The maximum allowable rail load for 100 ton cars in unrestricted interchange service is currently 263,000 lbs. A number of our coal-hauling utility customers have asked us to advise them on the consequences of raising this limit to 286,000 lbs., an increase in load of 8.7%.

Fatigue Life

The first step in our study was to take a look at the increased load on the bearing from the stand-point of "laboratory" fatigue life. More commonly called L10 Life, this is a calculated prediction of bearing life in terms of stress cycles (related to revolutions) based on a 10% failure rate. The subject of fatigue life was addressed previously in Technical Forum 90-1. Please refer to this publication for a detailed description of fatigue life and how it relates to actual service life. The following is offered as a brief review.

Spalling is the flaking of material from the rolling contact surfaces of a bearing. Spalls typically begin as small cracks below the surface of the bearing raceway. These cracks join each other and grow toward the race surface, eventually becoming a small spall. This type of failure is a direct result of repeated stress cycles and is referred to as a fatigue failure. A stress cycle is the application and removal (or reversal) of a load. An example of a common series of stress cycles is the repeated bending of a coat hanger. After several cycles, the coat hanger breaks. In a bearing, each passage of a roller over a point on a raceway, while in the load zone, is considered a stress cycle. While the number of stress cycles a bearing can endure is considerable, there is a finite limit for all but the most lightly loaded bearing applications.

Fatigue life is most heavily influenced by the applied load on the bearing. In the calculations of L10 Life, fatigue life is inversely proportional to the 10/3 power of the load. In other words, half the load does not result in 2 times the life but 2 10/3 or about 10 times the original life. Conversely, twice the load will result in 0.5 10/3 or about 1/10 of the original life. This general relationship can be seen on the graph in Figure 1. Note that this graph uses relative values. In this way, it can be used to compare any load condition having an unknown bearing life to a load where the expected life is known.

To look at the load change in question, let's zoom in to a smaller area of the graph. Figure 2 shows that an 8.7% load increase will result in an approximate 25% reduction in fatigue life, as defined. Because of other types of failures and factors affecting actual service life, this does not mean that there will be exactly a 25% drop in usable service life. It does, however, mean that the higher loading will significantly decrease the average time a bearing is in service before a spall appears. It also suggests that this result will appear more frequently and more readily in reconditioned bearings due to the presence of a variety of acceptable defects (under current AAR standards) and repaired defects which appear to have increased susceptibility to spalling.

Figure 1	Figure 2

The reduction in bearing L10 life is one of the most significant effects of the proposed increase in loading upon bearing life. What is important to bear in mind is that our predictions of increased fatigue failure are based on the performance of a new bearing operating in laboratory conditions. In the real world, we will be applying the heavier loads to a fleet of bearings that have already seen many years of service. While the acceleration of fatigue damage in such a varied population under actual service conditions cannot be predicted with any accuracy, it can most certainly be predicted to occur.

If reduced bearing life were the only consequence of heavier loading, we might not feel so concerned about customers' trading off higher bearing replacement costs for overall productivity gains in their operations. But the effects of the heavier loads on bearing performance raise more serious concerns, in particular the greatly increased fretting of bearing components and the resulting loss of bearing retention that can occur under heavily loaded conditions.

Fretting Wear

Fretting is a type of wear that has the appearance of corrosion. It is the result of relative motion between tightly fitted parts, such as bearing and axle components. These relative motions can lead to surface discoloration, pitting, and eventually, surface failure. (See Figure 3.) For a railroad application, this can mean costly axle damage; within the bearing assembly itself, it can lead directly to loss of the bearing's clamping force.

Figure 3

The major contributor to fretting is bearing operations with high axle deflections (high loads) present. When an axle deflects or bends under a load, the journal's outside surface gets a little longer on the top and a little shorter on the bottom. Depending on many variables, the bearing parts may or may not maintain their original position on the journal. When they do not, sliding occurs between the parts and the journal surface. This sliding occurs on a very small, even microscopic scale and is therefore very difficult to observe directly.

There are two main areas where fretting wear can occur. The first is the point at which the backing ring contacts the journal's fillet. At one time this was the area of greatest concern because it was demonstrated that even slight journal deflections led to the backing ring riding up and down the journal fillet. (See Figure 4.) The adoption of an interference-fit collar on the newer fitted style backing ring has resolved most of the problem when used in conjunction with a "fitted" axle.

