Failure analysis of gas turbine rotor blades

Sweety Kumari

, D.V.V. Satyanarayana, M. Srinivas

Defence Metallurgical Research Laboratory, PO: Kanchanbagh, Hyderabad 500 058, India
a r t i c l e i n f o
Article history:
Received 13 March 2014
Received in revised form 16 June 2014
Accepted 16 June 2014
Available online 15 July 2014
Keywords:
Failure analysis
Turbine blade
Superalloy
Fatigue
Internal object damage
a b s t r a c t
The failure of the stage I, II and III turbine rotor blades of an aircraft engine was investi-
gated by metallurgical analysis of the failed/damaged blades. The blades were made out
of Ni-based superalloys of different grades. The aeroengine has completed about 80% of
the assigned life since new and several hundred hours since the last overhaul before the
failure of the blades. Two blades of I stage were found broken at the top and several blades
of this stage were also found with deep cuts on one of the edges. Further, several II and III
stage turbine blades have dents and nicks on their leading/trailing edges. Detailed investi-
gation including visual examination of the blade surfaces, fractography, micro structural
examination, chemical analysis and hardness measurement was carried out to identify
the cause of the failure of the blades. The investigation has revealed that the damage of sur-
face coating has caused severe localized oxidation attack to I stage blade A, leading to the
formation of oxide at interface between coating and substrate and pits as well as dislodge-
ment of surface coating at several locations on the surface of the aerofoil. Fatigue cracks
have initiated at these pits and propagated during service and led to the fracture of the rst
stage blade. Subsequently, the broken pieces of the rst stage blade has caused further
damage (internal object damage) to other rst stage as well as the second and the third
stage turbine blades in the form of dents and nicks on leading/trailing edges by impact.
2014 Elsevier Ltd. All rights reserved.
1. Introduction
Turbine rotor blades are important components of aeroengines with complex geometry whose performance is directly
related to the reliability and life of engines [13]. Gas turbine blades are principally made out of nickel-base and cobalt-base
superalloys. Nickel-based superalloys are the most complex and the most widely used alloys for various gas turbine compo-
nents which operate under high temperature and high pressure conditions. A noteworthy feature of nickel-base alloys is that
they can be used for load-bearing applications at temperatures up to 0.8 T
m
, (T
m
is melting point in Kelvin), and where no
other class of engineering alloys can be used. Ni-based superalloys commonly used as turbine disc and blade material are
strengthened by precipitation of c
0
phase (Ni
3
(Al,Ti)). Mechanical properties, in particular creep properties, of these alloys
depend upon the volume fraction of c
0
as well as precipitation of carbides along the grain boundary. These alloys are prone
to microstructural degradation during service exposure, which affects the load bearing characteristics of the superalloy
adversely. The microstructural changes during high temperature exposure or service include coarsening or directional
growth of c
0
particles (known as rafting), formation of continuous carbide lm along grain boundaries and excess precipita-
tion of carbides within the grains.
http://dx.doi.org/10.1016/j.engfailanal.2014.06.003
1350-6307/ 2014 Elsevier Ltd. All rights reserved.

Corresponding author. Tel.: +91 040 2458 6545; fax: +91 040 2434 0683.
E-mail address: sweety24_k@yahoo.com (S. Kumari).
Engineering Failure Analysis 45 (2014) 234244
Contents lists available at ScienceDirect
Engineering Failure Analysis
j our nal homepage: www. el sevi er . com/ l ocat e/ engf ai l anal
The blades and other components of gas turbines suffer service induced damage which may be natural or accelerated. The
degradation may have metallurgical or mechanical origin. The factors responsible for damage in turbine blades are: (i) oper-
ating environment (high temperature, fuel and air contamination, solid particles, etc.), (ii) high mechanical stresses (due to
centrifugal force, vibratory and exural stresses) and (iii) high thermal stresses (due to thermal gradients). In general, blade
failures can be grouped into two categories (a) fatigue including both High Cycle Fatigue (HCF) and Low Cycle Fatigue (LCF)
[48] and (b) creep/stress rupture [9,10]. Further, these damages can be categorized: (i) external and internal surface damage
such as corrosion, oxidation, crack formation, erosion, foreign object damage (FOD), and fretting, and (ii) internal damage of
microstructure such as c
0
[Ni
3
(Al,Ti)] phase coarsening or rafting, grain growth, grain boundary creep voiding, carbide pre-
cipitation and brittle topologically close packed (TCP) phase formation. There are generally three possible damage mecha-
nisms affecting the life of turbine blades; mechanical damage through creep, multiaxial fatigue and high temperature
corrosion. Impact damage or foreign object damage (FOD) is a common failure mechanism in the compressor blades but
not so common in turbine blades. However, these can suffer impact by the lost fragments of other broken off parts of the
engine, known as internal object damage (IOD).
