Вы находитесь на странице: 1из 12

Journal of System Vol. 2, No.

1, 2008

Design and
Dynamics 【Review Paper】

Field Experiences of Gas Turbines Vibrations -


A Review and Case Studies*
M Salman LEONG**
**Institute of Noise & Vibration, Universiti Teknologi Malaysia,
Jalan Semarak 54100 Kuala Lumpur, Malaysia.
E-mail: salman.leong@gmail.com

Abstract
Blade failures represent the highest percentage of failures in gas turbines. This
paper presents some typical examples of blade related failures. A literature review
of common types of blade faults and research on detection methods are presented.
Some methods are however less feasible under practical operating conditions in the
plant. Three case studies of gas turbines vibrations are presented. The first case
reports on stator blades and labyrinth glands rubs resulting in recurring shaft
seizure during the coast down after load removal. Comparison of vibration
spectrum undertaken a day before failure with prior data showed increase in
specific blade passing frequencies (BPF) with increased side bands at intervals of
the synchronous rpm. These increases were from 13 to 28 times above standard
deviations of baseline datum. The stator blades rub was suspected to originate
from a distorted casing. Another case relates experiences of a cracked shaft which
resulted in severe rubbing during a run up. The unit experienced steady increase in
vibration levels on the compressor non-drive end bearing several months prior to
the incident. A full rub occurred as the unit passed through the second critical, with
further development of a thermal bent shaft aggravating the problem with
instantaneous severe vibration excursion. An approximate 200o phase shift was also
noted. A 100mm longitudinal crack on the shaft was found together with signs of
severe rubs on the compressor blades. The third case involved time varying
vibrations where vibration amplitudes and phase angles increased over a time
period which dropped off after each time period; and repeated in regular cycles.
This was due to oil leaks carbonization at the glands resulting in a rub. The paper
concludes with a brief discussion on issues relating to relatively poor detection of
blade faults in the plant.

Key words: Blade Failures, Rub, Cracks, Cracked Shaft, Gas Turbines

1. Introduction
Vibration failures in gas turbines can be rather catastrophic with serious economic
consequences. The turbines operate under relatively high vibratory and thermal cyclic
stresses, especially when used in situations requiring daily start-stops (hot and cold starts)
which are typical in peak load applications in power generation for example. The industry
has seen its fair share of dramatic failures in vibration and other mechanical related failures
in gas turbines.
This paper presents a review of common failures in gas turbines, with several case
studies of vibration related failures in gas turbines based on the experience of the author in
*Received 18 Sep., 2007 (No. 07-0567) the power generation industry.
[DOI: 10.1299/jsdd.2.24]

24
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
2. Examples of Blade Failures in Gas Turbines
The world’s largest insurance company Allianz Insurance (as cited by Meher-Homji,
1995d) reported that blade failures had accounted for as much as 42% of failures in gas
turbines. Some typical failures are presented below.
Fig. 1 shows a case from a 171 MW gas turbine which had 2 catastrophic failures totaling
more than US$ 25million in production downtime, Barnard (2006). Hot section disk cracks
lead to blade liberation (breakage) resulting in hot section blades and discs destroyed.
Regular dental mold taken on machine showed no excessive erosion. Subsequent root cause
analysis showed that the failure mechanism was fatigue driven by blade vibrations: either
from short duration rotating stall during start up and shut down exciting the first tangential
(sub-synchronous) mode of the blade or from synchronous vibration exciting the second
bending mode of the blade.
Photos in Fig. 2 shows lost of part in compressor blades, and rub marks on the rotor from
stator blades from 100 MW gas turbines used in combined and open cycle generation in
Malaysian power plants. Lost of parts and foreign object damage (FOD) are common
failures. The photo also shows rub marks on the rotor from stator blades.

