- Email: [email protected]
Failure analysis of shear pins in wind turbine generator
Engineering Failure Analysis 18 (2011) 325–339
Contents lists available at ScienceDirect
Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
Failure analysis of shear pins in wind turbine generator S. Sankar a,⇑, M. Nataraj b, V. Prabhu Raja c a b c
Department of Mechanical Engineering, Anna University, Coimbatore, India Department of Mechanical Engineering, Government College of Technology, Coimbatore, India Department of Mechanical Engineering, PSG College of Technology, Coimbatore, India
a r t i c l e
i n f o
Article history: Received 14 May 2010 Received in revised form 14 September 2010 Accepted 17 September 2010 Available online 25 September 2010 Keywords: Coupling Failure analysis Fractography FEA Nacelle
a b s t r a c t This paper examines the failure of shear pins that connect gearbox and generator in a wind turbine generator. Chemical and micro-structural together with hardness measurements have been performed to check any deviation in the material specification. The failure mechanism is analyzed by both visual and Scanning Electron Microscope (SEM) inspection on the fractured surface. Finite Element Analysis (FEA) of shear pin is carried out to determine the root cause for its failure. The neck diameter of shear pin is optimized for the safe operation of the wind turbine. It is inferred from the observations based on fractography study on the fractured surface of the shear pins that the misalignment between the driving and the driven elements in the wind turbine leads to low cyclic fatigue growth. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction The coupling connecting the gearbox assembly and an electric generator of the wind turbine generator (WTG) are fitted with shear pins numbering seven. The driver and driven ends of the coupling are shown in Figs. 1 and 2 respectively. The shear pin is a mechanical sacrificial component like an electric fuse designed to break itself as and when the mechanical overload arises to prevent the severe damage of expensive components in gearbox assembly and electric generator. The rubber buckles fitted in the coupling provide flexibility during operation. The nacelle comprises the gearbox assembly, coupling assembly, brake disc assembly, generator and cover. The sudden switch over of generator from high speed (1007 rpm – 6 poles) to slow speed (750 rpm – 8 poles), whenever there is a sudden change in wind velocity, is termed as down coupling. During an operation of the WTG, it was observed that the frequent pre-mature failure of the shear pins occurred during down coupling even the WTG did not attain the rated power. The gearbox in WTG is subjected to shock loads due to fluctuating wind force, sudden grid drop, non -synchronization of pitching and sudden braking. In the present work, a case study of failure occurring in a WTG of capacity 350 kW, commissioned on 31/12/2001, is undertaken. The WTG ceased to function on 06/06/2007 due to the failure of shear pins causing disconnection of transmission of drive from gearbox end (drive) to generator end (driven). During an inspection, it was observed that all the seven shear pins had failed at the neck diameter (Fig. 5). The shear pin is generally designed in such a way that it should fail as and when the mechanical overloads come up so as to protect the highly expensive units, namely, gearbox and generator. The sudden and frequent change in wind velocity, misalignment between gearbox and generator and bearing (rolling contact type) failure are the major sources for mechanical overload. Figs. 3 and 4 shows an assembly of drive train components in WTG and an arrangement of shear pins in the coupling respectively. The researchers in the field of wind energy have been ⇑ Corresponding author. Tel.: +91 994 2419636; fax: +91 422 2221951. E-mail address: [email protected] (S. Sankar). 1350-6307/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.engfailanal.2010.09.012
326
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
Fig. 1. Driver end of coupling.
Fig. 2. Driven end of coupling.
working on the development of advanced materials, new heat treatment methods, designing larger size couplings, to tackle the problem of failure of shear pins due to mechanical overload. Sreekumar investigated the deleterious effect of inclusion in steel, especially for fatigue behavior of a failed steel shear pin [1]. Magarotto indicated that the stable crack propagation due to fatigue caused by bi-directional bending was nucleated at machining marks under normal load [2]. Piramohammadi has recommended while analyzing a blade, a disk and a lock-pin in 3D finite element model that inadequate design and long service reduced the performance of lock–pin for sustaining a severe hot condition [3]. Smith suggested that misalignment during shear pin assembly associated with vibration might have promoted the pre-mature failure of the shear pin by fatigue [4]. The maximum shear strength is derived from the maximum shear stress theory, which states that the material shear strength is 0.5 times that of the yield strength [5]. Rahman investigated the prediction of fatigue life, effects of the stress combination for the proportional loading condition during assessment of multiaxial fatigue behavior of cylinder head for new free piston linear engine [6]. Goksenli investigated the reason for elevator drive shaft failure by performing stress analysis using finite element method and also determined the permissible force and torque acting on the shaft [7]. Bhaumik analyzed the failure of a hollow power transmission shaft and recommended for failure prevention [8]. Many research papers have been published on failure analysis of Kaplan turbine [9], Steam generators [10]. Elevator drive shaft [7] and Air compressor [11]; but an attempt has not been made to analyze the reasons for the failure of shear pins employed in wind turbine generator, to the best knowledge of the investigator. Therefore, it is imperative to investigate the problem of failure of shear pin in WTG.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
327
Fig. 3. Drive train components in WTG.
Fig. 4. Shear pins arrangement in coupling.
Fig. 5. Failure shear pin.