In this paper, the results of the detailed metallurgical investigation undertaken to identify the cause of the failure of tur-
bine blades of different stages of an aeroengine of a military transport aircraft, are presented.
2. Experimental procedure
A set of four blades from all the three stages of a gas turbine were examined. Detailed metallurgical analysis including
visual examination, stereo, optical and scanning electron microscopy, chemical analysis and hardness measurement, was
undertaken. The fracture as well as damaged portions of blades of different stages I, II and III were cut cleaned ultrasonically
and examined using stereo and scanning electron microscopy for recording fracture features. Also, samples were cut from
aerofoil region of blades of different stages, mounted in Bakelite, polished following standard metallographic practice and
etched using a chemical reagent for microstructural examination. Krolls agent was used for etching stage I and II blade sam-
ples, while 5% nital for (electrolytic etching) stage III blade samples.
3. Results
3.1. Chemical composition
Chemical analysis of one blade from each stage was carried out using instrumental analysis method. The concentration of
elements present in I, II and III stage blades are given in Tables 13, respectively.
3.2. Hardness
Bulk hardness measurements were made on the polished samples extracted from aerofoil region of I, II and III stage
blades using a universal hardness testing machine at a load of 30 kgf. The average hardness of I, II and III stage blades in
the aerofoil section were found to be 420, 380 and 275 HV, respectively.
3.3. Visual examination
During defect investigation after stripping of engine, one blade of turbine rotor of I stage was found broken (designated as
blade A) and other had a deep cut at the trailing edge (designated as blade B) as illustrated in Fig. 1. The blade A had frac-
tured at a height of about 4 cm from the blade root platform (Fig 1a). A piece of about 1 cm size was found chipped off at the
trailing edge of blade B (Fig. 1b). Also, several blades of II and III stages were found to have nicks and dents on their leading/
trailing edges as illustrated in Fig. 2. It can be seen that the extent of damage i.e., the depth of cut/dent/nick on edges of the
blades has progressively reduced from stage I through stage II to stage III.
3.4. Stereographic examination
Stereographic examination of the fracture surface of blade A revealed two distinct fracture regions as shown in Fig. 3.
Region I i.e., fracture surfaces near the leading (Fig. 3a) and the trailing edges (Fig. 3b), reveals beach marks a characteristic
of progressive fracture, while region II i.e., the region between leading and trailing edge has rough and dull appearance. Ste-
reo micrograph of the damaged region of stage I blade B reveal both rough and rubbed features as shown in Fig. 4.
Table 1
Chemical composition of I stage blade (wt%).
C Mo Nb Cr Al Ti W Co Ni
0.16 3.5 0.032 9.60 5.0 2.3 13.0 5.10 Balance
S. Kumari et al. / Engineering Failure Analysis 45 (2014) 234244 235
Stereo micrographs of damaged regions of II and III stage blades are shown in Figs. 5 and 6, respectively. The micrographs
in both the cases reveal severe plastic ow, ne surface cracks and rub marks in the dented/damaged regions.
3.5. Fractography
Fractographs shown in Fig. 7ad, reveal fracture features at different regions across the cross-section of the aerofoil of the
stage I blade A. Fig. 7a shows beach marks as well as the damage of surface coating at the tip of the leading edge which
could be the fracture initiating site. Further, it is observed that beach marks are present over most portions of the fracture
surface, except over small region near trailing edge (Fig. 7b) and the region away from the tip of the leading edge as marked
in Fig. 3a (region II). In both the regions overload fracture features are observed as shown in Fig. 7(c). Also, the entire fracture
surface is found to be covered with oxide scale (Fig. 7d). Further, fractograph shown in Fig. 8, reveals the presence of surface
coating of about 35 lm thickness, in some regions of blade A. The presence of Ni and Al in the coating as revealed by EDAX
analysis (Fig. 8b) suggests that the surface coating could be nickel aluminide. SEM fractographs given in Fig. 9, reveal
Table 2
Chemical composition of II stage blade (wt%).
