By Definition, Spalling is Fatigue Failure
First, it is important to understand that by definition, spalling represents rolling contact surface failure. Whether caused by an inclusion in the steel, impact fracture, repeated cyclic high stress loads, degradation of the lubricant, or any number of other causes, the fact of the matter is that the load bearing surface has failed. Does this mean that the spall will propagate and grow, leading to overheating of the bearing? Obviously that depends on the underlying cause of the spalling, as well as on the continued load-induced stresses the bearing must undergo at that point in the raceway. AAR testing has conclusively demonstrated that many spalls do in fact grow in continued service.
Spalling is the Primary Cause for Component Replacement
Spalling is the primary cause for component replacement in the reconditioning process, and a leading cause - second only to loose components - for bearings being set out in service. With the intensive attention being given in the industry today to loose bearings - including AAR rule changes to remove the cap screw seal rings, and many owners remedial programs to check cap screw torques in the field - bearing retention has been greatly improved, improving bearing performance on the one hand, but at the same time opening the way for spalling to now become the number one cause of bearing failure.
What could accelerate this trend? The recent increase in the permissible load limits for 100-ton equipment will almost certainly have a major impact. Studies by both U.S. bearing manufacturers concur that increasing the load limits to 286,000 pounds GRL will effectively reduce the relative fatigue life of the 6½ X 12 journal bearings by 25%. As a consequence, we can expect spalling of the bearing raceways to occur sooner and more extensively in heavy haul applications.
The second condition that will accelerate the role of spalling in bearing performance is the age of the bearing fleet. A large preponderance of the freight car bearings currently in service were manufactured during the freight-car building boom of the 1970's. Because of the great surplus of used bearings that developed during the 1980's, relatively few new bearings were introduced into North American fleets over the past thirteen years. An age profile of the bearings we are seeing today in our bearing reconditioning operations is shown in Figure 1 and reflects the rapidly aging condition of the bearing fleet. More importantly, a high percentage of these older bearings are approaching the end of their service life. Many of these bearings have been reconditioned several times, with a high percentage containing numerous repaired spalls and other raceway defects.
In view of these trends, the effects of spalling upon bearing performance, safety, service life and maintenance cost, should be of great concern to all railcar owners. Spalling failure of raceway components is almost certain to become the major cause of bearing failure in the 1990's.
Scope of this Discussion
This Technical Forum will:
- illustrate the modes of spalling, or surface fatigue;
- consider the consequences of spalling propagation in service; and
- suggest how railroads and car owners can realize a longer, more reliable service life for
their bearings.
Raceway Fatigue Modes
The several modes of rolling contact fatigue are usually categorized as either surface or subsurface in origin, and can usually be distinguished be post-failure inspection of the surface or by fractographic examination. Two or more modes of contact fatigue can act simultaneously to initiate and propagate cracks.
Role of Lubrication
Lubrication plays a vital role in extending fatigue life. The relatively thick elastohydrodynamic (EHD) films provided by an effective bearing lubricant create the ideal rolling contact conditions of an antifriction bearing by separating the components from actual metal-to-metal contact. The rollers rotate on this film of oil, enhancing the anti-frictional function of the bearing. When fatigue finally occurs in these ideal conditions, it is usually the result of a subsurface inclusion.
It also becomes apparent from this analysis that thin lubricant films permit the occurrence of real surface interactions and as a consequence, surface-origin fatigue. Because the thinner films cannot compensate for surface irregularities, they promote the rapid propagation of all modes of contact fatigue.
Subsurface Fatigue
As described in previous Technical Forums (90-1 and 91-1), subsurface fatigue initiates in the high shear stress range below the surface at non-metallic inclusions and then propagates to the surface, usually forming semi-ellipsoidal spalls such as shown in Figures 2A - 2E.
|
Figure 2A: Subsurface fatigue cracks forming in the area of maximum shear stress. Magnification 200X. |
|
Figure 2B: Photomicrograph of subsurface fatigue crack in ineer race. Magnification 50X |
Surface Fatigue
The several modes of surface-origin fatigue include point surface-origin, geometric stress concentration and various forms of microspalling. These modes generally occur when the roller/raceway interaction is not pure rolling, but contains a sliding component which causes surface stresses that initiate surface cracks.
|
Figure 2C: Photomicrograph of subsurface face crack in raceway at location of crack front. Magnification 100X |
|
Figure 2D: Photomicrograph of subsurface face fatigue crack in raceway after cracking hads progressed such that top layer has been lost. Magnification 100X |
|
Figure 2E: Fatigue spall on inner raceway. (Scale 1 division - 1/16") |
The distinguishing feature of point surface-origin fatigue is damage generally located at stress risers on the surface due to dents, scratches, or surface carbides. The resulting spalls have an arrowhead shape with the arrow pointing in the direction from which the rolling contact approaches.