Fig. 1: Photos of damaged turbine blades (left), and missing blade (Barnard)

Fig. 2: Photos of lost part of blades, and stator blade rubs

Less dramatic but more frequent failures are cracks and looseness in the intermediate
packing piece and blade roots which holds the blades to the shaft. Photos in Fig. 3 shows
a gas turbine with 21 stages axial flow compressor section on a single, composite welded
shaft. All the blades in the compressor stator and rotor were made of high-tensile, ferritic
chrome steel; and manufactured partly by milling and partly by precision forging. They are
fitted into radial grooves machined in the shaft and compressor casing, and separated from
one another by intermediate pieces. The rotor blades of the first ten rows were fitted directly
in the shaft by specially shaped roots, while the remaining blades were rooted in
intermediate pieces. There had been frequent experiences of cracks at the roots. These
cracks would cause the intermediate pieces and the blades to become loose, and potentially

25
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
detach from the rotor and consequently causing foreign object damage (FOD) to other
components downstream.

Fig. 3: Roots of compressor disk, and cracks at intermediate piece.

While failures in blades are on its own are already an issue of grave concern, the
consequential damages could be more catastrophic. Photo (a) given in Figure 4 shows a
fire started in a transformer that was hit by a blade that went through a wall which then hit
the transformer. The photo (b) shows a completely damaged compressor section (as a
result of damaged turbine blades for the same machine cited in Fig. 1).

(a) Transformer fire (b) Compressor section


Fig. 4: Consequential damages from blades failures: (a) fire after blade went through wall
and hit transformer; (b) damage in compressor section (Barnard).

3. Literature Review of Blade Related Failures


Gas turbines often operate at the design limit of blades, bearings and combustion
components under extreme conditions. This meant that these components are life limited
and are more likely to experience failures than other less stressed parts. Hot section blades
typically fail because of creep, oxidation, low-cycle fatigue (LCF), and high-cycle fatigue
(HCF). Contributing factors include cyclic loads, over firing, inadequate refurbishment
cyclic loads, corrosion and environmental attack. Hot section blades are life-limited items
and require refurbishment or replacement at intervals dependent upon thermal exposure.
High-cycle fatigue (HCF) is avoided in rotating turbine blades by controlling vibratory
stresses. Long blades are often designed by tuning the lower natural frequencies to avoid
operating near strong excitations and high vibratory stresses. When the natural frequency
variance is greater than an acceptable band, the blades must be designed with sufficient
stiffness to limit stresses when blades operate at resonance. HCF failures could occur if these
conditions are not met or if excessive excitations occur because of vane damage, rotating stall,
flutter, or blade tip rubbing. The variability of blade fabrication can influence natural

26
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
frequencies on assembly; wear and erosion can reduce resonant frequency margins over time.
Either of these conditions can result in high-cycle fatigue damage.
In principle, blade faults diagnosis can be undertaken from measurements and
monitoring of gas turbine operating parameters such as the pressure, vibration, strain and
stress, and acoustic signals in an attempt to obtain information to assess the blades’ condition.
This is however easier said than done under practical operating situations in the plant.
Common blade faults could be broadly classified into the following categories: rubbing,
cracking and foreign object damage (FOD) or lost part, blade deformation (twisting,
creeping, corrosion and erosion), blade fouling and rotating stall, blade fatigue failure, and
blade root attachment problems (root crack and loose blade). A review of each of such faults
and diagnostic methods proposed in the literature is presented.

3.1 Blade Rubbing


Meher-Homji (1995b) reported that blade rubbing in gas turbine could be detected from
increased rotor sub-harmonic, sub-synchronous vibration at the rotor natural frequency
observed in the vibration spectrum. It was also suggested that the acoustic diagnostic
method is viable to detect blade rubbing using stethoscopes during gas turbine start up and
shut down, Meher-Homji (1995b). Simmons et. al (1992) reported that the measurement of
the blade tip clearance using optical probe could provide early indication for blade rubbing
problem if the measured blade tip clearance has reduced. Fisher (2000) commented that
techniques such as the performance and vibration monitoring are dependent on a secondary
effect being produced by the fault. For instance, blade rubbing or other blade faults could
only be monitored by these techniques provided sufficient material loss had occurred and
results in an imbalance or loss of performance. Kubiak et. al (2001) reported that blade
rubbing could also be detected if the blade passing frequencies (BPF) amplitude is found to
be extremely high in the vibration spectrum.

