- Email: [email protected]
Cement kiln dust induced corrosion fatigue damage of gas turbine compressor blades – A failure analysis
Accepted Manuscript Cement kiln dust induced corrosion fatigue damage of gas turbine compressor blades- A failure analysis M.R. Jahangiri, A.A. Fallah, A. Ghiasipour PII: DOI: Reference:
S0261-3069(14)00400-2 http://dx.doi.org/10.1016/j.matdes.2014.05.031 JMAD 6509
To appear in:
Materials and Design
Received Date: Accepted Date:
11 December 2013 19 May 2014
Please cite this article as: Jahangiri, M.R., Fallah, A.A., Ghiasipour, A., Cement kiln dust induced corrosion fatigue damage of gas turbine compressor blades- A failure analysis, Materials and Design (2014), doi: http://dx.doi.org/ 10.1016/j.matdes.2014.05.031
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Cement kiln dust induced corrosion fatigue damage of gas turbine compressor blades- A failure analysis M.R. Jahangiri1, A.A. Fallah1, A. Ghiasipour2 1- Metallurgy Department, Niroo Research Institute, Tehran 14686, Iran Tel: +98-21-88079401-6, Fax : +98-21-88078296 E-mail address: [email protected] 2- Rey Power Plant, Rey City, Iran
Abstract Compressor of one of the gas turbines installed in a power plant was stopped under emergency conditions. Primary investigation showed that almost all of the first stage blades and some of the next stages were severely damaged. In this study, one of the first stage broken blades was failure analyzed. The results showed that the corrosion pits were formed on the compressor blade surface due to the presence of Cl and S elements in the compressor inlet air. Since the power plant located in the vicinity of a cement company and also an oil refining company, the inlet air of compressor had large amounts of Cl and S containing compounds. The corrosion pits acted as stress concentration sites, and facilitated fatigue crack initiation and propagation, leading to final fracture of the blades. Key words: Compressor blade; Corrosion fatigue; Cement kiln dust; Oil refinery
1- Introduction The required air for combustion and cooling applications in gas turbines is provided by the compressors. After passing through the filters, air is entered into the compressor and passes through different stages, as it becomes denser and hotter [1].
1
Dimensions of the compressor and aerodynamic shape of moving or stationary blades varies on the basis of gas turbine type and its manufacturer. In this study, failure of the first stage compressor rotor blades of a 30 MW gas turbine was investigated. The compressor had 18 blade rows. Based on the power plant documents, the blades had been made by a reputable company and they had worked more than 100 thousand hours before fracture. The blades were regularly washed during the periodic inspections (intervals around 30 thousand hours of operation). A survey of the compressor manufacturer documents showed that the material of the blades must be the martensitic stainless steel grade AISI 403, and no coating was applied on the surfaces of blades. Stainless steel grade AISI 403 is a modified turbine quality version of the well-known AISI 410 steel. AISI 403 is now widely used for the manufacturing of compressor blades of gas turbines, as well as steam turbine blades [2-3]. The main mechanisms of the failure of the compressor blades are low cycle and high cycle fatigue, corrosion, erosion, impact of foreign objects or FOD, and surge-induced heavy vibrations/deflections or sudden failure of blades [4-7]. In many cases, multiple mechanisms act simultaneously, resulting in the complete fracture of the component. At all events, precise analysis of the failure of the compressor blades is necessary to prevent future fractures. In this study, failure of the first stage rotor blades of a compressor of a 30 MW gas turbine was investigated, and the prevention methods were proposed for similar cases.
2- Experimental procedures To determine the failure mechanisms of the blades, following experiments were conducted:
2
2-1. Visual examination All rows of the compressor blades and vanes were visually examined to determine the type and degree of damage for each group. Also, the broken blade surfaces were inspected for the presence of corrosion products, foreign object and erosion damages.
2-2. Chemical analysis of blade material and some dusts To determine the chemical composition of the compressor blade alloy, after preparation of the suitable sample of the blade, chemical composition was determined by emission spectroscopy. Also, the chemical analyses of some cement kilns dusts were determined by X-ray fluorescence (XRF) and analytical chemical techniques.
2-3. Metallography To investigate the microstructures of the blades and the possible microstructural changes in different locations of the broken blades, various samples were prepared by grinding and polishing using standard metallographic methods and etched with Marble's reagent (10 g CuSO4, 50 mL HCl, 50 mL water). The etching time was approximately 5 seconds. Then, these metallographic specimens were examined by optical microscopy using a Leitz microscope model Aristomet, as well as by scanning electron microscopy (SEM) using a Tescan-Vega SEM. The SEM microscope equipped with an EDS detector for chemical analysis.
2-4. Fracture surface analysis Macroscopic and microscopic examinations of the fracture surfaces are needed to determine the failure mechanisms of the blades. For this purpose, the suitable fracture surfaces of the broken blade were selected and investigated. The fracture surfaces
3
were evaluated after ultrasonic cleaning, using a stereo microscopy, as well as a scanning electron microscopy (model Tescan-Vega).
2-5. Hardness measurements To evaluate the mechanical properties, Vickers hardness of the broken blade was measured using an Eseway hardness tester model DV RB-M under a load of 30 kg on the airfoil and root sections of the blade.