(See Figure 3)
|
Figure 3: Surface-initiated spalling showing distinct arrow-shaped spalls.
|
In the geometric stress concentration mode the fatigue damage has multiple origins at geometric discontinuities, such as at the end of a misaligned contact or an area of superficial pitting. Multiple adjacent pits are usually seen along the edge of the contact area - at the end of line contact. However, the resulting spalls in this mode can vary greatly in appearance, and they usually are indicative of high load levels above the recommended load.
Microspalling includes peeling, frosting and glazing. It is superficial fatigue damage caused by surface asperity interaction and is dependent on bearing surface roughness and lubricant film thickness. An example of microspalling is shown in Figure 4
|
Figure 4: Advanced microspalling (rough areas around major spalls) due to debris from primary subsurface spalls.
|
In comparing the frequencies of the different modes of spalling, it is worth pointing out that with the advent of cleaner, higher quality steels, surface-originated fatigue will become the most common form of spalling.
Propagation of Spalls
The nature of - and the rate of propagation of - rolling contact fatigue depends on the mode and origin of the spalling. Inclusion-caused subsurface fatigue involves the branching of cracks around the inclusions, with propagation often being most rapid in the direction perpendicular to rolling direction. Point surface origin cracks begin at the surface and propagate rapidly, both circumferentially and laterally. It should be obvious that once the structure of the load bearing surface has been impaired, the cyclic stresses will only serve to extend the damage. The mating rolling element surface also experiences these increased stresses and debris damage and can be expected to suffer earlier fatigue as well.
Role of Surface Macro and Micro Geometry
It may already be apparent that the condition of the rolling element surfaces - both the macro-geometry (the overall profiles or radii of roller/race contact) and the micro-geometry (surface finish and irregularities) - play important roles in determining both the origin and the propagation of rolling contact fatigue. The basic contacting geometry is, of course, determined by the manufacturer, but some aspects of the micro-geometry and raceway integrity are affected by service and repair practices. For example, significant indentations of the raceways (such as at roller brinells and indentations) can serve as the origin of subsequent contact fatigue development, as shown in Figure 5A & Figure 5B, since they create local stress concentrations and also cause irregularities in the lubrication film.
|
Figure 5A: Brinelling of cup raceway.
|
|
Figure 5B: Cross section of brinelled raceaway showing brinell only in top section and brinell plus surface cracking in bottom section. |
In a similar manner, raceway handling nicks, scratches and bruises or indentations from debris can cause stress concentrations that lead to point surface-origin spalling. Larger defects such as voids or holes in raceways that are left by the repair of spalls, for example, will themselves cause stress increases (as noted in Technical Forum 90-1) that can reinitiate spalling (Figure 6A & Figure 6B).
|
Figure 6A: Subsurface fatigue spalling which propogated from prior repaired spall.
|
|
Figure 6B: Close -up of repaired spall showing new spall growing from the repaired spall area. |
Raceway Fatigue Propagation in Service
Rolling contact fatigue damage of any mode will propagate in service as the result of several mechanisms. Branching cracks from hard, subsurface inclusions will propagate generally parallel to the surface, driven by the rolling-induced, maximum reversing shear and aided by more favorable residual stress orientation on these planes. When the subsurface cracks grow large enough, their presence disrupts the rolling contact stress field and will lead to a chipping or flaking off of material.
Once a subsurface crack breaks through to the surface (or a surface originated crack occurs) further propagation into and under the surface is promoted by the hydrodynamic pressure created in the entrained lubricant by the approach and over rolling of contacting surfaces. An ASME technical paper by Hoeprich in 1991 reports laboratory studies of the rate and extent of fatigue damage propagation, or spalled area growth in bearings, beyond the usual laboratory failure size. When the original spall and environmental conditions create a suitable stress situation, the secondary forms of propagation are initiated. Guidelines for examining the resulting spalled surface texture can help identify the original fatigue source or cause.