3.2 Blade Cracking, FOD and Lost Part


Beebe (1998) in a similar work for steam turbines reported that the breaking off of the
blade (lost part) could be detected using the vibration analysis and performance monitoring.
Yuan et. al (1999) used the short time Fourier Transform (STFT) and Mallat wavelet
decomposition method to analyze the vibration signal of a cracked blade. It was shown that
this method was viable in detecting cracked blades based on the presence of
higher-frequency impact signals in the STFT maps and wavelet decomposition map. These
impact signals were absent in these vibration maps if the blade is in good condition.
Doftman et.al (1999) presented in their paper that the turbine blade root cracks was
induced by the fatigue initiated by the rotor torsional vibration. Torsional vibration is
generally a sporadic, transient phenomenon resulting from a sudden load changes due to
changes in load. The measurement and monitoring of the torsional vibration could then be
used to provide early warning for the turbine blade root crack if the torsional vibration
detected differed from the baseline data. This method was also reported by Maynard (2001).
Damage of turbine blades, including cracks, tends to shift the blade resonant frequencies.
The monitoring of the blade’s resonance could therefore be adopted as a blade damage
indicator. Based on this principle, Mercadal et. al (2001) proposed a method using a
non-contacting stress monitoring system to monitor the blade’s resonance in an attempt to
diagnose the cracked blade. In addition, turbine blades crack susceptibility could also be
predicted based on the life evaluation method using fracture mechanics theory in computer
simulation and modeling. This method has been studied by Bernstein et.al (1992), Rao et. al
(2001) amongst others.

27
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
3.3 Blade Deformation (Twisting, Creeping, Erosion, and Corrosion)
Blade deformation faults in principle could be detected by measuring the pressure field
around the blades since the deformed blade would alter the pressure field around its body.
This method had been proposed by Mathioudakis et. al (1991). Beebe (1998) commented
that the erosion of the blade due to the solid particles could also be detected using the
performance monitoring method when turbine efficiency drops.

3.4 Blade Fouling and Rotating Stall


Meher-Homji (1990) presented a case study on blade fouling in axial compressor. He
found that compressor blade fouling could be detected based on performance monitoring. A
drop in the overall power output and thermal efficiency of the gas turbine could potentially
indicate the presence of fouling. Meher-Homji (1995a) commented that rotating stall in the
compressor section could also be detected by using performance monitoring. When stall
occurs, the overall performance of compressor decreased thus indicating a likely fault.
Mathioudakis et. al (1991) reported that measurement of the pressure field around rotating
blades in turbomachinery is more sensitive in detection of fouled blades than conventional
performance monitoring.

3.5 Blade Fatigue


Flow-induced resonant vibration is usually the cause for the blade fatigue failure in gas
turbine. Blade fatigue could be detected using the strain gauge method. This method was
studied by Mercadal et. al (2001) amongst others.

3.6 Blade Root Attachment Problems (Loose Blade)


Kuo (1995) studied the diagnosis of the loose blade with Fourier analysis of the
vibration signals. He used neural networks and fuzzy logic methods to develop a pattern
recognition algorithm to enhance the vibration analysis of the loose blade. Lim et.al (2005)
reported on laboratory studies on the use of wavelet analysis in the detection of loose
blades. Changes in the wavelet decomposition maps during run up coast down conditions
were demonstrated to show up loose blades which were not apparent in the FFT spectrum.