C Mo Nb Cr Al Ti W Co Ni
0.046 9.10 0.010 9.20 4.75 0.016 4.95 4.95 Balance
Table 3
Chemical composition of III stage blade (wt%).
C Mo Nb Cr Al Ti W Co Ni
0.081 0.030 0.012 19.12 0.63 2.60 0.009 0.016 Balance
Fig. 1. Photographs of (a) fractured and (b) damaged I stage blades.
Fig. 2. Photographs of damaged blades (a) II stage and (b) III stage blades.
236 S. Kumari et al. / Engineering Failure Analysis 45 (2014) 234244
Fig. 3. Fracture surfaces of I stage blade A showing (a) leading and (b) trailing edges revealing beach marks.
Fig. 4. Features in the damaged region of I stage blade B (shown in Fig. 1(b)).
Fig. 5. Stereographs at (a) low magnication and (b) high magnication showing features of severe plastic ow and ne surface cracks in the damaged
region of II stage turbine blade.
Fig. 6. Stereographs at (a) low magnication and (b) high magnication showing features of severe plastic ow and rub marks in the damaged region of III
stage turbine blade.
S. Kumari et al. / Engineering Failure Analysis 45 (2014) 234244 237
dimples, which are characteristic features of overload fracture and, also, the presence of abrasion marks in the damaged
region of I stage blade B.
Secondary electron images of the damaged regions (dents/nicks), of II and III stage blades are shown in Figs. 10 and 11,
respectively. As can be seen from Fig. 10, the magnied view of the dented region of II stage blade reveals deep cut into the
blade and the presence of multiple surface cracks on the surface. The micrographs show abrasion marks (Fig. 11) in the dam-
aged area of III stage blade. These features indicate that the blades might have been impacted by some object forming dent
on one of the edges. Further, except for oxides, no foreign material was found in the damaged regions of both the blades.
3.6. Microstructural examination
Fig. 12 shows optical micrograph revealing dendritic structure, a characteristic of cast nickel based superalloys, in the top
portion of I stage blade. The grains are very coarse and essentially equi-axed. Similar observations were also made in the
middle and lower portions of the aerofoil section as well as root of the blade. SEM micrographs of I stage blade revealing
Fig. 7. SEM micrographs of I stage blade A revealing fracture features at different regions across the cross-section of aerofoil (a) leading edge region
showing crack initiation site and beach marks (b) overload fracture region at trailing edge (c) overload fracture features and (d) oxide layer on fracture
surface.
Fig. 8. SEM fractographs showing (a) NiAl coating at the surface and (b) EDAX spectrum of surface coating revealing the presence of Ni and Al.
238 S. Kumari et al. / Engineering Failure Analysis 45 (2014) 234244
the presence of carbides in c matrix is shown in Fig. 13a. The magnied view in Fig. 13b shows ne cuboidal as well as par-
tially rafted (i.e., directional coarsening) c
0
precipitates in c matrix. Fig. 14 shows EDAX spectra of matrix (Fig. 14a) and car-
bide precipitates (Fig. 14b) which indicate the presence of elements such as Ni, Cr, Co, W, Ti, Al and Mo in matrix, while Ti, W
and Mo in carbides. Further, secondary electron micrograph in Fig. 15a reveals the absence of surface coating, but evidence of
oxidation damage of the surface up to a depth of 200 lm was observed. EDAX analysis of the surface (Fig. 15b) reveals the
presence of aluminum, chromium and titanium oxides in this region.
A detailed microstructural examination of a sample extracted from the region just below the fracture surface of I stage
blade was carried out and micrographs are shown in Fig. 16. These micrographs reveal that severe localized oxidation attack
(i.e., internal oxidation) of substrate in the regions where surface coating has got damaged during service. The oxidation
Fig. 9. SEM fractographs showing (a) dimple features and (b) abrasion marks in the damaged region of I stage blade B (Fig. 1(b)).
Fig. 10. SEM images showing deep cuts and surface cracks in the dented region of II stage blade.
Fig. 11. SEM images showing abrasion marks in the dented region of III stage blade.
S. Kumari et al. / Engineering Failure Analysis 45 (2014) 234244 239
attack has led to local dislodgement of surface coating thereby formation of pits on the aerofoil surface at several locations as
illustrated in Fig. 16.