Figure 4	Figure 5

The second principal area of fretting occurs in the "nose" area of the inboard seal wear ring. This is where the wear ring's inner diameter fits on the journal. Movement of the stack of bearing components is resisted by the axial (lateral) clamp forces exerted on the stack by the tension in the cap screws, together with the component interference fit. As the bearing is loaded, the journal bends, causing the top surface of the journal to stretch. When the load increases to a point where this axial force is overcome, the wear ring face will separate from the cone face. (See Figure 5.) This has two results. First, the wear ring slides on the journal, producing fretting in this area and damaging the axle by forming a rough groove. Second, as the two faces separate and re-contact, they bump and slide against each other, causing fretting wear on both contact faces. (See Figures 6A and Figures 6B). The face wear reduces the width of the components which results in a reduction of the overall length of the stack of bearing components. By this process, fretting can eventually lead to a total loss of bearing clamp, with the potential for rapid bearing failure with possibly catastrophic consequences. AAR studies have identified that as much as 50% of all bearing failures in rail service are related to loose components.

Figure 6 A & B

Because fretting occurs when heavy loading bends the journal enough to overcome the initial bearing clamp forces, it can be seen that higher clamp loads can withstand higher bending forces while lower clamp loads cannot withstand as much. Theoretically, for each initial clamp force there is a critical load where separation first starts. (See Figure 7.) Below this load, fretting will not occur. Above this load, fretting will almost certainly occur. It is apparent that higher loads will only result in worse fretting. The process is degenerative. Over time, fretting will reduce the part length, which in turn reduces the initial clamp force, with the result that the fretting action will become even more severe.

Figure 7

Trying to determine if and how much fretting is present for a given load and bearing clamp is extremely difficult due to the number of variables involved. What has been established is that heavier loading will cause fretting to occur where none had been before, and any fretting which has been occurring will definitely get worse.

BRENCO's Product Engineering department is currently using computer aided Finite Element Analysis Modeling (FEM) to determine the exact relationship between bearing clamp, axle load, and fretting. The results of these studies of axle and component deflections confirm an expected increase in axle deflection and axle stress with increased loads of 8.7% ( See Figure 8). BRENCO's Performance Engineering group will be supplementing these studies by verifying results on a static test rig under development, using a full wheelset under load. Results of this study will be reported at a later date. Other testing in the area of bearing clamp includes bolt torque studies and sponsored cone bore growth studies.

Figure 8

Conclusions

The findings to date are clear; increased loading is not a free lunch. Reduced equipment life and increased maintenance costs must be carefully weighed in the balance with immediate productivity gains. The two main effects upon bearing life and performance are reduced fatigue life and increased fretting. Fatigue life will drop significantly, somewhere in the order of 25% with a maximum load increase of less than 9%. Of even greater concern in our view is the fact that fretting may start occurring where it has not been seen before, and where it has been previously observed in service, the severity and the number of cases will increase, perhaps as much as 200%.

For those rail users committed to a program of heavier loading, we strongly recommend that they at least take actions to mitigate the effects of the higher loads. It is important to understand that such actions will only partially offset the effects of heavier loading, and will not prevent the higher stress loads from adversely affecting both bearing life and performance. Those actions recommended would include:

• Use fitted axles and new fitted backing rings to reduce fretting motion
• Select cars with new bearings which provide tighter interference fit and longer fatigue life for this type of service
• Specify higher reconditioning standards for those bearings being applied to cars for use in heavier loaded service
• Initiate a program to accelerate the re-torquing of older bearings in service and removal of cap screw seal rings where present (refer to Technical Forums 88-1 and 88-2)
• Use only new cap screws to insure maximum clamp
• Use a higher load capacity bearing, such as the 7 X 12
• Phase in heavier loading after incorporating some or all of the above program maintenance practices

This last recommendation serves to underscore what in our mind is the greatest concern in adopting the higher load limits. That concern, simply stated, is that the increased maximum loads would be implemented overnight, without any of the above precautions. The existing fleets contain many older bearings, bearings with cap screw seal rings, that have been reconditioned several times under AAR minimum standards, or that have non-fitted backing rings or non-fitted axles. The consequences to bearing performance could be severe.

We recognize that in today's competitive world there is always the need to become more efficient. We also understand that in the drive for greater productivity, sometimes immediate gains are achieved at the expense of equipment life and performance. But accelerated wear and more frequent breakdowns may end up increasing maintenance costs and disrupting service to the point where there are no real economic gains. There may even be substantial losses. Until we have a clearer measure of what those costs are likely to be in practice, we believe the wisest action is to proceed with caution.

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