Wear in Rolling Element Bearings and Gears - How Age and Contamination Affect Them

Bruce T. Kuhnell, Monash University
Tags: bearing lubrication

This article summarizes the author’s findings on wear in rolling element bearings and gears. The work disovered that surface cracks can result in pitting only; subsurface cracks result in spalling; and spalling is failure caused by the material collapse at subsurface crack tip ligaments. Also discussed are bearing failure modes due to contamination as well as results of experiments with different filtration.

Plain Bearings
Plain, journal, sleeve or shell bearings, and piston rings and bushes, have conformal interacting surfaces (area contact surfaces) that support the moving parts of machinery with sliding action (on an oil film) at the interfaces. Rolling element bearings have nonconformal contact surfaces (line or point contacts called hertzian contacts) that support moving parts of machinery with rolling action (on an oil film) at the interfaces. Gear drives also have nonconformal interacting surfaces and have both rolling and sliding action (on an oil film) at the tooth interfaces.

If ideal lubrication and operating conditions (such as load ranges, speed ranges, temperature ranges) are maintained, plain bearings will not fail. Lubricant considerations include cleanliness, allowable water content, viscosity, viscosity index improvers, wear protection additives, oxidation inhibitors, friction modifiers, seal swell additives, corrosion inhibitors and antifoam additives.

Even if proper lubrication and operating conditions are scrupulously maintained, rolling element bearings will eventually fail by fatigue because of subsurface voids, cracks or inclusions.

The different geometric nature of plain and rolling bearings creates different lubrication regimes: plain bearings have hydrodynamic (and sometimes hydrostatic) lubrication conditions while rolling element bearings (also cams and gears) operate with elasto-hydrodynamic (EHD) lubrication regimes. The “elasto” term in EHD refers to the effect of the hertzian line or point contacts, which produce extreme pressures and therefore elastically deform the surfaces to provide small elliptical contact areas. It is the repetitive formation of these elastically deformed contacts that eventually leads to surface fatigue (at subsurface defects).

There are many ways that lubrication and/or other operating conditions can be inadequate. The main considerations here are the influence and nature of contamination on bearing degradation.

Plain bearings generally do not seal as well as rolling element bearings. They rely on grooves in the bearing surfaces to supply the correct amount of oil over the surface of the journal to float it on a hydrodynamic film, which is drawn into the load zone by viscous drag (Figure 1).


Figure 1. Hydrodynamic Lubrication

The minimum separation of the journal from the shell is shown as h0 in Figure 1. Plain bearings are lubricated with five to 10 times more lubricant than rolling bearings, using flushing to remove contaminants from the load zone.

Contamination
There are many types of contaminants. These include hard particles (silica), bauxite, wear debris (surface fatigue, cutting wear, flakes), soft particles (fibers, polymers, cellulose), water and corrosive chemicals, which can be ingested or generated within the machine. Hard, abrasive contaminants generally cause three-body sliding, indentation or cutting wear damage to the softer material. Soft babbitt metal coatings are often provided on the shell of plain bearings to protect the journal by allowing generation of a small number of negligible grooves, which do not degrade the bearing function seriously.

The critical particle sizes are those that have the same dimensions as the minimum separation, h0 from Figure 1. Particles less than h0/5 will mostly pass through and particles too large to be dragged into the lubricant entrance wedge will find their way out the ends of the shell. A filter with at least ßh0/5 rating of 75 should be specified if economically feasible. A beta ratio: ßx = 75 means that of 75 particles size x or larger arriving at the filter, only one gets through. It is not a trivial task to estimate h0 and usually involves solving rather complex equations.

Rolling Element Bearings
Even when operating correctly, rolling element bearings will eventually fail as a result of a surface fatigue phenomenon. Rolling element bearing surface fatigue is characterized by spalling. It starts after some variable time of service as embryonic particles that are liberated from the surface of a race or rolling element in the load zone. Surface fatigue leaves craters that act as stress concentration sites. Subsequent contacts at those sites cause progression of the spalling process.

The duration of satisfactory performance depends largely on the durability of bearing surfaces. Generally, there are three types of surface contact damage that can occur under proper operational conditions: surface distress, fatigue pitting, and fatigue spalling. Other surface damage can occur due to improper mounting or improper operating conditions.

Surface distress appears as a smooth surface resulting from plastic deformation in the asperity dimension. This plastic deformation causes a thin work-hardened surface layer (typically less than 10 µm).

Pitting appears as shallow craters at contact surfaces with a depth of, at most, the thickness of the work-hardened layer (approximately l0 µm), as shown in Figure 2.