3. Results and discussion 3-1. Visual investigations The investigated compressor had 18 rows of rotor blades, 18 rows of stationary blades (vanes), and one row of stationary guide vanes. A great number of first stage rotor blades and some of the blades and vanes of other stages were damaged. The first stage stationary blades (vanes) had not suffered significant damage. This can be a sign that the surge phenomenon has not been occurred [8-9]. The first stage rotating blades were suffered more damage, and some of them were completely broken in the middle. Bottom section of a broken blade was studied in this study (Fig. 1). As can be seen in Fig. 1, the fracture surface makes an angle of 45° with respect to the longitudinal blade axis. There was a black greasy deposit on the surfaces of the blade. After cleaning the blade surfaces, small traces of erosion were observed on the blade leading edge.
3-2. Chemical composition
4
Table 1 presents the results of the chemical composition of the blades measured by emission spectroscopy. Also, the chemical composition of standard AISI 403 stainless steel in accordance with ASTM: A982 and ASTM: A1028 is given in this table. The comparative study of these results shows that the compressor blade alloy has chemical composition similar to that of AISI 403 stainless steel, but the amount of molybdenum in the composition of the blade is higher than that of the standard AISI 403 Steel. Molybdenum can improve the impact resistance and pitting corrosion resistance of these steels [10-11]. A literature survey shows that despite the relatively good corrosion resistance of AISI 403 steel in most environments, it have no satisfactory corrosion resistance in wet environments containing Cl and S ions. So, it is prone to pitting corrosion and intergranular corrosion in these environments [12-13].
3-3. Fracture surface analysis Figs. 2 and 3 show stereo and scanning electron microscopy images of the fracture surface of the blade. The macroscopic signs of the progressive growth of cracks (curved beach marks), indicating the occurrence of fatigue in the blades, can be seen on the fracture surfaces. The fracture surface resembles a quasi-cleavage fracture, and the river patterns are observed to some extent in these figures. The quasi-brittle fracture associated with the river patterns likely occurred due to the relatively low ductility of tempered martensite microstructures. These river patterns demonstrate that fatigue cracks have progressed on different planes during the growth stage. These patterns show the fast growing stage of the fatigue crack growth stages. The direction of the river patterns represents the direction of the crack propagation [14-15]. The origin of the fatigue cracks was identified by backward following the beach marks and river patterns. Such procedure is shown in Fig. 4-a to 4-d. As can be seen,
5
the origins of the cracks initiation are on the surface of the blade and are located on the surface pits. Chemical analysis by EDS showed that the Cl and S elements have been concentrated in these regions (Figs. 4-e and4-f). On some areas of the fracture surface (region A in Fig. 3-b), brittle striations and microcracks formed in parallel with these striations are clearly visible (Fig. 5). It should be mentioned that these striations were mainly parallel to the beach marks and perpendicular to the river patterns. Although, further studies are needed to identify the main causes of formation of numerous microcracks in the fracture surface parallel to striations, these microcracks may be formed due to the relatively brittle microstructure (tempered martensite) of the alloy and/or because of the presence of corrosive agents in the compressor environment.
3-4. Microstructural investigations Fig. 6 shows the optical microstructures of the blade samples. As can be seen, the microstructure consists of the tempered martensite, and the delta ferrite content in different regions of the microstructure is less than 2% (Fig. 6-c). The low delta ferrite content of the AISI 403 steel used in the manufacturing of the compressor blades is very important. The amount of delta ferrite should not exceed 2-5 percent [8-9, 16], because the tensile strength, resistance to corrosion, and the impact strength of the alloy would be significantly reduced [17-18]. Metallographic studies confirm that there was no coating on the blade surfaces. The blade surfaces showed severe localized corrosion (pitting). The depth of these pits was around 50-150 μm (Figs. 6-a and 6-b). In order to accurately evaluate the microstructures, various samples prepared from the leading and trailing edges, and also from the central portion of the airfoil, were
6
studied by SEM microscopy. Fig. 7 shows the results. Numerous pits and localized corrosion products are observed around the leading edge, as well as the suction or pressure sides of the blade surfaces. The depth of corrosion pits in some regions was approximately 150 μm. The corrosion products in the pits were analyzed by EDS method. As can be seen (Figs. 7 and 8), in addition to the alloying elements oxides, the presence of other elements such as Cl and S is considerable in these regions.
3-5. Identification of Cl and S sources To identify the Cl and S entrance sources and sites into the investigated compressor, the operational history and the service conditions of the compressor were further considered. Fig. 9 shows an average dust concentration of air entered to gas turbines for different power plants in Iran. These data are for the summer season of year 2001. As can bee seen, the investigated power plant in this study, the rey power plant, has the maximum dust concentration. A study of the geographic location of the power plant showed that this power generation plant has been located in the vicinity of two large companies: a cement company and an oil refinery. The exhaust gases exit the cement kiln, carrying along some of kiln dust. This dust usually contains significant amounts of chloride, sulfate, and alkali [19-20]. The chemical analysis range of the collected dusts from the investigated cement company kilns is given in Table 2. As can be seen, a large amount of Cl containing compounds is present in these dusts. On the other hand, the investigated power plant is located near an oil refining company. The refinery exhaust gases, as well as the power plant and cement company exhaust gases, contain a lot of S containing compounds. These S compounds can enter
7
the compressor through the inlet air. These Cl and S containing compounds act as the main causes for formation of corrosion pits on the blade surfaces.
