Presentation for the 8th international
Conference on Engineering failure analysis
-
ROOT CAUSE FAILURE ANALYSIS OF A TORSIONAL
FATIGUE FAILURE OF THE PROPULSION LINE OF A
SUEZ-MAX CLASS VESSEL
A MULTIDISCIPLINARY APPROACH
1
Techno Fysica is a combination of two departments
of two individual former large Dutch shipyards,
which merged as a result of the yards going into
bankruptcy
Techno Fysica has two individual business units:
Techno Monitoring
Techno Diagnosis
2
Techno Monitoring:
• Designs, produces and installs high-
end custom made load monitoring
systems and installations, mainly for
offshore applications and dredging.
• Service and calibration and retrofit of
third party load cells
• Software for specialized applications
3
Most common type: shear pin, in all sorts and sizes
4
Some typical applications:
5
With the latest achievement the measuring installation of Allseas “Pioneering Spirit”,
here during calibration of the system in Rotterdam harbour
6
Techno Diagnosis has the following main activies:
• Measurements of a large array of physical paramaters, among which:
•stress
•Strain / deflection
•Vibration measurements and analys
•Pressure / combustion behaviour
•Condition monitoring
•And more
• Laboratory for metallurgical failure examination.
• Calculations, mainly related to dynamically loaded installations, for
example torsional vibration calculations, Finite element (FEM)
calculations, alignment calculations etc.
7
8
The subject for this presentation:
ROOT CAUSE FAILURE ANALYSIS OF A TORSIONAL FATIGUE
FAILURE OF THE PROPULSION LINE OF A SUEZ-MAX CLASS
VESSEL - A MULTIDISCIPLINARY APPROACH
9
In the propulsion installation of a Suez Max class tanker, cracking was
noted in the corners pockets of the intermediate shaft of the propulsion
installation.
The propeller of this installation was a mechanically controlled pitch
propeller, and control of the pitch occurred through a slot, present in the
intermediate shaft.
10
The failure was investigated by DNV , with main conclusions:
• low-cycle torsional fatigue, and initiating from the corners of
the slot, which had acted as stress raisers.
• No flaws in the material or manufacturing were found to be
present.
The vessel was one of a series of 4 newly build shuttle tankers,
equipped with comparable installations.
11
The calculated service life for these vessels was 75 year, well in excess
of projected service life of the vessel. However, the true service life was
found to be less than 6 year.
The ship met with all class regulations and the necessary design
calculations were made (and found to be in order)
The question of course was : Why…?
And: can this happen on the other ships too?
12
Approach:
Trial measurements,
The main purpose of these measurements was to verify design conditions.
In addition, a data logging installation was installed on board, with the
purpose of obtaining an operational profile of the installation during 10
weeks, rather than measuring during “clean” trial conditions.
The main purpose of this was to establish if off-design circumstances could
attribute to the occurrence of the damage.
13
Stress measurements:
Stresses (using strain gauges) were measured in an
undisturbed section of the shaft, as well as in the
inner radius of the slot.
The stress as measured in the undisturbed section
is used as reference. The stress as measured in the
slot was measured for evaluation only.
14
Torsional Vibration Calculation and measurements results during trials
The graph (next page) contains a summary of the TVC, adapted to the
measured dynamic load on the instrumented location.
15
Torsional vibratory torque in Stb OD shaft of HN 1749 SHI
Measured 08 July 2010
Torque [kNm]
800 Barred
750
700
650
600
550
500
450
400
350
300
250
200
150
100
50
0
0 10 20 30 40 50 60 70 80 90 100 110
Shaft speed [rpm]
Dyn. Torque
Speed up slow down Propeller curve Nominal Calculated
Ballast condition
16
Torsional Vibration Calculation and measurements results during trials
The maximum calculated dynamic torque amounts to 760 kNm at 41.2 rpm
compared to the maximum measured dynamic torque of 748 kNm at 44
rpm in the undisturbed diameter of ø390xø125 mm which is a good fit.