Optical micrographs of II and III stage blades as shown in Fig. 17a and b, reveal equiaxed grains of c phase (matrix), a
structure typical of wrought nickel based superalloys. Secondary electron images of II stage blade revealing the presence
of carbides along grain boundary as well as within the grains (Fig. 18a) and cuboidal c
0
precipitates (Fig. 18b) are shown
in Fig. 18. Fig. 19 shows the microstructure of III stage blade. It can be seen that the microstructure comprises carbides along
grain boundary as well as within grains.
4. Discussion
The Chemical composition as given in Tables 13 and microstructures illustrated in Figs. 12, 13 and 1719 suggest that
stage I, II and III blades are made out of nickel base superalloys of different grades. The composition and the microstructure
Fig. 12. Optical micrograph of airfoil region of I stage blade showing dendritic structure.
Fig. 13. Microstructures of I stage blade in the aerofoil region (a) low magnication image showing carbides in c matrix and (b) magnied view showing
cuboidal and partially rafted c
0
structure.
Fig. 14. EDAX spectrum of (a) matrix and (b) carbides.
240 S. Kumari et al. / Engineering Failure Analysis 45 (2014) 234244
of I stage blade indicate that the alloy is used in as-cast and heat treated condition, whereas the alloy in II and III stage blades
are used in wrought and heat treated condition. Further, the hardness of I stage blade was found to be the highest i.e.
420 HV for I stage blade, intermediate (380 HV) for II stage blade, while the lowest (275 HV) for III stage blade. This
Fig. 15. SEM micrograph showing (a) oxidation damage close to surface of I stage blade and (b) EDAX analysis showing oxides of Cr, Al and Ti.
Fig. 16. BSE images showing (a) absence of coating, surface oxidation as well as severe localized oxidation attack and formation of pits caused by oxidation
of I stage blade, (b) optical micrographs showing dislodgement of surface coating near leading edge and (c) magnied view of encircled region in (a).
Fig. 17. Optical micrograph showing equi-axed grain structure in (a) II stage and (b) III stage blades.
S. Kumari et al. / Engineering Failure Analysis 45 (2014) 234244 241
Fig. 18. SEM images showing (a) carbides along grain boundary and within the grains and (b) cuboidal c
0
in II stage blade.
Fig. 19. SEM image showing carbides along grain boundary and within grains in III stage blade.
Fig. 20. Schematic illustrating various stages involved in dislodgement of surface coating (a) damaged surface coating, (b) oxygen diffusion through
damaged coating during service and formation of discrete oxides at coating/substrate interface at several locations, (c) initiation of continuous oxide
Fig. layer formation at the interface and dislodgement of surface coating and formation of surface pits and (d) dislodgement of surface coating and
establishment of oxide layer on substrate surface.
242 S. Kumari et al. / Engineering Failure Analysis 45 (2014) 234244
is consistent with the amount of total alloying element content (in particular refractory elements like W and Mo) and volume
percentage of c
0
precipitate determined from (Al + Ti) content of superalloys used in various stages i.e., the total alloying ele-
ment content as well as (Al + Ti) content is the highest for the superalloy used in I stage, intermediate for the alloy used in II
stage and the lowest for the alloy used in III stage.
4.1. Failure of I stage blade
The existence of beach marks over most portion of the fracture surface (Figs. 3 and 7) as revealed by the detailed fracto-
graphic examination suggest that the fracture of I stage turbine blade A has occurred by fatigue. Both fractographic and
microstructural examination reveal the dislodgement of surface coating as well as formation of oxide layer and pits at sev-
eral locations on the surface of the aerofoil is shown in micrographs (Fig. 16). The oxidation damage illustrated in Fig. 16 may
be well understood in terms of the schematic shown in Fig. 20, which clearly elucidates the process of dislodgement of the
surface coating during service exposure. The damage of the protective coating during service has facilitated severe localized
oxidation attack of the substrate by inward diffusion of oxygen leading to formation of oxides at different locations in a dis-
crete manner as illustrated in Fig. 20(b). These discrete oxides grow at the coating/substrate interface with time and pro-
moted the dislodgement of the coating locally and formation of pits on the surface of the substrate (Fig. 20(c)). Over a
period of time a continuous oxide layer has formed at the interface thereby leading to complete dislodgement of the coating
at several locations as evident from Fig. 20(d). Thus, the formation of surface pits by oxidation could have facilitated the ini-
tiation of fatigue cracks at several locations and propagated, as typically shown in Fig. 16(c), during service and led to frac-
ture of I stage blade A. Further, different orientations of the beach marks observed suggest that there are multiple origins for
fatigue crack initiation. These origins are the locations on the aerofoil surface where severe localized oxidation attack has
occurred to form pits due to the absence of the protective coating (Fig. 16). Hence, the initiation of fatigue fracture of I stage
blade A can be attributed to either premature damage of the coating or inability to produce good coating on the surface of the
blades. It has been informed that the rotor turbine blades have failed after several hundred hours of service following over-
hauling. The usual practice during overhauling is that the blades are subjected to sand blasting to remove the damaged coat-
ing and they are given a new protective coating. However, the fact that the fatigue fracture of blades has been caused by
severe oxidation damage suggests that the damaged coating might not have been properly restored during overhauling
process.