Figure 2. Pitting and Spalling

Spalling leaves deeper cavities at contact surfaces with a depth of 20 µm to 100 µm as shown in Figure 2. It must be noted here that no common definitions have been established to distinguish spalling from pitting in the literature. In most of the literature, spalling and pitting have been used indiscriminately, and in some other literature, spalling and pitting were used to designate different severities of surface contact fatigue. For instance, Tallian defined “spalling” as macroscale contact fatigue caused by fatigue crack propagation and reserved “pitting” as surface damage caused by sources other than crack propagation.

One of the reasons for the confusing definitions is probably due to the fact that the physical causes of pitting and spalling have not yet been established. To discuss spalling and pitting on a common ground, the following discussion rests on the definitions according to the phenomena as described in the foregoing; that is, pitting is the formation of shallow craters by surface-defect fatigue, and spalling is the formation of deeper cavities by subsurface-defect fatigue.


Figure 3. Well-developed Fatigue
Spalls on Bearing Inner Race

Figure 3 shows an example of advanced fatigue wear. The shaft in this tapered roller bearing was approximately 200 mm in diameter and some of the advanced spalling from multiple sites is 30 mm across.


Figure 4. Typical Spall Crater
(Scale Bar = 400 µm)

Figure 4 shows a large single spall some 250 µm across. Initial spall particles are typically 30 µm to 50 µm, but it is common for several particles to be generated from individual spall sites. Note at the sharper crater wall (near the top edge of the spall in this micrograph) there are several cracks associated with the spall.

Though both spalling and pitting are the common forms of surface contact fatigue, spalling results in more rapid deterioration of surface durability when compared to pitting. Spalling often induces early failure by severe secondary damage. It has been repeatedly reported as the more destructive surface failure mode for gear contacts. Such secondary damage can result in roller or race breakage, initiated from a severe spall on the contact surface, as well as friction- or heat-induced surface seizure, or complete spalling over all of the contact surfaces.

Gear Teeth
The kinematics of an involute gear meshing pair are best described as sliding/rolling contact with pure rolling at the pitch lines.


Figure 5. Gear Tooth Action Between
Two Gears During Their Engagement

As shown in Figure 5, the contact point moves along a straight line DE through the pitch point (P), which is tangent to the base circles. The line DE is called the pressure line because it is the common normal to the contact point for involute teeth. Sliding is maximum at the start of the engagement; toward the gear’s pitch line and away from the pinion’s pitch line. It reduces to zero at the pitch point (pure rolling as in rolling bearings) and then increases again to another maximum at the end of engagement.

Once the contact point moves into the vicinity of the pitch line, the number of gear pairs in contact becomes minimal, so that the normal Hertzian contact load is maximum on this region. Spalling is normally found in this region. Thus there is an intuitive similarity between the rolling bearing spalling and gear pitch line spalling phenomena.


Figure 6. Destructive Pitch-Line Pitting
of Gear Teeth from Subsurface Defect Origin

Pitch line surface fatigue spalling is illustrated in Figure 6. It is common for initial pitch line spalling (usually labeled initial pitting) to occur soon after gear drive commissioning and then stop suddenly and for the gears to continue a long life of service. It is a relief mechanism for small-tooth geometry errors. In some cases, however, more severe geometrical errors can continue, leading to premature tooth cracking and early failure. The drive in Figure 6 has been operating with severe misalignment, which has the same detrimental effect as improper tooth geometry.

Theories of Contact Fatigue
Surface-Defect-Origin. Way’s hypothesis postulated that lubricating oil in a surface crack was trapped when the approaching contact reached the surface opening and pinched the crack closed. As a result, the crack tip was extended by the hydraulic pressure of the oil sealed between the crack surfaces. Subsequent work by Keer and Bryant found that the dominant mechanism for surface-breaking crack growth was Mode II (shear) propagation which contradicts Way’s assumption of Mode I (tension) crack propagation. Bower performed a fracture-mechanics analysis of crack propagation in the presence of lubricating oil. His results do not appear to support Way’s hypothesis, either. Furthermore, the experimental results obtained by Cheng and others showed that the surface crack growth was very slow.

According to Ding and Kuhnell, surface crack growth can only be in Mode II and can result only in shallow craters.

To better understand spalling/pitting mechanisms, many researchers have also studied the behavior of subsurface cracks under contact loads. Fleming and Suh used fracture mechanics methods to analyze the propagation of subsurface cracks parallel to the contact surface. Their results showed that the stress intensity factors (SIFs) for Mode I and Mode II were quite low. Kaneta and others studied the growth mechanism of subsurface cracks by numerically analyzing the behavior of a three-dimensional subsurface crack parallel to the contact surface. They concluded that the propagation of subsurface cracks is mainly by Mode II.