2. Problem investigation A study was carried out to investigate the frequency of failure of shear pin used in the coupling. Table 1 gives the history of shear pin failure in the WTG under consideration. It is obvious from Table 1 that the shear pin fails within 6 years of usage against the recommended life of 10 years. In the event of such pre-mature failure, the shear pin has to be replaced
328
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339 Table 1 History of shear pin failure. Commissioning of turbine
Failure frequency
Time taken in hours for shear pin replacement
31/12/2001
06/01/2007 16/06/2007 20/06/2008 02/01/2009 15/02/2009 06/03/2009 17/03/2009 04/04/2009 07/04/2009 10/04/2009 16/06/2009
7 4 5 12 5 4 4 4 5 7 5
immediately for getting continuous power supply from the wind turbine. The frequent replacement of shear pin results in increase in bore diameter of locating bush leading to weakening of the coupling assembly by way of increased clearance between the bush and shear pin. The replacement of the defective coupling necessitates the dismantling of generator to be brought down from wind turbine nacelle located at the top of the tower. The de-erection of generator needs a crane of high capacity of the order of 200–400 Tonne at the wind turbine site to swap the coupling in the turbine nacelle. A team of operation and maintenance engineers and product development engineers has been trying to explore the possibility of designing stronger pins and employing new manufacturing process for trouble free operation of the wind turbine. This research study is intended to predict how and why the shear pins fail in wind turbine and 1 remedy to minimize the occurrence of failures either by design changes or design modification of neck diameter or introducing surface treatment at the neck portion of shear pin. More than 150 gearboxes were examined for radial and axial play (Fig. 7) to check the possibility of failure of shear pins due to radial play between the bearing outer race and gearbox housing. The examination revealed that a radial play of 0.05 mm and above is found to increase the failure rate. 3. Failure analysis The gearbox has been designed for a power of 350 kW and output speed of 1007 rpm to achieve the final step up speed ratio of 1:31.5. The high speed shaft of the gearbox is coupled with the driver end of the coupling and the driven end is connected with the generator by means of seven shear pins. The driven end of the coupling is connected with generator shaft by a key with shrink fit. The technical team inspected the damaged shear pins (Figs. 5 and 6) and presumed that the failures may be due to over load by wind force or misalignment between gearbox and generator assembly. 3.1. Material and geometry The preliminary step in failure analysis is material identification. The shear pins are made of ETG 88 unique (drawn) special steel that has very good mechanical properties combined with extraordinary good machinability.
Fig. 6. Fatigue fracture of shear pin.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
329
Fig. 7. Radial play measurement in gearbox.
The material properties of ETG 88 are given below: 2.1105 MPa 800–950 N/mm2
Young’s modulus Tensile strength
Hardness Yield strength
29 HRC >685 N/mm2
The hardness of the shear pin and locating bush were measured in Rockwell Hardness Tester in ‘C’ scale and found to be 25–27 HRC and 35 HRC respectively. 3.2. Compositional analysis The chemical analysis of the shear pin was carried out by spectrophotometer and the results are presented in Table 2. It is observed from Table 2 that the composition of failed shear pin confirms to the specifications of ETG 88 material. 3.3. Visual inspection The visual inspection revealed that the fracture occurred at the neck region of the shear pin (Fig. 5). When the shear pin cuts into pieces, the two halves of fractured surface rub against each other leading to surface damage. Machining marks were observed at the neck portion of the failed shear pins due to poor manufacturing process followed during production. While analyzing the fractured surface (Fig. 6), typical torsional cyclic fatigue fracture was noticed. Fatigue crack initiated at the corner of the neck region where the machining marks were present and propagated along the whole surface (Fig. 8). The crack nucleates and propagates due to the cyclic loading until a critical crack length was attained resulting in pre-mature failure of the shear pin. 3.4. Metallography The specimen was prepared by rough, an intermediate and fine polishing and then etched with 2% Nital solution for examination. Microstructures of the damaged portions of two different shear pins were analyzed using optical microscope as well as Scanning Electron Microscope (SEM). The photographic views of the microstructure of shear pin based on Table 2 Chemical analysis of the shear pin material. Element
Specifications of ETG 88 (wt.%)
Composition of failed shear pin
Carbon (C) Silicon (Si) Manganese (Mn) Sulphur (S) Phosphorus (P) Lead (Pb) Chromium (Cr) Nickel (Ni) Molybdenum (Mo) Iron (Fe)
0.42–0.48 0.10–0.30 1.35–1.65 0.24–0.33 60.04 (max) 0.0 – – – Remaining
0.510 0.108 1.54 0.284 0.0160 0.00600 0.0780 0.0810 <0.002 Remaining
330
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
Fig. 8. Crack path (100).
Fig. 9. Crack on shear pin (200).
experimental study are shown in Figs. 9–11. Micro-cracks clearly seen on the shear pin surface (Fig. 9) initiated from the outer surface of the machined place, and then reached the centre portion of the shear pin. During micro-structure examination, ferrite plus over tempered martensite structure (Fig. 10) is clearly seen on some pins and retained austenite plus over tempered martensite structure (Fig. 11) is seen on some other pins. The specification of the material [12] includes tempering treatment; the same was clearly evident from the microstructure examination. Further, the shear pins were hardened and tempered after the notch (neck diameter) was made.
3.5. Scanning electron fractography The fractured shear pins were cleaned ultrasonically in acetone, and examined using SEM and the outcome of the investigation is illustrated in Figs. 12–16. No surface preparation and coating were done on the fractured surface of the shear pin during the examination. Furthermore, an accelerating voltage of 25 kV was applied during the SEM analysis. Micro-cracks and oxide particles (Figs. 12 and 13) were seen on the surface of the fractured shear pin due to overload. The mode of failure of shear pin is either to brittle fracture (Figs. 14 and 15) or fatigue fracture (Fig. 16) because of low cyclic fatigue phenomenon.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
331
Fig. 10. Microstructure of shear pin material. It consists of ferrite and over tempered martensite structure (200).
Fig. 11. Microstructure of shear pin material. It consists of retained austenite and over tempered martensite structure (200).