4.2. Damage of I, II and III stage blades
A few of I stage blades as well as II and III stage blades were found damaged on one of the edges of the aerofoil. The dam-
age is in the form of deep cuts to dents/nicks (Figs. 10 and 11) and the severity of the damage is found to decrease from I
stage to III stage. Such damage (cuts/dents/nicks) is expected to take place by impact of an object. The detailed microstruc-
tural examination has revealed the presence of only oxides but no foreign material in the damaged region. This suggests that
the damage to all stage of blades is not caused by any foreign object, but, might have been caused by the impact of the bro-
ken piece of I stage blade.
5. Conclusions
The damage of surface coating has caused severe localized oxidation attack to I stage blade A, leading to the formation of
oxide at coating/substrate interface and pits as well as dislodgement of surface coating at several locations on the surface of
the aerofoil. These surface pits have facilitated the initiation of fatigue cracks during service and led to the fracture of several
blades.
The impact of broken pieces of I stage blade in turn might have caused damage i.e. internal object damage (IOD) to other
blades (such as B) of I stage, as well as II and III stage blades in the form of deep cuts, dents and nicks on either of the edges.
Acknowledgements
The authors would like to thank DRDO for nancial support and for granting permission to publish this paper and Direc-
tor, DMRL, for constant encouragement. The help of Mr. Amit Bala in extraction of specimens from the blades, Mrs. E. Vijaya-
laxmi in metallographic sample preparation and Sri Venkat in scanning electron microscopy, is gratefully acknowledged.
References
[1] Yao Siwei, Zhang Lixian, Li Jianjun. Analysis and prevention of turbine breakage in an engine starting motor. Mater Mech Eng 2002;26(2):402.
[2] Liu Delin, Zhang Weifang, Li Chunguang, et al. Fracture failure analysis of high-pressure turbine blades of II stage. Heat Treat Met 2007;32(z1):714.
[3] Tao Chunhu, Tang Xing, Xu Qiusha. Analysis and prevention of turbine breakage in an engine starting motor. Mater Mech Eng 2002;26(2):402.
[4] Walls DP, Daleneuville RE, Cunningham SE. Damage tolerance based life prediction in gas turbine engine blades under vibratory high cycle fatigue. J
Eng Gas Turbines Power 1997;119:1436.
[5] Burns J. Gas turbine engine blade life prediction for high cycle fatigue. The Technical Cooperation Program (TTCP), P-TP1; 1998.
[6] Conor PC. Compressor blade high cycle fatigue lifecase study. The Technical Cooperation Program (TTCP), P-TP1; 1998.
S. Kumari et al. / Engineering Failure Analysis 45 (2014) 234244 243
[7] Reddy TSR, et al. A review of recent aeroelastic analysis methods for propulsion at NASA Lewis research centre. NASA Technical Paper 3406; 1993.
[8] Rao JS. Natural frequencies of turbine blading a survey. Shock Vib Dig 1973;5(10):1.
[9] Persson C, Persson PO. Evaluation of service-induced damage and restoration of cast turbine blades. J Mater Eng Perform 1993;2(4):5659.
[10] Hou J, Wicks BJ, Stocks GJ, Slater SL, Antiniou RA. Creep failure assessment of a turbine disc using non-linear nite element method. In: IS-121, 24th
ISABE conference proceedings, Italy; 1999.
244 S. Kumari et al. / Engineering Failure Analysis 45 (2014) 234244