3.7 Use of Vibration Measurements in Blade Faults Diagnosis


Under practical field conditions, vibration inevitably represents the most readily and
widely used method for blade fault diagnosis. Vibration signal is conventionally analyzed
with Fourier analysis to obtain the vibration spectrum. Mitchell (1975) presented that the
relative changes in the blade passing frequency (BPF) and its harmonics could provide
useful information. BPF amplitude was reported to increase at both low flow (near surge)
and at high (approaching choke) conditions. Baines (1987) however commented that
vibration analysis could only detect the blade faults if severe damage had occurred at the
blade, while a minor shape modification of a few blades would not be evident in the
spectrum. The use of vibration analysis for blade faults diagnosis was also studied by
Simmons et. al (1986, 1987) and Parge (1990). It was reported that blade faults could be
detected from observations of relative changes in the BPF and its harmonics amplitude.
Mathioudakis et. al (1990) conducted an experimental investigation of the compressor
casing vibration correlated to the pressure field around the compressor blade for different
blade faults condition. In the investigation, he found that the unsteady pressure field
measurements provided a clearer picture of the blade faults than the conventional vibration
signals. Using non-parametric identification methods, acceleration outputs were correlated
to unsteady pressure measurements taken by fast response transducers at the inner surface
of the compressor casing. The transfer functions allow reconstruction of unsteady pressure
signal features from the accelerometers readings. It was demonstrated that information

28
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
could be deduced from casing measurements without intrusive measurements in the inner
working sections. Meher-Homji (1995c) commented that a considerable amount of
experimental work still needs to be done to correlate blade problems with the vibration
signatures, but there is little doubt that useful information does exist in the vibration
signatures as demonstrated from the case studies presented below.

4. Shaft Seizure as From Stator Blade Rubs


Four gas turbines each rated 100MW were used in a combined cycle plant. The turbines
had been in service for several years, but had no baseline vibration spectrum taken during
new conditions. Baseline spectra blade faults diagnosis was nevertheless obtained under
existing operating conditions. Periodic monitoring was undertaken over several months.
Detailed vibration spectra at base load condition were obtained on a particular unit GT4
before a scheduled overhaul. The measurements were then correlated to the blades’
condition during overhaul inspections. During the overhaul, minor maintenance works on
the compressor blades were conducted. No major defects were found on rotor and stator
blades. Measurements immediately after the overhaul were undertaken to be used as
baseline information for blade faults diagnosis based on the assumption that all blades in
GT4 were in good condition immediately after the overhaul. Comparison of vibration
spectra of GT4 to the three other operating units was also undertaken. Assessment of BPF
and sidebands with statistical analysis to determine standard deviations under normal
conditions were undertaken to establish alarm limits for blade faults diagnosis.
Three months of operation after the overhaul, rubbing like noise was audible from the
compressor section during the coast down of GT4. Under steady state operating
conditions (with load and full speed no load), this noise was inaudible. Plant personnel
also noted that the coast down time of the unit from load to stand still conditions reduced
significantly. Under prior normal conditions, the coast down time was approximately four
hours which had now reduced to approximately 20 minutes instead.
A vibration investigation was initiated. Typical vibration spectra of the measurement
are given in Fig. 5(b) & (d). Vibration spectrum at the same measurement location and
direction undertaken six months earlier are given in Fig. 5(a) and (c). Comparison of the
spectrum showed distinct increase in specific blade passing frequencies (BPF), and with
increased side bands of the BPF at intervals of the synchronous rpm. BPF at 3050 Hz,
3250 Hz, 3850 Hz and 3900 Hz were exceptionally high in vertical axis. Measurements in
axial axis showed that the BPF of 3050 Hz was extraordinary high. The BPF of 3050 Hz
(corresponding to compressor blades row 11, 12 and 14) in the vertical and axial directions
showed increase with standard deviations as high as 13 times. For BPF of 3250 Hz (row
13), an equally high amplitude exceeding 28x standard deviations was noted. Other BPF of
3850 Hz (row 15-21) and 3900 Hz (stator blades row 15-20) also increased above 7x
standard deviation. Typical trends of the BPFs are shown in Fig. 6. Multiple harmonics of
3400 Hz, 3500 Hz, and 3600 Hz were also noticeable. Correlated with the physical evidence
of reduced coast down time (which was obviously a strong indicator of increased friction
i.e. rub) and with an audible rubbing sound, it was suspected that there were blade rubs. It
could in fact be concluded with sufficient confidence that blade faults had occurred in the
latter rows of compressor section of GT4.
Despite these concerns on the unit, the turbine was to remain in operations since the
overall vibration levels (in velocity mm/s) as recorded in the DCS were substantially below
alarm levels. The rotor of GT4 seized up the following day after this measurement during
its coast down. The unit could not be turned (barring) until the subsequent day after
cooling down to cold conditions. Due to load demands from the electricity grid, the unit was
brought back to service. This resulted in a recurring shaft seizure during coast down.