More recently, Ding and others studied the behavior of subsurface cracks beneath the pitch line of a gear tooth, focusing on developing a fundamental understanding of the mechanisms of spalling in gears. Using the finite element method, the potential modes of crack propagation and failure were analyzed and the values of the stress intensity factors (SIFs) of the subsurface cracks were below the critical SIF, Kc. Consequently, ligament collapse at crack tips was hypothesized as the cause of spalling from subsurface cracks. Elastic-plastic finite element analysis was also performed to further evaluate the hypothesis as the failure mechanism of spalling in gears.

According to Ding and Kuhnell, subsurface spalling by crack propagation mechanisms would be too slow. Stress intensity factors for both Mode I and Mode II never exceed the critical stress intensity of crack failure in their study. Therefore, spalling is not caused by crack propagation of subsurface cracks.

Ding and others calculated the mean stress, sm, in a ligament region between the crack tip and the contact surface, and concluded that spalling results from ligament collapse at subsurface crack tips. The angles between the direction of the maximum shear stress and the crack line were 33 degrees at the trailing tip and 53 degrees at the leading tip of the subsurface crack.


Figure 7. Sectioned Micrographs of
Spalling on Gear Teeth Surfaces
Near Pitch Line

Therefore, a spall cavity should have a shallow wall at an angle of approximately 33 degrees at the trailing end and a steep wall of 53 degrees at the leading end of rolling direction. This finding was supported by the results of the experimental evidence as were the spall depth predictions. Figure 7 provides sectioned micrographs of three spall sites. Figure 7a shows a spall site with the material of the potential spall particle(s) still attached. Figure 7b is a spall which has progressed and a number of spall particles have detached. Figure 7c is a cross-section of a spall from which the particle(s) have been liberated. Note the cracks at the steep walls of Figure 7b, Figure 7c and Figure 4. These indicate the readiness for the spalling to continue on subsequent contacts at these sites.

Conclusion
There is still much that is not known about the physical mechanism of surface damage phenomena. The theoretical studies outlined here offer plausible explanations for nonconformal contact fatigue spalling under clean conditions and based on two-dimensional analyses. They did not directly include the effects of lubrication. Extension of the work to three-dimensional analyses and inclusion of lubricant effects still need to be addressed. Nevertheless, the theories are compatible with experimental work both in-house and by others.

References

  1. Tallian, T.E. Failure Atlas for Hertz Contact Machine Elements. New York, N.Y., 1992.
  2. Alban, L.E. Systematic Analysis of Gear Failures. Metals Park, Ohio: ASM International, 1985.
  3. Way, S. “Pitting Due to Rolling Contact.” Journal of Applied Mechanics, Transactions of the ASME, Vol. 2, 1935.
  4. Keer, L.M. and Bryaut, M.D. “A Pitting Model for Rolling Contact Fatigue.” Journal of Lubrication Technology, Transactions of the ASME, Vol. 105, April 1983.
  5. Bower, A.F. “The Influence of Crack Face Friction and Trapped Fluid on Surface-initiated Rolling Contact Fatigue Cracks.” Trans. ASME, JOT, Vol. 110, 1988.
  6. Cheng, H.S., Keer, L.M. and Mura, T. “Analytical Modeling of Surface Pitting in Simulated Gear-Teeth Contacts.” SAE Technical Paper No. 841086, 1984.
  7. Ding, Y. and Kuhnell B.T. “The Physical Cause of Spalling in Gears.” Machine Condition Monitoring, The Research Bulletin of the Centre for Machine Condition Monitoring, Vol. 9. Monash University, 1997.
  8. Fleming, J.F. and Suh, N.P. “Mechanics of Crack Propagation in Delamination Wear.” Wear, 1977.
  9. Kaneta, K., Murakami, Y. and Okazaki, T. “Growth Mechanism of Subsurface Crack Due to Hertzian Contact.” Journal of Tribology, Transactions of the ASME, Vol. 108, 1986.
  10. Sin, H.C. and Suh, N.P. “Subsurface Crack Propagation Due to Surface Traction in Sliding Wear.” Journal of Applied Mechanics, Transactions of the ASME, Vol. 51, June 1984.
  11. Ding. Y., Jones. R. and Kuhnell B.T. “Numerical Analysis of Subsurface Crack Failure Beneath the Pitch Line of a Gear Tooth During Engagement.” Wear, 1995.
  12. Ding, Y., Jones, R. and Kuhnell B.T. “Elastic-Plastic Finite Element Analysis of Spall Formation in Gears.” Wear, 1996.
  13. Ding, Y., Kuhnell. B.T. and Jones, R. “An Experimental Investigation of Gear Spalling Phenomenon.” Surface Modification Technologies X, Editors T.S. Sudarshan, K.A. Khor and M. Jeandin. London: The Institute of Materials, 1997.