3-6. Influence of Cl containing environments on fatigue properties of AISI 403 stainless steel According to the microscopy studies, pitting corrosion is demonstrated on the blade surfaces. Corrosion pits were formed due to the presence of Cl and S elements enriched on the compressor blade surfaces. The pits having large depths can provide initiation sites for primary cracks on the surface. These primary cracks can further grow due to the subsequent cyclic loads, leading to final failure of the component. Fig. 10 shows the effect of the presence of Cl containing compounds on fatigue properties of AISI 403 stainless steels [21-22]. As can be seen, the decrease of fatigue strength of the alloy due to 3% NaCl at 2×107 cycles is about 75%. A similar reduction in fatigue strength can be predicted as a result of presence of 3×10-2 % NaCl at 109 cycles. Schonbauer et al [23] have shown that the fatigue strength of AISI 403 stainless steel decreases significantly with increasing the pit size in aerated 6 ppm Clsolutions. Table 3 shows their results. From these results it can be concluded that the fatigue strength of AISI 403 stainless steel used in the manufacturing of compressor blade decreases even if very small amounts of Cl are present at local points of the blade surface.
3-7. Hardness results Hardness was measured in different regions of the blade (Fig. 11), and it was found that the hardness values of the blade are close to each other and about 239-246 Vickers. A comparison of measured hardness values with those reported for AISI 403
8
stainless steel after different heat treatments (Table 4) shows that the alloy was used in the quenched and tempered conditions, and its mechanical properties were in the standard range. Additionally, the mechanical properties of the alloy had undergone no changes during service. So, the alloy for the manufacturing of the blades had suitable conditions.
4. Conclusions According to the experiments conducted in this study, the following conclusions are drawn: 1- Corrosion pits are formed on the blades surfaces due to the vicinity of the power plant to a cement company and an oil refinery. 2- Cl and S containing compounds acted as the main causes for formation of corrosion pits on the blade surfaces. 3- Corrosion pits acted as stress concentration sites for nucleation of cracks. 4- Beach marks, striations, transgranular fracture, and river patterns were observed on the fracture surfaces. 5- After initiation of primary cracks, these cracks were grown by fatigue mechanism through the dynamic stresses applied on compressor blades, leading to final fracture of the blades. 6- The first stage stationary blades (vanes) had not damaged. This shows that the surge phenomenon has not been occurred.
To prevent similar incidents, the followings are recommended: 1- Prevention of the entrance of Cl and S containing compounds into the compressor, as far as possible.
9
2- The use of corrosion and erosion resistant coatings on the blades surfaces. These coatings can be of types such as nickel-cadmium coatings or aluminum-ceramic coatings. 3- The use of more resistant alloys against the corrosion in wet environments containing the Cl and S compounds. 4- Complete, accurate and non-destructive inspection of blades which have a long service life, to ensure the absence of surface cracks. 5- Selection and destructive inspection of one the blades’ having long lifetimes, to ensure the absence of pits and cracks due to the corrosion.
References [1] P. Hanlon, ‘Compressor Handbook’, McGraw-Hill Professional; 1 ed., January 2001. [2] P. Dowson, D. Bauer, S. Laney, ‘Selection of materials and material related processes for centrifugal compressors and steam turbines in the oil and petrochemical industry’, 37th Turbomachinery Symposium, Texas A&M University, 2008. [3] B. Mahmoudi, M.J. Torkamani, A.R. Sabour Aghdam, J. Sabbaghzadeh, ‘Effect of laser surface hardening on the hydrogen embrittlement of AISI 420: Martensitic stainless steel’, Materials & Design, (2011), 32(5): 2621-2627. [4] R.D.C. Passey, ‘Reliability of compressor aerofoils’, Progress in Aerospace Sciences, (1976), 17: 67-92. [5] Z. Huda, P. Edi, ‘Materials selection in design of structures and engines of supersonic aircrafts: A review’, Materials & Design, (2013), 46: 552-560. [6] L.M. Amoo, ‘On the design and structural analysis of jet engine fan blade structures’, Progress in Aerospace Sciences, (2013), 60: 1-11. [7] Q. Li, J. Piechna, N. Mueller, ‘Simulation of fatigue failure in composite axial compressor blades’, Materials & Design, (2011), 32(4): 2058-2065. [8] M.P. Boyce, ‘Gas turbine engineering handbook’, Butterworth-Heinemann, 4 ed., 2011.