29
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics 3.5
3.5
Overall 1.300 G RMS Overall 3.847 G RMS
3.0 3.0
0.386 mm/s RMS 0.911 mm/s RMS

Am plitude G 's R M S
Amplitude G's RMS
2.5 2.5

2.0 2.0

1.5 1.5

1.0 1.0

0.5 0.5

0.0 0.0
0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000
Frequency Hz Frequency Hz

(a) Axial direction: 6 months before seizure (b) Axial direction: 1 day before seizure
1.50 1.50
Overall 1.307 G RMS Overall 2.273 G RMS
0.723 mm/s RMS 0.800 mm/s RMS
1.25 1.25

Amplitude G's RMS


Amplitude G's RMS

1.00 1.00

0.75 0.75

0.50 0.50

0.25 0.25

0.00 0.00
0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000
Frequency Hz Frequency Hz

(c) Vertical direction: 6 months before seizure (d) Vertical direction: 1 day before seizure

Fig. 5: Acceleration spectrum as measured on bearing 6 months and 1 day before shaft seizure

The extremely high BPF (particularly 3050 Hz, 3250 Hz of rotor rows 11-21 and 3850
Hz, 3900 Hz of stator rows 15-20) implied that the clearance between the rotor with stator
and casing on these rows had been reduced. This was consistent with published works of
Kubiak et.al (2001) that reported that if the BPF had increased significantly in the vibration
spectrum, blade rubbing could probably be present. The multiple harmonics noticed
around 3400 Hz to 3600 Hz (which is not the BPF) in the vibration spectrums provided
further evidence of rubbing. The rubbing between the rotor and stationary components
could generate excessive localized heat and potentially causing the rotor to bow. A bowed
rotor could further increase the rubbing. This therefore explained why the rotor seized up.
On the next day, with cooling the bowed rotor in all likelihood straighten which made it
possible for the rotor to turn - only to be brought back to service with obvious consequences
of a repeat shaft seizure that obviously confirmed the vibration assessment.
An emergency outage was
initiated on the unit for a physical
inspection. It was found that some
compressor stator blades particularly
at row 20 and row 21 at the upper
part of compressor casing showed
traces of rub marks at the tip of the
blades. This finding was in good
agreement with the assessment that
the higher BPF seen on the vibration
spectrum was due to the rubbing in
gas turbine. Fig. 6 Trending of selected BPFs

30
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
Further inspections confirmed more severe and underlying faults on labyrinth glands.
The labyrinth glands immediately after the last row of compressor blades experienced a
severe rubbing against the rotor shaft. The rubbing was approximately 3 mm in depth which
could have reduced the efficiency of the unit due to the pressure leakage in the compressor.
Root cause analysis of the rubbing in the unit suggested that rotor eccentricity was the
likely reason of the rub. Radial blade rubs are a more common type of blade tip rubbing
caused by excessive differential expansion of the rotor and casing components during
transient event (shut down, emergency trip, start up). Neither of these anticipated conditions
was found after the inspection. There was a suggestion by the OEM that the root cause
could be due to the compressor casing distortion. Casing distortion in turn could be due to
the non-uniform insulation on the compressor casing. Major correction works on the unit
was carried out to improve the insulation on the compressor casing. The unit was put into
service again although the root cause of the rubbing was not conclusively identified.