The small difference (less than 7%) between measured- and calculated 1st
node natural frequency is most likely caused by a difference in the
assumed amount of entrained water in the propeller.
17
With the measured dynamic torque of 748 kNm (0-peak) at 44 rpm, the
measured shear stress in the OD shaft at the diameter ø530xø210 mm
(where also the slot is located) amounts to 26.3 N/mm².
Note that this is the stress in an undisturbed diameter, not in the slot
itself
The OD shaft was modelled in a
FEM model, based on drawings
of the shaft, submitted by the
manufacturer.
The model is shown here:
18
The maximum calculated shear stress in the slotted area amounts to a
value of 108 N/mm².
Given the measured maximum shear stress in the undisturbed diameter
in the area of the slot (ø530xø210 mm) of 26.3 N/mm² , this results in a
SCF of 4.1.
The empirical calculated SCF amounts to 3.84 which proves that this
calculated value is conservative compared to the value that follows from
the FEA (+7%).
This same conclusion was also drawn by DNV in their investigations.
19
20
The results from the FEA calculations were that the SCF that
will be used in the provisional life time calculation is 4.1, the
maximum shear stress amplitude at resonance at 44 rpm in the
undisturbed diameter is 26.3 N/mm² (based on manufacturer
measurements) and the maximum shear stress in the slot is
108 N/mm².
These values are used to compare the findings after the sea
trials by the manufacturer to the findings after the trials by
Techno Fysica to compare life time expectance.
Again, note that these are design loads.
21
Instrumentation
To link theory to every day practice, Techno Fysica has performed the
following instrumentation:
1. Mean- and dynamic torsional (shear)stress measurements in the OD
shaft (ø390 mm x ø125 mm) by means of strain gages on both
installations.
2. High resolution shaft speed measurement on both installations.
3. Mean- and dynamic stress measurement on the location of the cracks by
means of strain gages on the SB installation (ø530 mm x ø210 mm).
4. Measurement of the pitch setting (mechanical) on both installations.
All signals are recorded on line and simultaneously on a 16 channel Data
recorder for further analysis.
22
Such measurements have an impact on operations.
Therefor, due to operational restrictions, the measurement
program was cut into three pieces:
1. Measurements under low load conditions.
2. Measurements under loaded conditions in the North Sea.
3. Automatic data acquisition during 10 weeks.
23
Measurement results
Mean torque- and power curves
To have a reference concerning developed torque and shaft
power we measured the mean torque and shaft power on
both installations under ballast conditions for the following
operational modes:
• Transit combinator mode
• Harbour combinator mode
• Constant speed mode
• Separate rpm mode
24
Measured load during controlled operation
MEAN SPEED POWER % OF
MODE TORQUE [RPM] [KW] NOMINAL HANDLE
[KNM] PS/SB PS/SB PS/SB
PS/SB
Transit
combinator 682 / 677 104.5 / 104.5 7461 / 7405 96 / 95 9
Harbour
combinator 497 / 519 105 / 105 5446 / 5705 70 / 73 10
Constant
speed 686 / 673 104.5 / 104.9 7562 / 7391 96 / 95 9
Separate rpm 548 / 560 104.3 / 104.4 5986 / 6123 77 / 79 -
25
Measured load during automatic data acquisition
MEASUREMENT MAX TORQUE MIN SPEED MAX TENSILE/SHEAR
[KNM] TORQUE [RPM] STRESS SLOT
[KNM] [N/MM2] (0-PEAK)
Manufacturer Sea trials 748 -730 44 189/109
TF in Stavanger 1032 -1000 46 261/151
TF Automatic data 1499 -1388 48 379/219
acquisition
26
27
Fluctuations
during speed
up…….
28
….And during
speed down
29
Clearly undesirable regulator behaviour during a speed-up
30
In the last graph it can clearly be seen that the installation reaches the
desired speed setting of 63 rpm from standstill in 2.5 seconds.