A number of load and lubricant film thickness conditions, too complex to detail here, are found to influence the rate of lateral and circumferential spall propagation over the raceways.
The conditions that cause shallow microspalling are the same as those that cause point surface-origin spalling. Sometimes the deeper point surface-origin spalling mode has been observed to develop from microspalling.
In addition to the mechanical effect of a repaired spall in increasing the local contact stress, as mentioned above, the presence of such a hole in the raceway will lead to local lubricant film collapse and an increase in detrimental real surface asperity contact around the hole. This can cause microspalling or point surface-origin fatigue. In addition, if all the branching subsurface cracks were not removed by the repair, further propagation in service can be expected to occur. (Figure 6A & Figure 6B)
Service Life After Fatigue Occurs
There is a vast difference between the extremely small fatigue failure that defines bearing L10 life in laboratory testing and the gross spalling that determines the end of bearing life in actual railroad service.
Lets discuss the various limits to rolling contact fatigue that occur in North American service for journal roller bearings. First is the AAR limit for repairable spalling. The AAR Roller Bearing Manual defines the largest acceptable repaired spall in a race (no spall is acceptable on a roller), as well as the maximum number per raceway and the nearest proximity of two repaired spalls. Careful inspection for cracks and rounding of edges of ground-out spalls are required. Under AAR rules, only the original bearing manufacturer can reapply the cage if a cone assembly is disassembled to repair cone race spalls.
The next practical spalling limit in operation would be the development of sufficient spalling to cause the bearing to be removed from service. The removal might result from a detectable increase in bearing temperature or noise picked up be wayside detectors. No dimensional or spalled area criteria have yet been established for this limit, but it might reasonably be called the end of useful life.
Finally, gross raceway and roller spalling can contribute, in the extreme, to final bearing failure and seizure of the bearing while in service. When gross raceway and roller spalling occurs, there may be sufficient debris of large enough size to interfere with the normal rolling action, leading to roller locking and cage failure, sliding, seizure and burn-off. Alternatively, the debris may contaminate the grease, leading to a loss of the lubricant's effectiveness. Metal particles in the grease may also damage the elastomer seals, resulting in loss of lubricant and contributing to thermal instability.
There are obviously many uncertainties about the modes of rolling contact fatigue and service conditions that make it extremely difficult to predict the initiation of spalling, the rate of its propagation, or even its participation in final bearing failure. Nevertheless, we know that spalling conditions represent fatiguing of the critical load-bearing surfaces and adversely affect both bearing performance and life. In such a critical application as a moving freight car, relying on spalled components - especially in heavy haul service - presents unquestionable risks.
Anyone experienced in bearing reconditioning operations has at one time or another seen bearings come out of service under routine maintenance with the raceways heavily spalled. It is astonishing to think that those bearings were performing normally, without apparent distress, in this condition. Yet we know in these instances that the anti-frictional properties of the bearing have been seriously impaired and that bearings typically overheat and eventually fail in such condition, sometimes catastrophically if not caught in time by a hot box detector.
In 1988 at the AAR's test facilities in Pueblo, Colorado, an instrumented bearing with known defects (heavily spalled rollers) experienced thermal runaway with the outer race temperature increasing 290º F. in 54 seconds. The increased friction was observed to cause bearing seizure (in which the cup turned in the adapter under the loaded car). These results were published by the ASME Rail Transportation Division in 1989 (ASME Book No. H00578, pages 33-37). It has therefore been demonstrated, contrary to some opinion that spalling does not lead to catastrophic failure, that bearings with spalled conditions can overheat and enter the final failure mode in very rapid progression.
Because the raceways are almost invariably destroyed in bearings that have overheated to the point of burn off, accurate statistics on the number of burnoffs caused by spalled components are not available. Nevertheless, the bearing manufacturers recognize the importance of the anti-frictional, load-bearing surfaces to the performance of their bearing products and continue to add millions of dollars annually to the cost of their products by using only the cleanest steels and expensive grinding techniques to perfect raceway and roller geometry and finishes. This enormous investment in quality is for the sole purpose of ensuring that their bearings perform reliably under load without undergoing premature fatigue failure.