10
[9] R.E. Dundas, ‘Engineering and metallographic aspects of gas turbine engine failure investigation: identifying the causes’, Aviation Mechanics Bulletin, (1994) 42: 1-11. [10] H. Amaya, T. Mori, K. Kondo, H. Hirata, M. Ueda, ‘Effect of chromium and molybedenum on corrosion resistance of super 13Cr martenistic stainless steel in CO2 environment’, CORROSION 98, March 22 - 27, 1998, San Diego Ca. [11] R. Guillou, M. Guttmann, Ph. Dumoulin, ‘Role of molybdenum in phosphorusinduced temper embrittlement of 12%Cr martensitic stainless steel’, Metal Science, (1981) 15(2): 63-72. [12] I. Thorbjornsson, ‘Corrosion fatigue testing of eight different steels in an Icelandic geothermal environment’, Materials & Design, (1995), 16(2): 97-102. [13] X. Wang, J. Xu, C. Sun, ‘Effects of sulfate-reducing bacterial on corrosion of 403 stainless steel in soils containing chloride ions’, International Journal of Electrochemical Science, (2013), 8: 821-830. [14] Metals Handbook, Vol. 12: Fractography, ASM International, 9 ed., 1987. [15] B. Li, Y. Shen, W. Hu, ‘Casting defects induced fatigue damage in aircraft frames of ZL205A aluminum alloy–A failure analysis’, Material & Design, (2011), 32(5): 2570-2582. [16] AMS5611F, ‘Steel, Corrosion and Heat-Resistant, Bars, Wire, Forgings, Tubing, and Rings, 12Cr, Ferrite Controlled, Consumable Electrode Melted’, SAE International, 2006. [17] X.P. Ma, L.J. Wang, B. Qin, C.M. Liu, S.V. Subramanian, ‘Effect of N on microstructure and mechanical properties of 16Cr5Ni1Mo martensitic stainless steel’, Materials & Design, (2012), 34: 74-81. [18] G. Alkan, D. Chae, S.J. Kim, ‘Effect of δ ferrite on impact property of hot-rolled 12Cr–Ni steel’, Materials Science and Engineering A: (2013), 585: 39-46. [19] J.N. Asaad, S.Y. Tawfik, ‘Polymeric composites based on polystyrene and cement dust wastes’, Materials & Design, (2011), 32(10): 5113-5119. [20] P. Chaunsali, S. Peethamparan, ‘Influence of the composition of cement kiln dust on its interaction with fly ash and slag’, Cement and Concrete Research, (2013), 54: 106-113. [21] R. Ebara, ‘Long-term corrosion fatigue behaviour of structural materials’, Fatigue & Fracture of Engineering Materials & Structures, (2002), 25: 855–859.
11
[22] R. Ebara, ‘The present situation and future problems in ultrasonic fatigue testingMainly reviewed on environmental effects and materials’ screening’, International Journal of Fatigue, (2006), 28(11): 1465-1470. [23] B.M. Schonbauer, A. Perlega, S.E. Stanzl-Tschegg, ‘Pit-to-crack transition and corrosion fatigue of 12% Cr steam turbine blade steel’, 13th International Conference on Fracture, June 16–21, 2013, Beijing, China. [24] Metals Handbook, Vol. 4: Heat Treating, ASM International, 10 ed., 1991. [25] J.R. Davis, Stainless Steels (ASM Specialty Handbook), ASM International, 1994.
12
Table 1- Chemical composition of the compressor blade alloy and standard AISI 403 stainless steel (Wt%) Element
Fe
C
Cr
Si
Mn
P
S
Ni
Mo
Compress
Bas
0.115
12.9
0.175
0.42
0.015
0.013
0.1
0.54
or Blade
e
4
4
AISI 403
Bas
0.15ma
11.5- 0.5ma
1ma
0.04ma
0.03ma
-
-
e
x
13
x
x
x
5
x
Table 2- Chemical composition range of cement kilns dusts collected from the cement manufacturing company in the vicinity of the power plant Constituent
SiO2
Al2O3
Fe2O3
CaO
MgO
SO3
K2O
Na2O
Cl
-
L.O.I
Percent
11.24-
6.17-
1.93-
46.01-
0.48-
1.68-
2.02-
0.84-
2.81-
4.25-
16.41
7.36
2.62
56.44
2.58
8.7
3.21
2.65
6.93
8.26
Table 3- Fatigue strength (Δσ/2) of AISI 403 stainless steel determined in aerated 6 ppm Cl- solution at 90°C (in MPa). R = σmin/σmax [23] Pit Size
R= -1
R= 0.05
R= 0.5
R= 0.8
Smooth 100 µm 250 µm
300 160
265 175
165 110
72.5 65 52.5
Table 4- Hardness of AISI 403 steel after selected heat treatments [24-25] Heat Treatment Type
Full Annealing Quenching and Tempering
Tempering
Vickers Hardness Number
Temperature (°C)
(HV)
205-370 565-605
137-170 372-471 266-310
13
600-650
230-272
Figures Captions: Fig.1. (a-b) Images of the broken blades from different views Fig.2. (a-b) Stereo microscopy images of different regions of the broken blade fracture surface Fig.3. SEM images of the fracture surface of the blade, (a-b) low magnification, (c-d) medium magnification Fig.4. (a-d) SEM image of the nucleation site for fatigue cracks showing the presence of corrosion pits in these regions, (e) secondary and back-scattered SEM images of surface corrosion pits, (f) EDS analysis of corrosion products in pits Fig.5. High magnification SEM images of the fatigue crack growth regions in fracture surface of the broken blade indicating the brittle striations and also the microcracks Fig.6. (a-b) Optical microscopy images of the metallographic samples before etching, (c) microstructure of the alloy after etching Fig.7. (a-c) SEM images of the blade microstructure before etching at different magnifications, (d) EDS analysis of region A in Fig. 7-c Fig.8. (a) SEM image of a corrosion pit on the surface of the broken blade, (b) EDS analysis of region C in Fig. 8-a Fig.9. A comparison of mean dust concentration in air entered to gas turbines for different power plants in Iran Fig.10. Effect of different contents of Cl containing compounds on fatigue properties of AISI 403 stainless steels [21-22]. Rotating bending test, 60 Hz, R (σmin/σmax) = -1 Fig.11. Vickers hardness values in different regions of the broken blade
14
15
16
17
18
19
20
21
22
23
24
25
Highlights: - One of the first stage broken blades of a compressor was failure analyzed - The power plant located near a cement company and an oil refining company - Hence, the inlet air of compressor had large amounts of Cl and S compounds