4. Finite element analysis of existing design 4.1. Force and stress calculation Power (P) = 350 kW Speed (N) = 1007 rpm Coupling PCD (D) = 314 mm Diameter of the shear pin (d) = 10 mm Neck diameter of shear pin (d1) = 5 mm Total number of shear pins (n) = 7
P 60 103 ¼ 3319 Nm 2 pN
ð1Þ
2T 2 3319 ¼ ¼ 21141 N D 0:314
ð2Þ
Torque ðTÞ ¼ Force ðF t Þ ¼
332
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
Fig. 12. SEM fractograph at the neck portion. Micro-crack and oxides are observed.
Fig. 13. SEM fractograph at the neck portion. Micro-crack observed.
Nominal shear stress for 5 mm neck diameter ðsÞ ¼
Nominal shear stress for 10 mm shear pin ðsÞ ¼
4T 2 103
p
2 d1
nD
4T 2 103
p d2 n D
¼
¼
4 3319 2 103 3:14 ð5Þ2 7 314
4 3319 2 103 3:14 ð10Þ2 7 314
¼ 154 N=mm2
¼ 38:47 N=mm2
ð3Þ
ð4Þ
The assumption made during the calculation: All shear pins take up the total load equally and hence the load on each shear pin is = 21141 ¼ 3020:14 N 7
4.2. Finite element analysis Fig. 17 shows the dimensions of the shear pin. The 20 node SOLID 192 element type chosen for meshing has three degrees of freedom per node. Fig. 18 shows the meshed model of the shear pin and Table 3 gives the material properties and mesh details used for the analysis. The shear pin is assumed as a cantilever and loads (tangential force and bending shear forces) are applied at the free end. The total number of elements used in 5 mm neck diameter model is 41616, 41960 for 5.5 mm diameter model, 50534 for 6 mm diameter model, 54944 for 6.5 mm diameter model and 57134 for 7 mm diameter model.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
Fig. 14. SEM fractograph at the neck portion. Brittle fracture observed.
Fig. 15. SEM fractograph at the neck portion. Brittle fracture is seen.
Fig. 16. SEM fractograph at the neck region. Fatigue fracture observed.
333
334
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
Fig. 17. Dimensions of the shear pin.
Fig. 18. Finite element meshed model of the shear pin.
Table 3 Mechanical properties and mesh details. Material properties Young’s modulus (E) Poisson’s ratio
210,000 MPa 0.3
Mesh details Element type Mesh type and size No of nodes for 5 mm neck diameter shear pin No of elements for 5 mm neck diameter shear pin
20 Node SOLID 192 Very fine mesh, mesh size 1 59,850 41,616
4.2.1. Loads on the shear pin Two cases of loading, namely, (i) twisting force and (ii) twisting force combined with bending force are considered for investigation. In case (i), shear pins are subjected to pure shear loading (Tangential force) and in case (ii) shear pins are subjected to shear load combined with bending due to misalignment between the gearbox and generator. The analysis is carried out to determine the shear stress and von Mises stress. Fig. 19 shows the distribution of shear stress and Fig. 20 shows the distribution of von Mises stress for case (i) where shear pin with 5 mm neck diameter is considered. The combination of shear and bending loads was applied in different percentage levels. The maximum shear strength is derived from the maximum shear stress theory, which states that the material shear strength is half of the material yield strength [5]. The maximum shear strength is 400 N/mm2 as maximum yield strength of the shear pin material is 800 N/ mm2 (Section 3.1). It is observed from Table 4 that the von Mises stress exceeds beyond 400 N/mm2 when the misalignment between gearbox housing and the high speed shaft bearing outer race is 0.05 mm and above, whereas shear stress is crossing the nominal value 154 N/mm2 when the misalignment is around 0.20 mm. Fig. 21 shows the distribution of shear stress and Fig. 22 shows the distribution of von Mises stress respectively for shear loading combined with 5% bending load.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
335
Fig. 19. Shear stress distribution in shear pin with 5 mm neck diameter.
Fig. 20. Von Mises stress distribution in shear pin with 5 mm neck diameter.
5. Finite element analysis of modified design A brainstorming session was held to investigate the reason for pre-mature failure of shear pin. It was decided to undertake field study to investigate the radial play between gearbox and generator. The pre-mature failure of the shear pin may be due to the cyclic fatigue load because of the radial play (Section 2). Hence, a modified design which will facilitate the shear pin to withstand cyclic fatigue loading is made and validated using FEM. The chemical analysis of the failed shear pin has proven that the composition of the shear pin is in accordance with the specifications of the ETG 88 steel and hence there is no materials defect. Therefore, design alternatives are tried by geometrical modification since the neck region of the shear pin is the critical zone. The design modification was done based on the neck diameter. Fig. 23 shows the dimensions of the proposed design. The shear pin in accordance with the new design is modeled and analyzed using ANSYS software. Fig. 24 shows the shear stress distribution and Fig. 25 depicts von Mises stress distribution for the modified shear pin. The Finite
336
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339 Table 4 Shear stress and von Mises stress against various misalignments. Misalignment (mm)
Shear stress (N/mm2)
von Mises stress (N/mm2)
0 0.05 0.1 0.15 0.20
74.50 78.099 87.773 101.842 171.223
287.94 668.534 688.086 989.238 1276
Fig. 21. Shear stress distribution in shear pin with 5 mm neck diameter for 0.05 mm misalignment.
Fig. 22. Von Mises stress distribution in shear pin with 5 mm neck diameter for 0.05 mm misalignment.