However, the speed drops again and remains in the barred speed range for
approximately 12.5 seconds before stabilising at 63 rpm which leads to an
increase in the number of cycles that contribute to the fatigue load from 4
(single pass) to 25 which means an increase by a factor of 6.25.
31
It is worth noticing that the maximum torque (and therefore stress in the slot) that
occurred during the period of automatic data acquisition was higher by a factor of 2
compared to the value that was measured by the manufacturer during initial sea trials.
This means that during normal operation and without active aware intervention by the
crew, the dynamic load can be high without being noticed which of course has a
negative effect on expected life time of the OD shaft.
This behaviour was discussed with the engine manufacturer, and they found it very
hard to believe that dynamic torque could be twice as high as during sea trials.
However, the setting of the torque meter was checked and double checked and found
to be correct which was proven by the realistic values of the mean torque during the
measurements which comes from the same measurement system.
Therefore the conclusion that dynamic torque as high as 1500 kNm can occur in the
shaft line is justified.
32
Average dynamic load during the measurement period
The time frame during which the system has recorded data stretches to 1551
hours.
During this period, the main engines were stopped for 441 hours which leaves
1110 hours of data with the main engines in operation over a period of 10.5
weeks.
Measurements were also taken during DP operation.
Under these conditions, the engine speed is approximately 63 rpm and does not
pass the BSR.
This means that the life time of the OD shaft is mainly determined by the number
of start-stops and the way these cycles are passed, i.e. how much fuel is injected
and does the engine pass the barred speed range once or, as it is now, three to
five times during each pass
33
Life time prediction
Based on the measurements performed until now, the life time
prediction of the OD shaft can be summarised as follows:
MEASUREMENT PALMGREN-MINER TOTAL START-STOP/WK LIFETIME
30 YEARS [YEARS]
Manufacturer sea trials 0.2 10 > 30
TF in Stavanger 0.61 6 15
TF Automatic data acquisition 0.59 2.6 15
34
•Conclusions
1. Based on a continuous measurement of the operational profile of the
ship over a period of 10.5 weeks, the expected lifetime of the OD
shafts, based on the Palmgren-Miner criterium, will be approximately
15 years.
2. This estimate is based on an average of 2.6 starts/stops each week
and will be reduced if the number of starts/stops increases.
3. On one occasion the maximum measured tensile stress in the slot of
the OD shaft exceeded the specified yield strength of 300 N/mm² for
the used material (CK 45 mod) by 26% (380 N/mm² 0-peak).
4. Stresses in the range of the yield stress were measured. Techno
Fysica is of the opinion that the yield strength should never be
exceeded under any given condition.
35
5 The maximum level of vibratory torque is approximately twice as high as
during the measurements by the manufacturer during sea trials (1500
kNm 0-peak compared to 748 kNm 0-peak).
6 The stress concentration factor between the undisturbed diameter of
the OD shaft (ø530 mm x ø210 mm) and the slot for shear stress is
4.174 which was calculated and supported by measurements.
7 Given this factor we would advise to find a different way to get feed
back from the pitch setting because this slot has a very high
influence on fatigue life of a shaft in a direct coupled two-stroke
installation.
8 Without the slot, the expected lifetime would be infinite, no matter
the realistic level of vibratory torque.
36
The main cause of the problem:
The speed control of the engine during passing the BSR is unstable,
resulting in several passes during a single speed up or slow down
action which results in higher amplitudes and up to 6 times the number
of cycles with high amplitude compared to a single, straight pass.
Although the other ships were not measured by Techno Fysica, we
conclude that the behaviour will quite possibly be the same, based on
the sea trial data as provided by the manufacturer.
Based on the outcome of the investigation, the advice was given to
change the settings of the engine with regards to transient behaviour.
37
The end
Thank you for your attention
For more information,
please check:
www.technofysica.com
or
Email us at:
info@technofysica.com
38
