Failure Analysis Case Studies II Episode 13 ppsx

In the case of gears, three types of contact fatigue damage were identified, depending on the relative movement of the contacting bodies, and the resulting stress distribution in the sur

Failure Analysis Case Studies II

D.R.H Jones (Editor)

CONTACT FATIGUE IN ROLLING-ELEMENT BEARINGS

shafts in contact with sliding bearings [2] In the case of gears, three types of contact fatigue damage were identified, depending on the relative movement of the contacting bodies, and the resulting stress distribution in the surface and near-surface material [l] The characteristics of each type of failure were discussed in detail in [l]

Rolling-element bearings consist of balls or rollers positioned between raceways which conform

to the shape of the rolling element Depending on the bearing design, the loads acting on the bearing may be radial, angular or axial [3] These loads lead to elastic deformation at the points of contact between the rolling elements and the raceways The stress distribution in the surface and near- surface material under these conditions depends on the loads and the curvature and relative movement between the contacting bodies

When bearing operation leads to pure rolling contact between the rolling elements and the raceway, the maximum shear stress occurs at some distance below the surface This situation is similar to that encountered along the pitch-line of gear teeth [ 11 In the early stages of damage, pure rolling forms a highly polished surface, as shown in the case of a bearing cup from a large thrust

Fig 1 Schematic illustration of counterformal (a) and conformal (b) surfaces in contact

Reprinted from Engineering Failure Analysis 4 (2), 155-160 (1997)

410

bearing (Fig 2) [4] Under repeated loading, cracks ultimately initiate at the point of maximum stress, and propagate parallel to the surface At some stage, these cracks deviate and grow towards the contact surface, resulting in the formation of steep-sided pits These pits are usually microscopic, but may, with continued bearing operation, act as stress concentration sites for further damage Under normal bearing operation, it is more common that contact between the rolling elements and the raceway includes both rolling and sliding The resulting stress distribution in the near- surface material under these conditions changes, and the maximum stress point moves closer to the surface Again, this situation is similar to that encountered in the addenda and dedenda of gear teeth [l] Cracks initiate at the contact surface, and propagate to form small, irregular-shaped pits

In some cases, the pits may form in the shape of an arrow-head pointing in the direction of load approach [3] This is similar to the “cyclone pitting effect” also observed in gear teeth [l]

The initiation of surface cracks under rolling-sliding contact can be significantly accelerated by the presence of stress concentration sites on the contact surfaces [3] These include corrosion pits, handling damage, surface inclusions, and dents formed by solid particles entrapped in the lubrication fluid These geometric inhomogeneities lead to high localized stresses, rapid crack initiation, and the formation of contact fatigue pits In some cases, the cracks initiated in this way may propagate through the bearing rings to cause complete fracture An example of this is given in Fig 3, which shows the inner ring of a thrust bearing [ 5 ] Extensive surface damage, probably resulting from the action of solid particles entrapped in the lubricating fluid, is clearly noticeable, as is the through- crack emanating from this damage Figure 4 shows the crack face in the vicinity of the region marked with an arrow in Fig 3, and clearly indicates that crack growth was by fatigue

3 FLAKING AND SPALLING

Under continued operation, the pits formed by rolling and rolling-sliding contact fatigue may progress to form a more severe form of damage known as flaking [3] This results in the formation

of large, irregular pits which cause rapid deterioration and failure of the bearings Flaking is usually first observed on the stationary ring of a bearing, since the surface of this ring is subjected to the maximum stress every time a rolling element passes over it In the case of the rotating ring, the

Fig 2 Thrust bearing cup showing highly polished surfaces typical of the initial stages of rolling contact

412

Fig 5 Flaking damage on the outer raceway of a roller bearing

Fig 6 Localized flaking damage on the inner ring of a thrust bearing

generally found to be associated with pitting originating from surface stress concentration sites In roller bearings, flaking sometimes occurs along a ring on a plane corresponding to the end of the rollers This indicates that the bearing is misaligned, and the loads unevenly distributed Finally, flaking damage is occasionally found at regular intervals corresponding to the rolling element spacing In these cases, damage is associated with indentations produced when the stationary bearing

is loaded, these indentations being referred to as true brinnelling

Another form of severe contact fatigue damage is known as spalling As in the case of gear teeth [l], spalling occurs as a progression of the pits formed by rolling and rolling-sliding contact fatigue,

413

Fig 7 Severe spalling on the inner ring of a thrust bearing

or as a result of cracking at the case-core interface in case-hardened components The damage in this case results in the formation of large, deep pits with sharp edges, steep sides, and flat bases A

good example of this is given in Fig 7, which shows the inner raceway of the failed thrust bearing [6]