5. Cracked Shaft and Rotor Rub


This case was for a 100 MW in another power generation plant in Malaysia. Several
months before the crack was detected, steady but significant increase in vibrations were
noted at the control room monitoring displays (overall levels).
12

10

8
Velocity (mm/s Pk)

Brg No.1
Brg No.2
6 Brg No.3
Brg No.4
Brg No.5

0
6-Jan

7-Jan

8-Jan

9-Jan

10-Jan

11-Jan

12-Jan

13-Jan

14-Jan

15-Jan

16-Jan

17-Jan

18-Jan

19-Jan

20-Jan

21-Jan

22-Jan

23-Jan

24-Jan

25-Jan

26-Jan

27-Jan

28-Jan

29-Jan

30-Jan

31-Jan

1-Feb

2-Feb

3-Feb

4-Feb

5-Feb

Date

Fig. 7: Overall levels for gas turbine with (subsequent detected) cracked shaft

Fig. 7 shows trends of the x1 synchronous component for bearing casings in vertical
direction over a month time span. Bearing No. 1 (compressor non-drive end), No. 3
(turbine drive end) showed marked increase over the latter two weeks period. Vibration
levels in the four months period prior to this were less than 1 mm/s for bearings No. 1 and
3, and 3.3 mm/s for bearing No. 5 (generator non-drive end). The alarm limit was 12mm/s
and a trip level at 25mm/s (0-Pk). A vibration investigation was initiated a month later.
The synchronous amplitude and phase angle plots against rpm are shown in Fig. 8.
The vibration spectra during the run up past the first critical speed and as the unit
approached the second critical speed were dominated by the synchronous x1 RPM
component. Vibration amplitudes increased substantially as the second critical of 2280
rpm was approached. At this stage, high harmonics of x2, x3 and x4 rpm were noted.
This suggested a rotor rub at the second critical due to excessive amplification as the unit
approached this critical. After passing the second critical, vibration levels dropped with less
prominent harmonic components.

31
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
This suggested the rub being relieved at this point in time. As the unit approached full
speed no load (FSNL) conditions at 3000rpm vibration levels were below trip levels. The
vibration spectrum at speeds before and just at full speed no load is shown in Fig. 9. While
settling at full speed (with marginal over speed to 3008 rpm) an instantaneous increase of
vibration levels most severe at the outboard (non-drive end) compressor bearing with peak
velocity levels of up to 194 mm/s occurred. This was accompanied by a phase shift of
approximately 200o. A manual trip of the turbine was in progress.

40

30
Velocity mm/s
20

10

0
0 500 1000 1500 2000 2500 3000 3500
RPM

180

120
Phase (degree)

60

0
~200o phase shift
-60

-120

-180
0 500 1000 1500 2000 2500 3000 3500
RPM

Fig. 8: Run up amplitude (x1 RPM) and phase angles


During the coast down, the dynamic amplification at the second critical was more
severe as compared to the run up; Fig. 10 (a) and (b). The vibration response in the
vertical directions was significantly higher than response in the horizontal direction
particularly in the compressor NDE (Brg. 1) and turbine DE bearings (Brg. 3). A cracked
shaft and/or thermally bent shaft were suspected to the root cause of the increasing
vibrations particularly at the compressor bearing. A full rub was thought to have occurred
as the unit passes through the second critical, with further development of a thermally bent
shaft aggravating the problem and hence the severe vibration excursion.

(a) During run up at 2826 rpm (b) During run up at 3008 rpm (full speed no load) with rub
Fig. 9: Forced vibration response during run up (with rotor rub in progress)

32
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics

(a) During run up

(b) During coast down


Fig. 10: Peak hold spectra during run up and coast down after a rub on Brg. 1 (vertical)

A subsequent run up investigation of the gas turbine was initiated by the plant with the
generator uncoupled to isolate any fault from the generator. This run-up re-confirmed
higher response in the vertical directions than the horizontal, with highest response at the
compressor bearing. The unit was removed for an inspection where a 100mm longitudinal
crack was found in the compressor section, together with signs of severe rubs on the
compressor blades. This resulted in a downtime of more than 8 months awaiting a
replacement shaft at a time when there was almost no spare capacity in the national grid.