- Therefore, the corrosion pits were formed on the compressor blade surface - Corrosion pits acted as stress concentration sites for fatigue fracture
26
S0261-3069(14)00400-2 http://dx.doi.org/10.1016/j.matdes.2014.05.031 JMAD 6509
To appear in:
Materials and Design
Received Date: Accepted Date:
11 December 2013 19 May 2014
Please cite this article as: Jahangiri, M.R., Fallah, A.A., Ghiasipour, A., Cement kiln dust induced corrosion fatigue damage of gas turbine compressor blades- A failure analysis, Materials and Design (2014), doi: http://dx.doi.org/ 10.1016/j.matdes.2014.05.031
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Cement kiln dust induced corrosion fatigue damage of gas turbine compressor blades- A failure analysis M.R. Jahangiri1, A.A. Fallah1, A. Ghiasipour2 1- Metallurgy Department, Niroo Research Institute, Tehran 14686, Iran Tel: +98-21-88079401-6, Fax : +98-21-88078296 E-mail address: [email protected] 2- Rey Power Plant, Rey City, Iran
Abstract Compressor of one of the gas turbines installed in a power plant was stopped under emergency conditions. Primary investigation showed that almost all of the first stage blades and some of the next stages were severely damaged. In this study, one of the first stage broken blades was failure analyzed. The results showed that the corrosion pits were formed on the compressor blade surface due to the presence of Cl and S elements in the compressor inlet air. Since the power plant located in the vicinity of a cement company and also an oil refining company, the inlet air of compressor had large amounts of Cl and S containing compounds. The corrosion pits acted as stress concentration sites, and facilitated fatigue crack initiation and propagation, leading to final fracture of the blades. Key words: Compressor blade; Corrosion fatigue; Cement kiln dust; Oil refinery
1- Introduction The required air for combustion and cooling applications in gas turbines is provided by the compressors. After passing through the filters, air is entered into the compressor and passes through different stages, as it becomes denser and hotter [1].
1
Dimensions of the compressor and aerodynamic shape of moving or stationary blades varies on the basis of gas turbine type and its manufacturer. In this study, failure of the first stage compressor rotor blades of a 30 MW gas turbine was investigated. The compressor had 18 blade rows. Based on the power plant documents, the blades had been made by a reputable company and they had worked more than 100 thousand hours before fracture. The blades were regularly washed during the periodic inspections (intervals around 30 thousand hours of operation). A survey of the compressor manufacturer documents showed that the material of the blades must be the martensitic stainless steel grade AISI 403, and no coating was applied on the surfaces of blades. Stainless steel grade AISI 403 is a modified turbine quality version of the well-known AISI 410 steel. AISI 403 is now widely used for the manufacturing of compressor blades of gas turbines, as well as steam turbine blades [2-3]. The main mechanisms of the failure of the compressor blades are low cycle and high cycle fatigue, corrosion, erosion, impact of foreign objects or FOD, and surge-induced heavy vibrations/deflections or sudden failure of blades [4-7]. In many cases, multiple mechanisms act simultaneously, resulting in the complete fracture of the component. At all events, precise analysis of the failure of the compressor blades is necessary to prevent future fractures. In this study, failure of the first stage rotor blades of a compressor of a 30 MW gas turbine was investigated, and the prevention methods were proposed for similar cases.
2- Experimental procedures To determine the failure mechanisms of the blades, following experiments were conducted:
2
2-1. Visual examination All rows of the compressor blades and vanes were visually examined to determine the type and degree of damage for each group. Also, the broken blade surfaces were inspected for the presence of corrosion products, foreign object and erosion damages.
2-2. Chemical analysis of blade material and some dusts To determine the chemical composition of the compressor blade alloy, after preparation of the suitable sample of the blade, chemical composition was determined by emission spectroscopy. Also, the chemical analyses of some cement kilns dusts were determined by X-ray fluorescence (XRF) and analytical chemical techniques.
2-3. Metallography To investigate the microstructures of the blades and the possible microstructural changes in different locations of the broken blades, various samples were prepared by grinding and polishing using standard metallographic methods and etched with Marble's reagent (10 g CuSO4, 50 mL HCl, 50 mL water). The etching time was approximately 5 seconds. Then, these metallographic specimens were examined by optical microscopy using a Leitz microscope model Aristomet, as well as by scanning electron microscopy (SEM) using a Tescan-Vega SEM. The SEM microscope equipped with an EDS detector for chemical analysis.
2-4. Fracture surface analysis Macroscopic and microscopic examinations of the fracture surfaces are needed to determine the failure mechanisms of the blades. For this purpose, the suitable fracture surfaces of the broken blade were selected and investigated. The fracture surfaces
3
were evaluated after ultrasonic cleaning, using a stereo microscopy, as well as a scanning electron microscopy (model Tescan-Vega).
2-5. Hardness measurements To evaluate the mechanical properties, Vickers hardness of the broken blade was measured using an Eseway hardness tester model DV RB-M under a load of 30 kg on the airfoil and root sections of the blade.