4 SUMMARY

Surface contact fatigue is a common cause of failure in rolling-element bearings The extent of damage observed depends on the contact loads, the curvature of the rolling elements, and the relative motion between the contacting surfaces The characteristics of the various types of contact fatigue are as follows:

(a) Microscopic pits form under pure rolling contact These may act as stress concentration sites for further damage

(b) Under rolling-sliding contact, irregular-shaped pits are formed This type of damage is accel- erated by the presence of geometric inhomogeneities such as corrosion pits, handling damage, and dents

(c) Flaking occurs as a progression of the pits formed under rolling and rolling-sliding contact fatigue, and leads to the formation of large, irregular-shaped pits

(d) Spalling refers to the formation of large, deep pits with sharp edges, steep sides, and flat bases,

or to cracking at the case-core interface in case-hardened surfaces

A number of practical examples of bearing failure have been used to illustrate the various types of contact fatigue damage

Acknowledgement-The assistance of the staff of the Advanced Engineering and Testing Services Programme, Mattek, CSIR,

in the preparation of this communication is gratefully acknowledged

REFERENCES

1 Fernandes, P J L and McDuling, C., Engineeering Failure Analysis, 1997, 4(2), 99-107

2 Wulpi, D J., Understanding How Components Fair American Society for Metals, Metals Park, OH, 1985, pp 183-204

414

3 ASM Metals Handbook, Failure of Rolling-element Bearings, Vol 11, Failure Analysis and Prevention, 9th edn American

4 James, A., Failure Analysis Report Advanced Engineering and Testing Services, CSIR, 1996

5 Fernandes, P J L., Failure Analysis Reporr Advanced Engineering and Testing Services, CSIR, 1995

6 James, A., Failure Analysis Report Advanced Engineering and Testing Services, CSIR, 1994

Society for Metals, Metals Park, OH, 1986, pp 490-513

Failure Analysis Case Studies Ii

D.R.H Jones (Editor)

AN AIR CRASH DUE TO FATIGUE FAILURE OF

A BALL BEARING

I SALAM, A TAUQIR*, A UL HAQ and A Q KHAN

Metallurgy Division, Dr A Q Khan Research Laboratorics, GPO Box 502, Rawdlpindi, Pakistan

(Receitied I5 June 1998)

Abstract-The failure analysis of an air crash conclusively shows that the cage of the central main bearing of the compressor region failed due to fatigue The broken piece of the cage got struck between the bearing balls and the races and impaired the function of the bearing resulting in the crash Q 1998 Elsevier Science Ltd A11

ball bearing type The bearings are lubricated by way of a closed loop lubrication system and metallic filters are inserted to clean the debris The life of the CMB, which is suspected to have

caused the accident, is 600 h The bearing of the engine which failed had completed a total of 467 h

and 5 h had elapsed since it was last inspected The inspection report shows that the parameters measured were within the specified limits The CMB of the other engine had completed about 130

h and was found undamaged There are some previous cases in which an aircraft landed safely after

an indication of a problem in the engine; in all these cases failures in the cages of the CMB were

detected The cage of the CMB from the plane under consideration in this study was retrieved in a

broken condition (Fig Ib)

To isolate the cause of failure, the bcaring components were subjected to detailed analysis The main components of the bearing included the steel balls, inner and outer races and the cage The material of these components was investigated and the details are summarized in Table I The cage material was subjected to detailed metallurgical investigations

The retrieved balls exhibited significant wearlerosion They were studied for smearing of foreign

material Similarly the surfaces of the races were extensively deformed and these were analyzed to

investigate the deformation details and the material smeared on them

2 CMB ASSEMBLY

The bearing is mounted on the compressor shaft and is located in a housing [I] The main

components of the CMB are inner and outer races on which the bearing ball roll and a soft cage to

keep the balls at a distance from each other The arrangement of these components is shown in Fig 2

*Author to whom correspondence should be addressed

Reprinted from Engineering Failure Analysis 5 (4), 26 1-269 (1 998)

416

(b) Fig 1 (a) The plane after the accident (b) Bearing components on the compressor shaft after the accident

Table 1 Chemical composition of the different parts of the CMB

Composition wt.% of components Element Cage Balls Inner race Outer race Bearing lock

1.67 k0.12 0.40 k 0.10 0.25 k 0.04 0.92+0.01 0.001

1 0 2 ~ 0 0 1 0.003

Bal

38 Cr Alt

-

t Closest standard

2.1 After the accident

Examination of the compressor region revealed that the cage was broken and significantly deformed The inner and outer races were put together with the cage and the probable configuration