Element Analysis envisages that the shear pin with 6.5 mm neck diameter would not fail since the induced stress is well within the specified stress limit.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
Fig. 23. Proposed shear pin with 6.5 mm neck diameter.
Fig. 24. Shear stress distribution in shear pin with 6.5 mm neck diameter.
Fig. 25. Von Mises stress distribution in shear pin with 6.5 mm neck diameter.
337
338
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339 Table 5 Shear stress and von Mises stress for different neck diameter. Neck diameter (mm)
Shear stress (N/mm2)
von Mises stress (N/mm2)
5 5.5 6.0 6.5 7.0
74.50 58.214 46.46 37.678 29.896
287.94 222.828 172.87 139.17 111.787
Table 5 gives the shear stress and corresponding von Mises Stress obtained using ANSYS corresponding to various neck diameters considered for analysis. It is evident from Table 5 that the von Mises stress (139.17 N/mm2) for 6.5 mm neck diameter is less than that of the nominal shear stress (154.0 N/mm2) of 5.0 mm neck diameter (Section 4.1), also the induced stress is within the safe limit of the material yield stress (400 N/mm2) 6. Discussion The visual observation and the micro-fractographic examination confirmed that the fracture was nucleated at the machining marks near the neck region of the shear pin. The results of the chemical analysis (Table 2) showed that the shear pin material is in accordance with ETG 88 specifications. Failure analysis suggests that the crack propagation is due to the cyclic fatigue load until a critical crack length was attained. The Finite Element Analysis considering pure shear loading effect and the various levels of combined shear and bending loads showed that the maximum stress was found along the neck region of the shear pin. The shear stress for the pure shear loading was 74.5 N/mm2 for 5 mm neck diameter with negligible misalignment (Table 4) and it is found to be 171.223 N/mm2 for 0.20 mm misalignment which exceeds the nominal shear stress for 5 mm neck diameter. Further, according to maximum shear stress theory the material shear stress is 400 N/ mm2, but the von Mises stress is 668.537 N/mm2 for 0.05 mm misalignment which is more than the material shear stress. It is well understood that the frequent failure of the shear pin happened due to the combined bending and shear loading which caused reversed fatigue loading in the shear pin. 7. Conclusion The major conclusions of the present work are as follows: Based on evidence collected at the site and the investigation carried out at the laboratory, the failure of the shear pin is attributed to the growth of the pre-existing fatigue crack to the critical size scale which triggered the frequent tripping of the wind turbine during operation with misalignment. The metallographic and chemical analyses reveal that the failure is not related to any defect in material or with any abnormal operating conditions like temperature. Visual observation and the micrographic analysis indicated that the fracture was nucleated on machining marks along the surface of the shear pin. The pre-mature and frequent failure of the shear pin was because of fatigue due to misalignment (radial play) between the gearbox and generator assembly in nacelle. Finite element analysis reveals that for pure shear loading condition, the von Mises stress exceeds the tensile yield strength of the ETG 88 material. It is because of the reduced neck diameter and misalignment between the driving and driven shafts. Due to misalignment and incorrect neck diameter, fatigue loading was produced with combined bending and shear load, which causes the pre-mature failure of the shear pin.
Acknowledgements The authors thank the Department of Metallurgy of PSG College of Technology, Coimbatore, India for Microscopic examination and Metallographic support. The authors are immensely thankful for the support rendered by M/s Suzlon Infrastructure Services Limited, Coimbatore, India to formulate this research problem and offering their technical expertise for the successful completion of this research study. References [1] Abhay Jha K, Sreekumar K, Mittal MC. Metallurgical studies on a failed EN19 steel shear pin. Eng Fail Anal 2008;15:922–30. [2] Azevedo CRF, Magarotto D, Araujo JA, Ferreira JLA. Bending fatigue of stainless steel shear pins belonging to a hydroelectric plant. Eng Fail Anal 2009;16:1126–40. [3] Poursaeidi E, Pirmohammadi AA, Mohammadi Arhani MR. Mechanical investigation of a failed lock-pin. ASME 2009;131:042501. [4] Smith M, Fisher F, Romios M, Es-Said OS. On the redesign of a shear pin under cyclic bending loads. Eng Fail Anal 2007;14:138–46. [5] Shigley JE. Mechanical engineering design. 3rd ed. New York: McGraw hill; 1989.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
339
[6] Rahman MM, Ariffin AK, Abdullah S, Rosli AB, Sani MSM. Assessment of multiaxial fatigue behavior of cylinder head for a new free piston linear engine. In: Proceedings of the int. conference on mechanical engineering (ICME2007) Dhaka, Bangladesh; 2007. [7] Goksenli A, Eryurek IB. Failure analysis of an elevator drive shaft. Eng Fail Anal 2009;16:1011–9. [8] Bhaumik SK, Rangaraju R, Parameswara MA, Venkataswamy MA, Bhaskaran TA, Krishnan RV. Fatigue failure of a hollow power transmission shaft. Eng Fail Anal 2002;9:457–67. [9] Luo Yongyao, Wang Zhengwei, Zeng Jidi, Lin Jiayang. Failure of piston rod caused by unsteady, unbalanced, unsynchronized blade torques in a Kaplan turbine. Eng Fail Anal 2010;17:192–9. [10] Cicero S, Lacalle R, Cicero R, Garcia J. Failure analysis of a steam generator super heater drain tube used in a dump. Eng Fail Anal 2010;17:301–12. [11] Siva Prasad S, Narasaiah N, Das SK, Das G, Tarafder S, Gupta KK, et al. Investigation on the failure of air compressor. Eng Fail Anal 2010;17:150–7. [12] Metals hand book. 8th ed., vol.7. Metals Park (OH): ASM; 1972. p. 152.