6. Time Varying Vibrations from Seal Rubs


This study while less dramatic was initially a little puzzling. Scheduled maintenance
work was undertaken on a 100 MW gas turbine. When the unit was brought back to
service vibration of the turbine bearing was observed to increase gradually after start up;
increasing from amplitudes of less than 2 mm/s initially to 7 mm/s. Trim balancing works
were attempted by the OEM which did not reduce the vibrations. It was also noted that the
turbine vibrations after some time (up to 8 hours operation) in fact dropped and increased
with time over a cycle of 30 to 150 minutes. A typical chart recording showing bearing
vibrations and exhaust temperature (confirming steady load conditions) are given in Fig. 11.
It was also observed that phase angles were changing with time together with the amplitude
time variations. This phenomenon was repeated with daily operations.
Sacchi et al (1982) reported a similar case where an oil gland rub resulted in a vibration
with continuous phase angles increase. The fault was identified as oil leaks where the seal
was completely covered by a carbonized oil layer whose thickness was such that it reduced
the diametrical clearance between the seal and shaft. The problem was related to
insufficient seal pressurization, where oil leakage from the seal into a hot ambient was
subjected to a carbonization process with a progressive formation of a carbonized oil layer.
High vibration levels resulted in wearing of the carbonized oil layer; and the process was

33
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
repeated. Inspection of the oil glands on this unit confirmed signs of leaks, and the problem
resolved with rectifications of the oil gland and restoration of proper pressurization.

Time varying
vibrations over ~30 to

Fig. 11 Station chart showing bearings vibrations and process parameters

7. Issues in Poor Detection of Blades Related Failures


There ought to be concern that the single largest failure in gas turbines are often not
readily detectable by conventional overall vibration monitoring as evident from the field
experiences presented here. Protection systems as used in most installations in the field
are traditionally based on overall vibration amplitudes. These overall amplitudes are more
often than not dominated by the synchronous and possibly second and third harmonics of
rpm. As such only x1, x2 and higher components associated with mass unbalance and
misalignment would be readily detected. Blades related failures do not generate sufficient
excitation that would show up as a substantial increase in the overall vibration levels.
Plant operators are often concerned only if the vibration levels approaches the alarm or trip
levels. This therefore often results in blade related failures not detected.
The case studies also showed that blade related faults can indeed be detected before
catastrophic failure provided adequate analysis was undertaken. This involved analysis of
blade passing frequencies, sidebands of BPF, and wavelet analysis. Some faults such as
looseness at blade roots are also not readily detected under steady state conditions, but only
during transients with wavelet decomposition maps for example. There is still a gap
before some of the diagnostic techniques that showed promise undertaken in research find
its way into practice for operational machines.

References
(1) Baines, N., Modern Vibration Analysis in Condition Monitoring; Noise and Vibration
Control Worldwide, May 1987, pp.148-151.
(2) Barnard, I., Engineering Asset Management: An Insurance Perspective; 1st World
Congress on Engineering Asset Management, Australia, July 2006.
(3) Beebe, R., Condition Monitoring of Steam Turbines by Performance Analysis; 52nd
Conference of the Machinery Failure Prevention Society, Virginia Beach, 1998.
(4) Bernstein, H.L., Allen, J.M., Analysis of Cracked Gas Turbine Blades, Journal of
Engineering for Gas Turbines and Power, Vol. 114, April 1992, pp.293- 300.
(5) Dorfman, L.S., Miroslav, T., Torsional Monitoring of Turbine-Generators for Incipient
Failure Detection; Sixth EPRI Steam Turbine/Generator Workshop. August 17-20, 1999.
St. Louis, Missouri.