417

Housing

Inner ra

Fig 2 Sketch of the bearing assembly

of bearing components after mishap is shown in Fig 3a The compressor and then the CMB components were disassembled; the latter were retrieved in a severely distressed condition The condition of the retrieved components was compared with ones which were in normal use in another engine The retrieved cage compared with a new one is shown in Fig 3b A detailed study of the retrieved components was conducted using stereo, optical and electron microscopes to establish the mechanism of damage Detailed views of the sections of interest of each component are discussed

in the following with the help of the micrographs in Fig 4

The worst condition was that of the cage; Fig 4a-i Its outer surface is compared with that of another used cage During the accident the cage was fractured and the results of fractography are discussed in Section 3 The inner surface of the cage, as evident from the crushed hole peripheries

in Fig 4a-ii, had been severely deformed by the hard Cr-steel bearing balls

In Fig 4b-i the inner surfaces of the retrieved outer races are shown where significant deformation and smearing is visible The sections of the edges of a used and of a retrieved outer race are compared

in Fig 4b-ii and b-iii The cross sections of the edges clearly show extensive deshaping indicating high stresses acting on them; this is quite significant on the left-hand side (LHS) edge shown in Fig 4b-iv

The photograph in Fig 4c-i shows delamination of surface layers of the steel inner race It can

be seen, at a higher magnification in Figs 4c-ii and c-iii, that a layer of the cage material was smeared

on the surface and the subsequent deformation was so severe that inner race material covered it up, sandwiching the cage material completely

The view of the left and right edges in Fig 4c-iv was obtained following necessary sectioning of the component The degree of deformation is compared with the edges of another used inner race

in Fig 4c-v A closer look at the edges in Fig 4c-vi shows serious deshaping of the edges It is noted

that LHS edges of the inner and the outer races were much more deformed than the respective RHS

edges

418

The rolling elements, Cr-steel balls, also experienced severe deformation and wear; see Fig 4d-i One of these balls, which was stressed to a relatively small extent was sectioned to study the locations where stresses were overbearing The etched sample in Fig 4d-ii shows a band of the stressed region which is -4 mm wide and generated while the ball was rolling

To summarize the above observations, the inner and one of the outer races of the retrieved bearing were heavily deformed from one side The off-center extensive deformation indicates high stresses operating in these areas prior to the failure The inner and outer races put together with the cage in Fig 3a show the probably configuration in which the cage broke and was trapped between the balls and the races The balls shifted towards one side and severely deformed the races The function of the bearing was impaired resulting in the air crash The observations concerning the deformation and smearing of the components are sketched in Fig 4e to show the possible position and condition of the bearing components just before the accident

2.2 Material of CMB components

ray analysis (EDX), atomic absorption spectroscopy (AAS) and carbon/sulphur (C/S) analyzer

The chemical analysis of different parts of the bearing was carried out using energy dispersive X-

The cage is fabricated from a Cu-A1 alloy; its chemical composition is summarized in Table 1

420

Fig 5 Optical micrograph of the cage and the silver coating

Optical microscopy of the cage shows a dual phase structure of the material (Fig 5) The cage is most probably cast and has a - 38 pm thick silver coating Silver coating provides good resistance

to fretting and improves the bedding-in and running properties of harder bearing materials [2] Microscopic study revealed that there were darkly etched spherical inclusions in the cage material; the inclusions are up to - 8 pm in diameter and are Al-rich

The bearing balls, outer and inner races were manufactured from a Cr-steel in the hardened and tempered condition, having a fine distribution of Cr-carbides; the composition is summarized in Table 1 From the point of wear similar material to the components which are in mutual contact have a tendency to “smear” during the wear process [3]

Material of the bearing lock was found to be 38 Cr AI; Table 1 A section of the lock material observed under the optical microscope revealed - 18 pm thick copper coating on the surface

2.3 Material smeared on components

The material smeared on the surfaces of the components was analyzed following preparation of smaller cross-sectional specimens of the component These were analyzed in the scanning electron microscope equipped with energy dispersive spectroscopy The results are summarized in Tables 2 and 3

The higher magnification view of the bearing ball specimen in Fig 4d-iii shows the presence of deposited material; this is especially true in the regions where the material close to the surface shows heavy deformation bands The smearing, confirmed from EDX analysis, was found rich in Fe, Cu,