Contents lists available at ScienceDirect
Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
Failure analysis of shear pins in wind turbine generator S. Sankar a,⇑, M. Nataraj b, V. Prabhu Raja c a b c
Department of Mechanical Engineering, Anna University, Coimbatore, India Department of Mechanical Engineering, Government College of Technology, Coimbatore, India Department of Mechanical Engineering, PSG College of Technology, Coimbatore, India
a r t i c l e
i n f o
Article history: Received 14 May 2010 Received in revised form 14 September 2010 Accepted 17 September 2010 Available online 25 September 2010 Keywords: Coupling Failure analysis Fractography FEA Nacelle
a b s t r a c t This paper examines the failure of shear pins that connect gearbox and generator in a wind turbine generator. Chemical and micro-structural together with hardness measurements have been performed to check any deviation in the material specification. The failure mechanism is analyzed by both visual and Scanning Electron Microscope (SEM) inspection on the fractured surface. Finite Element Analysis (FEA) of shear pin is carried out to determine the root cause for its failure. The neck diameter of shear pin is optimized for the safe operation of the wind turbine. It is inferred from the observations based on fractography study on the fractured surface of the shear pins that the misalignment between the driving and the driven elements in the wind turbine leads to low cyclic fatigue growth. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction The coupling connecting the gearbox assembly and an electric generator of the wind turbine generator (WTG) are fitted with shear pins numbering seven. The driver and driven ends of the coupling are shown in Figs. 1 and 2 respectively. The shear pin is a mechanical sacrificial component like an electric fuse designed to break itself as and when the mechanical overload arises to prevent the severe damage of expensive components in gearbox assembly and electric generator. The rubber buckles fitted in the coupling provide flexibility during operation. The nacelle comprises the gearbox assembly, coupling assembly, brake disc assembly, generator and cover. The sudden switch over of generator from high speed (1007 rpm – 6 poles) to slow speed (750 rpm – 8 poles), whenever there is a sudden change in wind velocity, is termed as down coupling. During an operation of the WTG, it was observed that the frequent pre-mature failure of the shear pins occurred during down coupling even the WTG did not attain the rated power. The gearbox in WTG is subjected to shock loads due to fluctuating wind force, sudden grid drop, non -synchronization of pitching and sudden braking. In the present work, a case study of failure occurring in a WTG of capacity 350 kW, commissioned on 31/12/2001, is undertaken. The WTG ceased to function on 06/06/2007 due to the failure of shear pins causing disconnection of transmission of drive from gearbox end (drive) to generator end (driven). During an inspection, it was observed that all the seven shear pins had failed at the neck diameter (Fig. 5). The shear pin is generally designed in such a way that it should fail as and when the mechanical overloads come up so as to protect the highly expensive units, namely, gearbox and generator. The sudden and frequent change in wind velocity, misalignment between gearbox and generator and bearing (rolling contact type) failure are the major sources for mechanical overload. Figs. 3 and 4 shows an assembly of drive train components in WTG and an arrangement of shear pins in the coupling respectively. The researchers in the field of wind energy have been ⇑ Corresponding author. Tel.: +91 994 2419636; fax: +91 422 2221951. E-mail address: [email protected] (S. Sankar). 1350-6307/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.engfailanal.2010.09.012
326
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
Fig. 1. Driver end of coupling.
Fig. 2. Driven end of coupling.
working on the development of advanced materials, new heat treatment methods, designing larger size couplings, to tackle the problem of failure of shear pins due to mechanical overload. Sreekumar investigated the deleterious effect of inclusion in steel, especially for fatigue behavior of a failed steel shear pin [1]. Magarotto indicated that the stable crack propagation due to fatigue caused by bi-directional bending was nucleated at machining marks under normal load [2]. Piramohammadi has recommended while analyzing a blade, a disk and a lock-pin in 3D finite element model that inadequate design and long service reduced the performance of lock–pin for sustaining a severe hot condition [3]. Smith suggested that misalignment during shear pin assembly associated with vibration might have promoted the pre-mature failure of the shear pin by fatigue [4]. The maximum shear strength is derived from the maximum shear stress theory, which states that the material shear strength is 0.5 times that of the yield strength [5]. Rahman investigated the prediction of fatigue life, effects of the stress combination for the proportional loading condition during assessment of multiaxial fatigue behavior of cylinder head for new free piston linear engine [6]. Goksenli investigated the reason for elevator drive shaft failure by performing stress analysis using finite element method and also determined the permissible force and torque acting on the shaft [7]. Bhaumik analyzed the failure of a hollow power transmission shaft and recommended for failure prevention [8]. Many research papers have been published on failure analysis of Kaplan turbine [9], Steam generators [10]. Elevator drive shaft [7] and Air compressor [11]; but an attempt has not been made to analyze the reasons for the failure of shear pins employed in wind turbine generator, to the best knowledge of the investigator. Therefore, it is imperative to investigate the problem of failure of shear pin in WTG.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
327
Fig. 3. Drive train components in WTG.
Fig. 4. Shear pins arrangement in coupling.
Fig. 5. Failure shear pin.