3. Results and discussion 3-1. Visual investigations The investigated compressor had 18 rows of rotor blades, 18 rows of stationary blades (vanes), and one row of stationary guide vanes. A great number of first stage rotor blades and some of the blades and vanes of other stages were damaged. The first stage stationary blades (vanes) had not suffered significant damage. This can be a sign that the surge phenomenon has not been occurred [8-9]. The first stage rotating blades were suffered more damage, and some of them were completely broken in the middle. Bottom section of a broken blade was studied in this study (Fig. 1). As can be seen in Fig. 1, the fracture surface makes an angle of 45° with respect to the longitudinal blade axis. There was a black greasy deposit on the surfaces of the blade. After cleaning the blade surfaces, small traces of erosion were observed on the blade leading edge.
3-2. Chemical composition
4
Table 1 presents the results of the chemical composition of the blades measured by emission spectroscopy. Also, the chemical composition of standard AISI 403 stainless steel in accordance with ASTM: A982 and ASTM: A1028 is given in this table. The comparative study of these results shows that the compressor blade alloy has chemical composition similar to that of AISI 403 stainless steel, but the amount of molybdenum in the composition of the blade is higher than that of the standard AISI 403 Steel. Molybdenum can improve the impact resistance and pitting corrosion resistance of these steels [10-11]. A literature survey shows that despite the relatively good corrosion resistance of AISI 403 steel in most environments, it have no satisfactory corrosion resistance in wet environments containing Cl and S ions. So, it is prone to pitting corrosion and intergranular corrosion in these environments [12-13].
3-3. Fracture surface analysis Figs. 2 and 3 show stereo and scanning electron microscopy images of the fracture surface of the blade. The macroscopic signs of the progressive growth of cracks (curved beach marks), indicating the occurrence of fatigue in the blades, can be seen on the fracture surfaces. The fracture surface resembles a quasi-cleavage fracture, and the river patterns are observed to some extent in these figures. The quasi-brittle fracture associated with the river patterns likely occurred due to the relatively low ductility of tempered martensite microstructures. These river patterns demonstrate that fatigue cracks have progressed on different planes during the growth stage. These patterns show the fast growing stage of the fatigue crack growth stages. The direction of the river patterns represents the direction of the crack propagation [14-15]. The origin of the fatigue cracks was identified by backward following the beach marks and river patterns. Such procedure is shown in Fig. 4-a to 4-d. As can be seen,
5
the origins of the cracks initiation are on the surface of the blade and are located on the surface pits. Chemical analysis by EDS showed that the Cl and S elements have been concentrated in these regions (Figs. 4-e and4-f). On some areas of the fracture surface (region A in Fig. 3-b), brittle striations and microcracks formed in parallel with these striations are clearly visible (Fig. 5). It should be mentioned that these striations were mainly parallel to the beach marks and perpendicular to the river patterns. Although, further studies are needed to identify the main causes of formation of numerous microcracks in the fracture surface parallel to striations, these microcracks may be formed due to the relatively brittle microstructure (tempered martensite) of the alloy and/or because of the presence of corrosive agents in the compressor environment.
3-4. Microstructural investigations Fig. 6 shows the optical microstructures of the blade samples. As can be seen, the microstructure consists of the tempered martensite, and the delta ferrite content in different regions of the microstructure is less than 2% (Fig. 6-c). The low delta ferrite content of the AISI 403 steel used in the manufacturing of the compressor blades is very important. The amount of delta ferrite should not exceed 2-5 percent [8-9, 16], because the tensile strength, resistance to corrosion, and the impact strength of the alloy would be significantly reduced [17-18]. Metallographic studies confirm that there was no coating on the blade surfaces. The blade surfaces showed severe localized corrosion (pitting). The depth of these pits was around 50-150 μm (Figs. 6-a and 6-b). In order to accurately evaluate the microstructures, various samples prepared from the leading and trailing edges, and also from the central portion of the airfoil, were
6
studied by SEM microscopy. Fig. 7 shows the results. Numerous pits and localized corrosion products are observed around the leading edge, as well as the suction or pressure sides of the blade surfaces. The depth of corrosion pits in some regions was approximately 150 μm. The corrosion products in the pits were analyzed by EDS method. As can be seen (Figs. 7 and 8), in addition to the alloying elements oxides, the presence of other elements such as Cl and S is considerable in these regions.
3-5. Identification of Cl and S sources To identify the Cl and S entrance sources and sites into the investigated compressor, the operational history and the service conditions of the compressor were further considered. Fig. 9 shows an average dust concentration of air entered to gas turbines for different power plants in Iran. These data are for the summer season of year 2001. As can bee seen, the investigated power plant in this study, the rey power plant, has the maximum dust concentration. A study of the geographic location of the power plant showed that this power generation plant has been located in the vicinity of two large companies: a cement company and an oil refinery. The exhaust gases exit the cement kiln, carrying along some of kiln dust. This dust usually contains significant amounts of chloride, sulfate, and alkali [19-20]. The chemical analysis range of the collected dusts from the investigated cement company kilns is given in Table 2. As can be seen, a large amount of Cl containing compounds is present in these dusts. On the other hand, the investigated power plant is located near an oil refining company. The refinery exhaust gases, as well as the power plant and cement company exhaust gases, contain a lot of S containing compounds. These S compounds can enter
7
the compressor through the inlet air. These Cl and S containing compounds act as the main causes for formation of corrosion pits on the blade surfaces.