Si, A1 and Cr while Ag and Mn were detected at isolated locations Si is probably from the lubricant

or debris At a location on the surface of the ball (location A in Fig 4d-i) where it seemed that material was removed due to impact, a high concentration of Fe, Cr, Si, A1 and Cu was detected

It seems that the material of the cage was smeared probably at the inner race first and then extensive compressive/shear stresses between the ball and the outer race resulted in chipping or shearing of the ball material

Smearing on the outer race is accompanied by deformation, as is evident from the deformation bands near the surface in Fig 6 Three different regions were analyzed and the results are summarized

in Table 2a; the first two regions were big enough to permit analysis at a couple of locations In all

the locations cage material was smeared Region I contained a significant amount of Ag showing

that the cage got smeared when the coating was intact Analysis at Region 111, on the other hand, did not show the presence of an Ag coating, indicating that the cage had already been distressed to the level that its coating was completely stripped off before smearing at Region 111

Micrographs in Fig 4c show the condition of smearing on the inner race The LHS was quite deformed and damaged Figure 4c-i shows the seat of the race where extensive smearing and deformation is clearly visible Four different locations, marked in Figs 4c-vi and c-vii, were analyzed

42 1

Table 2 Smeared material on the outer race

Fig 6 Deformation bands near the surface of the retrieved outer race

and they indicate the presence of smeared cage material at Regions I1 and IV; see Table 2b for the analysis report At Region 11, Ag is also present indicating that when the cage got smeared here the coating was at least partially intact No bearing component was found to contain Cr, Si, A1 and

Mg in the amounts detected in Region I11 of the inner race Particles rich in Si and A1 and containing

Mg were retrieved from the lubricating oil and could be transported to the inner race from some other source through the lubricant

The material protruding from the RHS of the race (Fig 4c-vii) contained more than 7% Cr which

was not found in any of the bearing components It seems that a component of high chromium content was severely damaged and hit the inner bearing race No obvious damage was seen on the inner race at the location of impact

Different portions of the retrieved cage, including the inside surface of the holes, showed smears

of steel; an iron content of up to 71 % was found near the surface regions The retrieved balls showed

422

Cr depletion of up to 1.40% near the surface region; this shows that the balls were running hot for quite some time

3 FRACTOGRAPHY O F THE CAGE FRACTURE STRUCTURE

The fractured surface of the retrieved cage was examined in the SEM (Figs 7a and b) The general features of the surface at low magnification are shown in Fig 7a The typical fatigue striations (Fig 7b) in the central portion of the sample were clearly visible The crack propagating lines lead to the corner marked ‘0’; from where the crack probably initiated The features disappeared near the exposed surface but bending marks could be observed Bending could be due to some foreign body impact, quite possibly during the accident after the failure It is quite clear from the observations that the failure was due to fatigue which was confirmed in the simulated laboratory experiments The latter were necessary to confirm that the striations were not slip bands which are sometimes observed in these materials [4]

Before the accident excessive wear of the bearing components took place in a short time of approximately 5 h Cage material was stuck between the ball and the race producing severe

(b)

Fig 7 SEM photographs of the fractured surface of the retrieved cage: (a) general view; (b) fatigue striations

423

misalignment of the outer race with respect to the compressor shaft The offcentering of the balls exerted overbearing stresses on one side (displaced from the middle) of the inner and outer races inducing heavy deformation on their surfaces The rotation of the bearing and the main compressor shaft became difficult and led to the accident To get in between the balls and the races, the cage had to break The fatigue failure features on the fracture surface of the cage confirmed the above hypothesis [5]

It has been conclusively shown that the cause of the accident was the failure of the CMB cage The cage failed due to fatigue No material defect could be traced at the site of crack initiation

Acknowle&emenrs-The authors are grateful to Mr Badar Habib, Mr Tahir Mehmood and Mr Altaf Hussain for their help

in photography and compilation of the paper

REFERENCES

1 Widner, R L., Metals Handbook, vol I1,9th edn., American Society for Metals, Metals Park, OH 44073, 1986, pp 490-

2 Neal, M J., ed., BearingeA Tribology Handbook, 2nd edn., Butterworth-Heinemann Ltd., Oxford, 1993, pp 97-1 16

3 Kossowskii, R., Emerging Technologies Inc., USA, private communication

4 Walker, C R and Starr, K K., Failure Analysis Handbook, Pratt & Whitney Report August 1989, Materials Laboratory,

5 Widner, R L and Wolfe, J O., Merul Progress, April 1968, pp 52-59

513

130-1 34

Wright Research and Development Center, OH 45433-6533, pp 206,267,272,354,358