2. Problem investigation A study was carried out to investigate the frequency of failure of shear pin used in the coupling. Table 1 gives the history of shear pin failure in the WTG under consideration. It is obvious from Table 1 that the shear pin fails within 6 years of usage against the recommended life of 10 years. In the event of such pre-mature failure, the shear pin has to be replaced
328
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339 Table 1 History of shear pin failure. Commissioning of turbine
Failure frequency
Time taken in hours for shear pin replacement
31/12/2001
06/01/2007 16/06/2007 20/06/2008 02/01/2009 15/02/2009 06/03/2009 17/03/2009 04/04/2009 07/04/2009 10/04/2009 16/06/2009
7 4 5 12 5 4 4 4 5 7 5
immediately for getting continuous power supply from the wind turbine. The frequent replacement of shear pin results in increase in bore diameter of locating bush leading to weakening of the coupling assembly by way of increased clearance between the bush and shear pin. The replacement of the defective coupling necessitates the dismantling of generator to be brought down from wind turbine nacelle located at the top of the tower. The de-erection of generator needs a crane of high capacity of the order of 200–400 Tonne at the wind turbine site to swap the coupling in the turbine nacelle. A team of operation and maintenance engineers and product development engineers has been trying to explore the possibility of designing stronger pins and employing new manufacturing process for trouble free operation of the wind turbine. This research study is intended to predict how and why the shear pins fail in wind turbine and 1 remedy to minimize the occurrence of failures either by design changes or design modification of neck diameter or introducing surface treatment at the neck portion of shear pin. More than 150 gearboxes were examined for radial and axial play (Fig. 7) to check the possibility of failure of shear pins due to radial play between the bearing outer race and gearbox housing. The examination revealed that a radial play of 0.05 mm and above is found to increase the failure rate. 3. Failure analysis The gearbox has been designed for a power of 350 kW and output speed of 1007 rpm to achieve the final step up speed ratio of 1:31.5. The high speed shaft of the gearbox is coupled with the driver end of the coupling and the driven end is connected with the generator by means of seven shear pins. The driven end of the coupling is connected with generator shaft by a key with shrink fit. The technical team inspected the damaged shear pins (Figs. 5 and 6) and presumed that the failures may be due to over load by wind force or misalignment between gearbox and generator assembly. 3.1. Material and geometry The preliminary step in failure analysis is material identification. The shear pins are made of ETG 88 unique (drawn) special steel that has very good mechanical properties combined with extraordinary good machinability.
Fig. 6. Fatigue fracture of shear pin.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
329
Fig. 7. Radial play measurement in gearbox.
The material properties of ETG 88 are given below: 2.1105 MPa 800–950 N/mm2
Young’s modulus Tensile strength
Hardness Yield strength
29 HRC >685 N/mm2
The hardness of the shear pin and locating bush were measured in Rockwell Hardness Tester in ‘C’ scale and found to be 25–27 HRC and 35 HRC respectively. 3.2. Compositional analysis The chemical analysis of the shear pin was carried out by spectrophotometer and the results are presented in Table 2. It is observed from Table 2 that the composition of failed shear pin confirms to the specifications of ETG 88 material. 3.3. Visual inspection The visual inspection revealed that the fracture occurred at the neck region of the shear pin (Fig. 5). When the shear pin cuts into pieces, the two halves of fractured surface rub against each other leading to surface damage. Machining marks were observed at the neck portion of the failed shear pins due to poor manufacturing process followed during production. While analyzing the fractured surface (Fig. 6), typical torsional cyclic fatigue fracture was noticed. Fatigue crack initiated at the corner of the neck region where the machining marks were present and propagated along the whole surface (Fig. 8). The crack nucleates and propagates due to the cyclic loading until a critical crack length was attained resulting in pre-mature failure of the shear pin. 3.4. Metallography The specimen was prepared by rough, an intermediate and fine polishing and then etched with 2% Nital solution for examination. Microstructures of the damaged portions of two different shear pins were analyzed using optical microscope as well as Scanning Electron Microscope (SEM). The photographic views of the microstructure of shear pin based on Table 2 Chemical analysis of the shear pin material. Element
Specifications of ETG 88 (wt.%)
Composition of failed shear pin
Carbon (C) Silicon (Si) Manganese (Mn) Sulphur (S) Phosphorus (P) Lead (Pb) Chromium (Cr) Nickel (Ni) Molybdenum (Mo) Iron (Fe)
0.42–0.48 0.10–0.30 1.35–1.65 0.24–0.33 60.04 (max) 0.0 – – – Remaining
0.510 0.108 1.54 0.284 0.0160 0.00600 0.0780 0.0810 <0.002 Remaining
330
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
Fig. 8. Crack path (100).
Fig. 9. Crack on shear pin (200).
experimental study are shown in Figs. 9–11. Micro-cracks clearly seen on the shear pin surface (Fig. 9) initiated from the outer surface of the machined place, and then reached the centre portion of the shear pin. During micro-structure examination, ferrite plus over tempered martensite structure (Fig. 10) is clearly seen on some pins and retained austenite plus over tempered martensite structure (Fig. 11) is seen on some other pins. The specification of the material [12] includes tempering treatment; the same was clearly evident from the microstructure examination. Further, the shear pins were hardened and tempered after the notch (neck diameter) was made.
3.5. Scanning electron fractography The fractured shear pins were cleaned ultrasonically in acetone, and examined using SEM and the outcome of the investigation is illustrated in Figs. 12–16. No surface preparation and coating were done on the fractured surface of the shear pin during the examination. Furthermore, an accelerating voltage of 25 kV was applied during the SEM analysis. Micro-cracks and oxide particles (Figs. 12 and 13) were seen on the surface of the fractured shear pin due to overload. The mode of failure of shear pin is either to brittle fracture (Figs. 14 and 15) or fatigue fracture (Fig. 16) because of low cyclic fatigue phenomenon.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
331
Fig. 10. Microstructure of shear pin material. It consists of ferrite and over tempered martensite structure (200).
Fig. 11. Microstructure of shear pin material. It consists of retained austenite and over tempered martensite structure (200).
