional vibration characteristics. Nevertheless, the
calculations were re-examined. These were
satisfactory, with the major torsional resonances
away from the operating speed.

Figure 14: Dye penetrant showing extent of cracking on
surface of number 7 crankpin.

185


The running surfaces of the other big end bearings
shells exhibited light scratching and scoring, and
some light cavitation damage, but were considered
to be in satisfactory condition.
Two factors which could have changed the
torsional characteristics were the condition of the
rubber elements of the flexible couplings fitted
between the engine and the alternator and the
condition of the spring-mass tuner-damper at the
free end of the crankshaft. The elements of the
former had been replaced about 2,500 hours
previously and the latter was opened up and found
to be in good condition. Measurements by the
engine builder on the other engines confirmed that
torsional vibration characteristics remained
satisfactory although some changes, due to
hardening of the rubber elements and some
deterioration of the damper, had affected the
system stiffness and natural frequencies.





Figure 16: Crack origin in crankpin 7 after being broken
open.

The crankshaft was cut up carefully to avoid
damaging the fracture faces. The part containing
the crack was sent to Lloyds Registers Materials
Laboratory and opened up for examination. The
crack was found to have two principal origins in the
surface damage caused by rubbing of the bearing
shell in the forward filler radius as seen in Figure 16.
This is not uncommon in cracks associated with
rubbing damage. The fillet radius is a highly
stressed area and, once provided with an initiation
site, the crack propagated under the prevailing
dominant torsional stress regime.

Figure 15: Close up view of crack origin and rubbing
marks in crankpin fillet radius from bearing shell.

The surface of number 7 crankpin had sustained
extensive abrasive and adhesive wear damage and
heat discolouration across the full width and
throughout the circumference of the journal surface.
Rubbing marks from the big-end bearing shells
were visible within the fillet radius around the full
circumference at the forward side of the journal
adjacent to the crack (Figure 15).

The big end bearing shells had suffered overheating
which resulted in distortion and consequent
movement of the shells. The condition of the
damaged big-end housing prevented an assessment
of its geometry. The most likely reason for the
failure is dirt.
The big-end bearing shells showed abrasive and
adhesive wear and heat discolouration in two
separate locations. Corresponding heat marks on
the bearing housing could not be made to line up
simultaneously, indicating that there had been two
separate overheating incidents. The upper shell had
remained in its correct axial location, but the lower
shell was deformed and exhibited ratchet marks
suggesting movement in the axial direction towards
the damaged fillet. Both shells had turned slightly
in the housing in the direction of rotation and the
locating tangs at the butts had been flattened.






186

10. ACKNOWLEDGEMENTS

The author wishes to thank the owners and
managers of the vessels on board which the
investigations were undertaken. Thanks also go to
the surveyors who undertook the investigations, in
particular: Dr. Kian Banisoleiman, Mr. Malcolm
Jowsey, Mr. Joe Stainsby and Mr. Brett Weintz.
Particular thanks must go to Mr. Colin Waylen and
the staff of the Lloyds Register Materials and NDE
Laboratory who undertook the metallurgical
examinations.

Notes
187


























Peter F.G. Filcek

Technical Manager
Technical Investigation Department
Lloyds Register EMEA
71 Fenchurch Street
London EC3M 4BS
Tel: +44 (0)207 423 1765
Fax: +44 (0)207 423 1804
Mob: +44 (0)7768 984443
peter.filcek@lr.org Email:


Lloyd's Register, its affiliates and subsidiaries and their respective officers, employees or agents are, individually and
collectively, referred to in this clause as the Lloyd's Register Group. The Lloyd's Register Group assumes no responsibility
and shall not be liable to any person for any loss, damage or expense caused by reliance on the information or advice in
this document or howsoever provided, unless that person has signed a contract with the relevant Lloyd's Register Group
entity for the provision of this information or advice and in that case any responsibility or liability is exclusively on the
terms and conditions set out in that contract.

Lloyds Register EMEA
T +44 (0)20 7709 9166
F +44 (0)20 7423 2057
E emea@lr.org

71 Fenchurch Street
London EC3M 4BS, UK
Lloyds Register Asia
T +852 2287 9333
F +852 2526 2921
E asia@lr.org

Suite 3501 China Merchants Tower
Shun Tak Centre
168-200 Connaught Road Central
Hong Kong, SAR of PRC
Lloyds Register Americas, Inc
T +1 (1)281 675 3100
F +1 (1)281 675 3139
E americas@lr.org

1401 Enclave Parkway, Suite 200
Houston, Texas, 77077, USA