The current practice in the rail industry is to rely upon the reconditioner to sort out spalled components according to AAR minimum standards under the reconditioning rules. As pointed out earlier, AAR minimum reconditioning standards allow raceways with a variety of spalled conditions to be returned to service, and permit certain spalls to be repaired with a hand grinder. This is in stark contrast to the high standards of quality we insist upon from the bearing manufacturers, for whom even the slightest imperfection in their product is cause for rejection. Yet the position that some in the industry have taken with respect to reconditioned bearings is that acceptable defective components are quite sufficient to ensure reliable performance for the full shell life, even in heavy haul service.
This apparent inconsistency is nowhere more striking than in the industry's present practices with respect to heavy haul service. Originally, when the proposed maximum load limit for 100-ton cars was increased to 286,000 lbs. GRL, recommendations coming out of the AAR technical staff called for only new bearings to be used on this equipment. However, as a practical matter, even new cars constructed specifically for the higher load limits and equipped with new bearings, routinely have wheel sets removed, and to this date the AAR has taken no action with respect to requiring either a new bearing or a premium reconditioned bearing as the appropriate replacement for any bearings removed. Consequently, current AAR minimum reconditioning standards are still the rule even for heavy haul service. Bearings with acceptable defects for general service are also acceptable for 286,000 lbs. axle loads, in spite of the greatly accelerated fatigue life predicted for this type of service.
Surveying the bearing population as a whole, it is reassuring to see that actions taken by the AAR and the rail industry with respect to improving bearing retention have already had a very favorable impact on bearing performance, as well as on reducing axle fretting wear damage. However, the large number of reported bearing set outs and early change outs for Why Made Code 04 (Defects - Internal Parts) clearly indicate that spalled components are a continuing threat to reliable bearing performance. This is further supported by the growing body of data demonstrating the superior performance of new bearings in their initial application, as compared to reconditioned product which has been returned to service with known defects.
The Car Repair Billing (CRB) data for the years 1990-1992 show that after Why Made Code 33 (Derailment Damage), Why Made Code 04 is the largest bearing related cause for bearing removal from service. Spalling is the most prevalent defect type in Why Made Code 04 removals. During this time period, the CRB data indicate that more the 235,000 bearings were removed for Why Made Code 04. At an estimated minimum average cost of $1,000 per incident, this translates into a cost in excess of $235,000,000 incurred by the rail industry in this short time frame for these removals. When you factor in the much higher costs of in-service set outs and derailments - not to mention the disruption in service - the cost to the industry of Why Made Code 04 defects is staggering. In an environment of increased loads, higher utilization, and an aging bearing population, these trends are likely only to grow more pronounced, and more costly.
Conclusions
In view of the aging condition of the bearing fleet and the high service demands of today's railroads, what can be done to improve bearing performance? While it is true that significant gains have been achieved through improved bearing retention, little has been done to address the spalling issue.
The only truly effective, long-term solution to spalling related performance problems is to eliminate spalls altogether in bearings returned to service. As a partial solution, we should reduce the size and frequency of spalls permitted in the reconditioning specifications. It is important to understand that this does not require AAR action. Within his own maintenance requirements, the car owner can specify any one of several premium reconditioning specifications that eliminate or reduce the extent of acceptable spalling. A set of reconditioning specifications that permits no spalling or defects in raceway components whatsoever has in fact been adopted by many coal-hauling utilities and at least one high mileage intermodal carrier. This is probably the best choice for unit coal train service, intermodal, and other high mileage, heavy service applications. New bearings are, of course, always an alternative. But, with defect-free reconditioning options now widely available in the industry, new bearings are not necessarily the best economic choice. After all, in order to achieve performance, what matters is the condition of the bearing, not whether it has ever been reconditioned.
Without such an approach to upgrade the condition of bearings going back into service, an aging bearing fleet is certain to face deteriorating performance in an environment of greatly increasing service demands. In today's intensely competitive transportation market, the costs of such poor performance are more than extremely onerous, they can be prohibitive for any carrier committed to providing efficient, dependable service to his customers.
The Technical Forum is an information resource for the rail industry and is provided as a courtesy of Amsted Rail Group. Suggestions, inquiries or comments are welcomed and should be directed to:
Editor, Technical Forum
BRENCO, Incorporated
P.O. Box 389
Petersburg, Virginia 23804
804-863-1713
BACK TO TECH SHEETS
|