4. Finite element analysis of existing design 4.1. Force and stress calculation Power (P) = 350 kW Speed (N) = 1007 rpm Coupling PCD (D) = 314 mm Diameter of the shear pin (d) = 10 mm Neck diameter of shear pin (d1) = 5 mm Total number of shear pins (n) = 7
P 60 103 ¼ 3319 Nm 2 pN
ð1Þ
2T 2 3319 ¼ ¼ 21141 N D 0:314
ð2Þ
Torque ðTÞ ¼ Force ðF t Þ ¼
332
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
Fig. 12. SEM fractograph at the neck portion. Micro-crack and oxides are observed.
Fig. 13. SEM fractograph at the neck portion. Micro-crack observed.
Nominal shear stress for 5 mm neck diameter ðsÞ ¼
Nominal shear stress for 10 mm shear pin ðsÞ ¼
4T 2 103
p
2 d1
nD
4T 2 103
p d2 n D
¼
¼
4 3319 2 103 3:14 ð5Þ2 7 314
4 3319 2 103 3:14 ð10Þ2 7 314
¼ 154 N=mm2
¼ 38:47 N=mm2
ð3Þ
ð4Þ
The assumption made during the calculation: All shear pins take up the total load equally and hence the load on each shear pin is = 21141 ¼ 3020:14 N 7
4.2. Finite element analysis Fig. 17 shows the dimensions of the shear pin. The 20 node SOLID 192 element type chosen for meshing has three degrees of freedom per node. Fig. 18 shows the meshed model of the shear pin and Table 3 gives the material properties and mesh details used for the analysis. The shear pin is assumed as a cantilever and loads (tangential force and bending shear forces) are applied at the free end. The total number of elements used in 5 mm neck diameter model is 41616, 41960 for 5.5 mm diameter model, 50534 for 6 mm diameter model, 54944 for 6.5 mm diameter model and 57134 for 7 mm diameter model.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
Fig. 14. SEM fractograph at the neck portion. Brittle fracture observed.
Fig. 15. SEM fractograph at the neck portion. Brittle fracture is seen.
Fig. 16. SEM fractograph at the neck region. Fatigue fracture observed.
333
334
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
Fig. 17. Dimensions of the shear pin.
Fig. 18. Finite element meshed model of the shear pin.
Table 3 Mechanical properties and mesh details. Material properties Young’s modulus (E) Poisson’s ratio
210,000 MPa 0.3
Mesh details Element type Mesh type and size No of nodes for 5 mm neck diameter shear pin No of elements for 5 mm neck diameter shear pin
20 Node SOLID 192 Very fine mesh, mesh size 1 59,850 41,616
4.2.1. Loads on the shear pin Two cases of loading, namely, (i) twisting force and (ii) twisting force combined with bending force are considered for investigation. In case (i), shear pins are subjected to pure shear loading (Tangential force) and in case (ii) shear pins are subjected to shear load combined with bending due to misalignment between the gearbox and generator. The analysis is carried out to determine the shear stress and von Mises stress. Fig. 19 shows the distribution of shear stress and Fig. 20 shows the distribution of von Mises stress for case (i) where shear pin with 5 mm neck diameter is considered. The combination of shear and bending loads was applied in different percentage levels. The maximum shear strength is derived from the maximum shear stress theory, which states that the material shear strength is half of the material yield strength [5]. The maximum shear strength is 400 N/mm2 as maximum yield strength of the shear pin material is 800 N/ mm2 (Section 3.1). It is observed from Table 4 that the von Mises stress exceeds beyond 400 N/mm2 when the misalignment between gearbox housing and the high speed shaft bearing outer race is 0.05 mm and above, whereas shear stress is crossing the nominal value 154 N/mm2 when the misalignment is around 0.20 mm. Fig. 21 shows the distribution of shear stress and Fig. 22 shows the distribution of von Mises stress respectively for shear loading combined with 5% bending load.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
335
Fig. 19. Shear stress distribution in shear pin with 5 mm neck diameter.
Fig. 20. Von Mises stress distribution in shear pin with 5 mm neck diameter.
5. Finite element analysis of modified design A brainstorming session was held to investigate the reason for pre-mature failure of shear pin. It was decided to undertake field study to investigate the radial play between gearbox and generator. The pre-mature failure of the shear pin may be due to the cyclic fatigue load because of the radial play (Section 2). Hence, a modified design which will facilitate the shear pin to withstand cyclic fatigue loading is made and validated using FEM. The chemical analysis of the failed shear pin has proven that the composition of the shear pin is in accordance with the specifications of the ETG 88 steel and hence there is no materials defect. Therefore, design alternatives are tried by geometrical modification since the neck region of the shear pin is the critical zone. The design modification was done based on the neck diameter. Fig. 23 shows the dimensions of the proposed design. The shear pin in accordance with the new design is modeled and analyzed using ANSYS software. Fig. 24 shows the shear stress distribution and Fig. 25 depicts von Mises stress distribution for the modified shear pin. The Finite
336
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339 Table 4 Shear stress and von Mises stress against various misalignments. Misalignment (mm)
Shear stress (N/mm2)
von Mises stress (N/mm2)
0 0.05 0.1 0.15 0.20
74.50 78.099 87.773 101.842 171.223
287.94 668.534 688.086 989.238 1276
Fig. 21. Shear stress distribution in shear pin with 5 mm neck diameter for 0.05 mm misalignment.
Fig. 22. Von Mises stress distribution in shear pin with 5 mm neck diameter for 0.05 mm misalignment.
Element Analysis envisages that the shear pin with 6.5 mm neck diameter would not fail since the induced stress is well within the specified stress limit.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
Fig. 23. Proposed shear pin with 6.5 mm neck diameter.