3-6. Influence of Cl containing environments on fatigue properties of AISI 403 stainless steel According to the microscopy studies, pitting corrosion is demonstrated on the blade surfaces. Corrosion pits were formed due to the presence of Cl and S elements enriched on the compressor blade surfaces. The pits having large depths can provide initiation sites for primary cracks on the surface. These primary cracks can further grow due to the subsequent cyclic loads, leading to final failure of the component. Fig. 10 shows the effect of the presence of Cl containing compounds on fatigue properties of AISI 403 stainless steels [21-22]. As can be seen, the decrease of fatigue strength of the alloy due to 3% NaCl at 2×107 cycles is about 75%. A similar reduction in fatigue strength can be predicted as a result of presence of 3×10-2 % NaCl at 109 cycles. Schonbauer et al [23] have shown that the fatigue strength of AISI 403 stainless steel decreases significantly with increasing the pit size in aerated 6 ppm Clsolutions. Table 3 shows their results. From these results it can be concluded that the fatigue strength of AISI 403 stainless steel used in the manufacturing of compressor blade decreases even if very small amounts of Cl are present at local points of the blade surface.
3-7. Hardness results Hardness was measured in different regions of the blade (Fig. 11), and it was found that the hardness values of the blade are close to each other and about 239-246 Vickers. A comparison of measured hardness values with those reported for AISI 403
8
stainless steel after different heat treatments (Table 4) shows that the alloy was used in the quenched and tempered conditions, and its mechanical properties were in the standard range. Additionally, the mechanical properties of the alloy had undergone no changes during service. So, the alloy for the manufacturing of the blades had suitable conditions.
4. Conclusions According to the experiments conducted in this study, the following conclusions are drawn: 1- Corrosion pits are formed on the blades surfaces due to the vicinity of the power plant to a cement company and an oil refinery. 2- Cl and S containing compounds acted as the main causes for formation of corrosion pits on the blade surfaces. 3- Corrosion pits acted as stress concentration sites for nucleation of cracks. 4- Beach marks, striations, transgranular fracture, and river patterns were observed on the fracture surfaces. 5- After initiation of primary cracks, these cracks were grown by fatigue mechanism through the dynamic stresses applied on compressor blades, leading to final fracture of the blades. 6- The first stage stationary blades (vanes) had not damaged. This shows that the surge phenomenon has not been occurred.
To prevent similar incidents, the followings are recommended: 1- Prevention of the entrance of Cl and S containing compounds into the compressor, as far as possible.
9
2- The use of corrosion and erosion resistant coatings on the blades surfaces. These coatings can be of types such as nickel-cadmium coatings or aluminum-ceramic coatings. 3- The use of more resistant alloys against the corrosion in wet environments containing the Cl and S compounds. 4- Complete, accurate and non-destructive inspection of blades which have a long service life, to ensure the absence of surface cracks. 5- Selection and destructive inspection of one the blades’ having long lifetimes, to ensure the absence of pits and cracks due to the corrosion.
References [1] P. Hanlon, ‘Compressor Handbook’, McGraw-Hill Professional; 1 ed., January 2001. [2] P. Dowson, D. Bauer, S. Laney, ‘Selection of materials and material related processes for centrifugal compressors and steam turbines in the oil and petrochemical industry’, 37th Turbomachinery Symposium, Texas A&M University, 2008. [3] B. Mahmoudi, M.J. Torkamani, A.R. Sabour Aghdam, J. Sabbaghzadeh, ‘Effect of laser surface hardening on the hydrogen embrittlement of AISI 420: Martensitic stainless steel’, Materials & Design, (2011), 32(5): 2621-2627. [4] R.D.C. Passey, ‘Reliability of compressor aerofoils’, Progress in Aerospace Sciences, (1976), 17: 67-92. [5] Z. Huda, P. Edi, ‘Materials selection in design of structures and engines of supersonic aircrafts: A review’, Materials & Design, (2013), 46: 552-560. [6] L.M. Amoo, ‘On the design and structural analysis of jet engine fan blade structures’, Progress in Aerospace Sciences, (2013), 60: 1-11. [7] Q. Li, J. Piechna, N. Mueller, ‘Simulation of fatigue failure in composite axial compressor blades’, Materials & Design, (2011), 32(4): 2058-2065. [8] M.P. Boyce, ‘Gas turbine engineering handbook’, Butterworth-Heinemann, 4 ed., 2011.