34
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
(6) Fisher, C., Gas Turbine Condition Monitoring Systems - an Integrated Approach; IEEE
Aerospace Conference 2000.
(7) Kubiak, J., Gonzalez, G., Garcia, G., Urquiza, B., Hybrid Fault Pattern for the Diagnosis of
Gas Turbine Component Degradation, International Joint Power Generation Conference,
New Orleans, Louisiana., June 4-7, 2001, Paper no: PWR-19112.
(8) Kuo, R. J., Intelligent Diagnosis for Turbine Blade Faults Using Artificial Neural Network
and Fuzzy Logic; Engineering Application of Artificial Intelligence, Vol. 8, 1995, 25-34.
(9) Lim, M.H., Leong, M.S., Diagnosis for Loose Blades in Gas Turbines Using Wavelet
Analysis, ASME Journal of Eng. for Gas Turbines and Power, Vol. 127, 2005, pp. 315-322
(10) Maynard, K., Application of Torsional Vibration Measurement to Blade and Shaft Crack
Detection in Operating Machinery; Maintenance and Reliability Conference, Gatlinburg,
Tennessee, May 6-9, 2001.
(11) Mathioudakis,K., Loukis, E, Papailiou, K.D., Casing Vibration and Gas Turbine Operating
Conditions; ASME Journal of Eng. for Gas Turbine and Power, Vol. 112, 1990, 478-485.
(12) Mathioudakis,K., Papathanasiou, A., Loukis, E., Papailiou, K.D.,Fast Response Wall
Pressure as a Means of Gas Turbine Blade Fault Identification; ASME Journal of
Engineering for Gas Turbine and Power, Vol. 113, 1991, 269-275.
(13) Meher-Homji, C.B., Gas Turbine Axial Compressor Fouling – a Unified Treatment of its
Effects, Detection and Troubleshooting; ASME Cogen Turbo IV, New Orleans, August
27-29, 1990.
(14) Meher-Homji, C.B., Blading Vibration and Failures in Gas Turbines: Part A – Blading
Dynamics & the Operating Environment, 1995, ASME paper no. 95-GT-418.
(15) Meher-Homji, C.B., Blading Vibration and Failures in Gas Turbines: Part B – Compressor
and Turbine Airfoil Distress, 1995, ASME paper no. 95-GT-419.
(16) Meher-Homji, C.B., Blading Vibration and Failures in Gas Turbines: Part C – Detection
and Troubleshooting, 1995, ASME paper no. 95-GT-420.
(17) Meher-Homji, C.B., Blading Vibration and Failures in Gas Turbines: Part D – Case
Studies; 1995, ASME paper no. 95-GT-421.
(18) Mercadal, M., VonFlotow, A., Tappert, P., Damage Identification by NSMS Blade
Resonance Tracking in Mistuned Rotors, IEEE Aerospace Conference Proceedings, Vol. 7,
2001.
(19) Mitchell, J., Examination of Pump Cavitation, Gear Mesh and Blade Performance Using
External Vibration Characteristics, Proceedings of the 4th Turbomachinery Symposium,
Texas A&M University, Oct 14-16, 1975, pp. 39-45.
(20) Parge, P., Non-Intrusive Vibration Monitoring for Turbine Blade Reliability, Proceedings
of Second International Machinery Monitoring and Diagnostic Conference, Los Angeles,
Oct 22-25, 1990, pp. 435-446.
(21) Rao, J.S., Pathak, A., Chwla, A, Blade Life: A Comparison by Cumulative Damage
Theories, Journal of Engineering for Gas Turbine and Power, Vol. 123, No. 4, October
2001, pp. 886-892.
(22) Saachi G., Moro, B., Vibratory Problems on Gas Turbines Rotors, International Federation
for Theory of Machine and Mechanisms Conference Proceedings, Italy, 1982, pp. 7-15.
(23) Simmons, H.R., Non-Intrusive Detection of Turbine Blade Resonance, Third EPRI
Conference on Incipient Failure Detection in Power Plants, Philadelphia, 1987.
(24) Simmons, H.R., A.J. Smalley, C.E. Edlund, and R.W. Frischmuth,. Monitoring
Compressor Blade Tip Clearance in Combustion Turbines, EPRI Steam and Combustion
Turbine Blading Conference, Orlando, Florida, January 29-31, 1992.
(25) Simmons, H.R., A Non-Intrusive Method for Detecting HP Blade Resonance, 1996,
ASME Paper No. 86-JPGC-Pwr-36.
(26) Yuan, Q., Xie, H., Meng, Q.J., Fault Diagnosis of Turbine Blade by Wavelet, A-PVC
Conference, Singapore, 1999, Vol. 1, pp. 170-175.

35

Оценить