Fig. 24. Shear stress distribution in shear pin with 6.5 mm neck diameter.
Fig. 25. Von Mises stress distribution in shear pin with 6.5 mm neck diameter.
337
338
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339 Table 5 Shear stress and von Mises stress for different neck diameter. Neck diameter (mm)
Shear stress (N/mm2)
von Mises stress (N/mm2)
5 5.5 6.0 6.5 7.0
74.50 58.214 46.46 37.678 29.896
287.94 222.828 172.87 139.17 111.787
Table 5 gives the shear stress and corresponding von Mises Stress obtained using ANSYS corresponding to various neck diameters considered for analysis. It is evident from Table 5 that the von Mises stress (139.17 N/mm2) for 6.5 mm neck diameter is less than that of the nominal shear stress (154.0 N/mm2) of 5.0 mm neck diameter (Section 4.1), also the induced stress is within the safe limit of the material yield stress (400 N/mm2) 6. Discussion The visual observation and the micro-fractographic examination confirmed that the fracture was nucleated at the machining marks near the neck region of the shear pin. The results of the chemical analysis (Table 2) showed that the shear pin material is in accordance with ETG 88 specifications. Failure analysis suggests that the crack propagation is due to the cyclic fatigue load until a critical crack length was attained. The Finite Element Analysis considering pure shear loading effect and the various levels of combined shear and bending loads showed that the maximum stress was found along the neck region of the shear pin. The shear stress for the pure shear loading was 74.5 N/mm2 for 5 mm neck diameter with negligible misalignment (Table 4) and it is found to be 171.223 N/mm2 for 0.20 mm misalignment which exceeds the nominal shear stress for 5 mm neck diameter. Further, according to maximum shear stress theory the material shear stress is 400 N/ mm2, but the von Mises stress is 668.537 N/mm2 for 0.05 mm misalignment which is more than the material shear stress. It is well understood that the frequent failure of the shear pin happened due to the combined bending and shear loading which caused reversed fatigue loading in the shear pin. 7. Conclusion The major conclusions of the present work are as follows: Based on evidence collected at the site and the investigation carried out at the laboratory, the failure of the shear pin is attributed to the growth of the pre-existing fatigue crack to the critical size scale which triggered the frequent tripping of the wind turbine during operation with misalignment. The metallographic and chemical analyses reveal that the failure is not related to any defect in material or with any abnormal operating conditions like temperature. Visual observation and the micrographic analysis indicated that the fracture was nucleated on machining marks along the surface of the shear pin. The pre-mature and frequent failure of the shear pin was because of fatigue due to misalignment (radial play) between the gearbox and generator assembly in nacelle. Finite element analysis reveals that for pure shear loading condition, the von Mises stress exceeds the tensile yield strength of the ETG 88 material. It is because of the reduced neck diameter and misalignment between the driving and driven shafts. Due to misalignment and incorrect neck diameter, fatigue loading was produced with combined bending and shear load, which causes the pre-mature failure of the shear pin.
Acknowledgements The authors thank the Department of Metallurgy of PSG College of Technology, Coimbatore, India for Microscopic examination and Metallographic support. The authors are immensely thankful for the support rendered by M/s Suzlon Infrastructure Services Limited, Coimbatore, India to formulate this research problem and offering their technical expertise for the successful completion of this research study. References [1] Abhay Jha K, Sreekumar K, Mittal MC. Metallurgical studies on a failed EN19 steel shear pin. Eng Fail Anal 2008;15:922–30. [2] Azevedo CRF, Magarotto D, Araujo JA, Ferreira JLA. Bending fatigue of stainless steel shear pins belonging to a hydroelectric plant. Eng Fail Anal 2009;16:1126–40. [3] Poursaeidi E, Pirmohammadi AA, Mohammadi Arhani MR. Mechanical investigation of a failed lock-pin. ASME 2009;131:042501. [4] Smith M, Fisher F, Romios M, Es-Said OS. On the redesign of a shear pin under cyclic bending loads. Eng Fail Anal 2007;14:138–46. [5] Shigley JE. Mechanical engineering design. 3rd ed. New York: McGraw hill; 1989.
S. Sankar et al. / Engineering Failure Analysis 18 (2011) 325–339
339
[6] Rahman MM, Ariffin AK, Abdullah S, Rosli AB, Sani MSM. Assessment of multiaxial fatigue behavior of cylinder head for a new free piston linear engine. In: Proceedings of the int. conference on mechanical engineering (ICME2007) Dhaka, Bangladesh; 2007. [7] Goksenli A, Eryurek IB. Failure analysis of an elevator drive shaft. Eng Fail Anal 2009;16:1011–9. [8] Bhaumik SK, Rangaraju R, Parameswara MA, Venkataswamy MA, Bhaskaran TA, Krishnan RV. Fatigue failure of a hollow power transmission shaft. Eng Fail Anal 2002;9:457–67. [9] Luo Yongyao, Wang Zhengwei, Zeng Jidi, Lin Jiayang. Failure of piston rod caused by unsteady, unbalanced, unsynchronized blade torques in a Kaplan turbine. Eng Fail Anal 2010;17:192–9. [10] Cicero S, Lacalle R, Cicero R, Garcia J. Failure analysis of a steam generator super heater drain tube used in a dump. Eng Fail Anal 2010;17:301–12. [11] Siva Prasad S, Narasaiah N, Das SK, Das G, Tarafder S, Gupta KK, et al. Investigation on the failure of air compressor. Eng Fail Anal 2010;17:150–7. [12] Metals hand book. 8th ed., vol.7. Metals Park (OH): ASM; 1972. p. 152.