10
[9] R.E. Dundas, ‘Engineering and metallographic aspects of gas turbine engine failure investigation: identifying the causes’, Aviation Mechanics Bulletin, (1994) 42: 1-11. [10] H. Amaya, T. Mori, K. Kondo, H. Hirata, M. Ueda, ‘Effect of chromium and molybedenum on corrosion resistance of super 13Cr martenistic stainless steel in CO2 environment’, CORROSION 98, March 22 - 27, 1998, San Diego Ca. [11] R. Guillou, M. Guttmann, Ph. Dumoulin, ‘Role of molybdenum in phosphorusinduced temper embrittlement of 12%Cr martensitic stainless steel’, Metal Science, (1981) 15(2): 63-72. [12] I. Thorbjornsson, ‘Corrosion fatigue testing of eight different steels in an Icelandic geothermal environment’, Materials & Design, (1995), 16(2): 97-102. [13] X. Wang, J. Xu, C. Sun, ‘Effects of sulfate-reducing bacterial on corrosion of 403 stainless steel in soils containing chloride ions’, International Journal of Electrochemical Science, (2013), 8: 821-830. [14] Metals Handbook, Vol. 12: Fractography, ASM International, 9 ed., 1987. [15] B. Li, Y. Shen, W. Hu, ‘Casting defects induced fatigue damage in aircraft frames of ZL205A aluminum alloy–A failure analysis’, Material & Design, (2011), 32(5): 2570-2582. [16] AMS5611F, ‘Steel, Corrosion and Heat-Resistant, Bars, Wire, Forgings, Tubing, and Rings, 12Cr, Ferrite Controlled, Consumable Electrode Melted’, SAE International, 2006. [17] X.P. Ma, L.J. Wang, B. Qin, C.M. Liu, S.V. Subramanian, ‘Effect of N on microstructure and mechanical properties of 16Cr5Ni1Mo martensitic stainless steel’, Materials & Design, (2012), 34: 74-81. [18] G. Alkan, D. Chae, S.J. Kim, ‘Effect of δ ferrite on impact property of hot-rolled 12Cr–Ni steel’, Materials Science and Engineering A: (2013), 585: 39-46. [19] J.N. Asaad, S.Y. Tawfik, ‘Polymeric composites based on polystyrene and cement dust wastes’, Materials & Design, (2011), 32(10): 5113-5119. [20] P. Chaunsali, S. Peethamparan, ‘Influence of the composition of cement kiln dust on its interaction with fly ash and slag’, Cement and Concrete Research, (2013), 54: 106-113. [21] R. Ebara, ‘Long-term corrosion fatigue behaviour of structural materials’, Fatigue & Fracture of Engineering Materials & Structures, (2002), 25: 855–859.
11
[22] R. Ebara, ‘The present situation and future problems in ultrasonic fatigue testingMainly reviewed on environmental effects and materials’ screening’, International Journal of Fatigue, (2006), 28(11): 1465-1470. [23] B.M. Schonbauer, A. Perlega, S.E. Stanzl-Tschegg, ‘Pit-to-crack transition and corrosion fatigue of 12% Cr steam turbine blade steel’, 13th International Conference on Fracture, June 16–21, 2013, Beijing, China. [24] Metals Handbook, Vol. 4: Heat Treating, ASM International, 10 ed., 1991. [25] J.R. Davis, Stainless Steels (ASM Specialty Handbook), ASM International, 1994.
12
Table 1- Chemical composition of the compressor blade alloy and standard AISI 403 stainless steel (Wt%) Element
Fe
C
Cr
Si
Mn
P
S
Ni
Mo
Compress
Bas
0.115
12.9
0.175
0.42
0.015
0.013
0.1
0.54
or Blade
e
4
4
AISI 403
Bas
0.15ma
11.5- 0.5ma
1ma
0.04ma
0.03ma
-
-
e
x
13
x
x
x
5
x
Table 2- Chemical composition range of cement kilns dusts collected from the cement manufacturing company in the vicinity of the power plant Constituent
SiO2
Al2O3
Fe2O3
CaO
MgO
SO3
K2O
Na2O
Cl
-
L.O.I
Percent
11.24-
6.17-
1.93-
46.01-
0.48-
1.68-
2.02-
0.84-
2.81-
4.25-
16.41
7.36
2.62
56.44
2.58
8.7
3.21
2.65
6.93
8.26
Table 3- Fatigue strength (Δσ/2) of AISI 403 stainless steel determined in aerated 6 ppm Cl- solution at 90°C (in MPa). R = σmin/σmax [23] Pit Size
R= -1
R= 0.05
R= 0.5
R= 0.8
Smooth 100 µm 250 µm
300 160
265 175
165 110
72.5 65 52.5
Table 4- Hardness of AISI 403 steel after selected heat treatments [24-25] Heat Treatment Type
Full Annealing Quenching and Tempering
Tempering
Vickers Hardness Number
Temperature (°C)
(HV)
205-370 565-605
137-170 372-471 266-310
13
600-650
230-272
Figures Captions: Fig.1. (a-b) Images of the broken blades from different views Fig.2. (a-b) Stereo microscopy images of different regions of the broken blade fracture surface Fig.3. SEM images of the fracture surface of the blade, (a-b) low magnification, (c-d) medium magnification Fig.4. (a-d) SEM image of the nucleation site for fatigue cracks showing the presence of corrosion pits in these regions, (e) secondary and back-scattered SEM images of surface corrosion pits, (f) EDS analysis of corrosion products in pits Fig.5. High magnification SEM images of the fatigue crack growth regions in fracture surface of the broken blade indicating the brittle striations and also the microcracks Fig.6. (a-b) Optical microscopy images of the metallographic samples before etching, (c) microstructure of the alloy after etching Fig.7. (a-c) SEM images of the blade microstructure before etching at different magnifications, (d) EDS analysis of region A in Fig. 7-c Fig.8. (a) SEM image of a corrosion pit on the surface of the broken blade, (b) EDS analysis of region C in Fig. 8-a Fig.9. A comparison of mean dust concentration in air entered to gas turbines for different power plants in Iran Fig.10. Effect of different contents of Cl containing compounds on fatigue properties of AISI 403 stainless steels [21-22]. Rotating bending test, 60 Hz, R (σmin/σmax) = -1 Fig.11. Vickers hardness values in different regions of the broken blade
14
15
16
17
18
19
20
21
22
23
24
25
Highlights: - One of the first stage broken blades of a compressor was failure analyzed - The power plant located near a cement company and an oil refining company - Hence, the inlet air of compressor had large amounts of Cl and S compounds
- Therefore, the corrosion pits were formed on the compressor blade surface - Corrosion pits acted as stress concentration sites for fatigue fracture
26