TESTING FOR CURRENT FREIGHT

CAR BRAKING RATIOS ON

UNION PACIFIC RAILROAD

For Presentation at the 92 nd Annual Technical Conference

of the Air Brake Association, Chicago, IL,
September 19, 2000

BY:

Todd Snyder, Ph. D. Cameron Lonsdale

Manager of Technical Services Technical Manager-Railway Products
Union Pacific Railroad Standard Steel
Omaha, NE Burnham, PA

David S. McConnell, P. E. James G. Rees, P. E.

Senior Manager-R&D Managing Director
Union Pacific Railroad Inter Swiss, Ltd.
Omaha, NE Chicago, IL
 TESTING FOR CURRENT FREIGHT
CAR BRAKING RATIOS ON
UNION PACIFIC RAILROAD
Todd Snyder, Ph. D. Cameron Lonsdale
Manager of Technical Services Technical Manager-Railway Products
Union Pacific Railroad Standard Steel
Omaha, NE Burnham, PA

David S. McConnell, P. E. James G. Rees, P. E.

Senior Manager-R&D Managing Director
Union Pacific Railroad Inter Swiss, Ltd.
Omaha, NE Chicago, IL

ABSTRACT

This paper describes the static dynamometer brake force (Jim Shoe) testing performed
on freight cars at a car repair facility of the Union Pacific Railroad. Several freight cars
were selected for testing, and cars were given a complete air brake system test. The
current hand-braking ratio, loaded braking ratio and empty braking ratio values were
determined for each car and the values were then compared to established Association of
American Railroads (AAR) braking ratio requirements for newly constructed and rebuilt
cars. Brake related wheel defects are reviewed and the implications of current freight car
braking ratios are discussed with regard to wheel removals and safety related issues.

INTRODUCTION

Railroads in North America remove large numbers of wheels each year for brake related
wheel defects such as shelling (AAR why made code 75), tread built-up (why made code
76) and tread slid flat (why made code 78). Car Repair Billing (CRB) data from 1999 show
that 80,298 wheels were removed for shelling, 27,018 wheels were taken out of service for
built-up tread and 9,131 wheels were removed for slid flats (Ref. 1). It is well known that
shelling has increased dramatically in recent years and represented 15.5% of CRB wheel
removals in 1999. Built-up tread and slid flat wheel removal percentages have remained in
the 2% and 5-6% range, respectively, for the last several years. Out-of-round wheels (why
made code 67) constituted another 2,554 wheel removals (0.5% of total) in 1999 (Ref. 1).

Since AAR CRB data represent only a portion of total North American wheel removals, the
true totals are even higher. System wheel repairs and wheel repairs performed by most
private car shops are not included in AAR CRB data. Therefore, AAR CRB figures should
be multiplied by a factor of 2.5 to obtain an estimate of the total number of North American
wheel removals. These removals represent an extremely large, and in the authors
opinion, partially avoidable cost to rail carriers. Authors from AAR/TTCI recently estimated
the annual cost of wheel shelling to be $180 million, excluding costs for lost service time
for equipment and the impact on customer service (Ref. 2). Efforts to reduce the numbers
of wheels being removed for brake related defects will free capital for other projects and
will improve safety by eliminating high impact wheels with tread damage from service.
 BACKGROUND - BRAKE RELATED WHEEL DEFECTS

Wheel shelling is condemnable under AAR rules according to the following definition in the
2000 AAR Field Manual (Ref. 3):

When the shell or spall is inch in diameter or larger and the

shells or spalls are more or less continuous around the periphery
of the wheel or whenever any shell or spall is 1 inch or more in
diameter the wheel must be removed from service. Islands of
original tread surface metal contained in the shell or spall will not
be considered as part of the area of shell or spall.

This AAR definition is accompanied by an illustration in Rule 41 showing that the above
dimensions apply to a circle and showing what constitutes a condemnable shelling defect.
In recent years there has been some confusion regarding the rejectable shell defect size
and the AAR thus made clarifications. A circular letter in June of 1999 indicated that the
dimensions applied to a square defect (Ref. 4).

True wheel shelling is a rolling contact fatigue phenomenon that leads to damage on the
wheel tread and eventually small pieces of the wheel tread break off. Thus, true shelling is
not related to braking systems or braking ratio, and is limited to certain specific heavy-haul
service lanes. A complete description of shelling, which is described in several references
including some listed here, is beyond the scope of this paper (Ref. 5-9). It is now generally
accepted that most wheel shelling in North America (why made code 75) is actually wheel
spalling.

Thermal mechanical shelling, known to exist in some service lanes, is a process that
requires elevated temperatures in conjunction with contact stresses (Ref. 10). In thermal
mechanical shelling, the fatigue cracking is a result of elevated temperatures (which
reduces the strength of the tread material) and the high contact stresses, which would do
no damage if the wheel tread was at its room temperature strength. Spheroidized pearlite
at the tread surface, resulting from brake shoe heating, can be an important clue when
looking for thermal mechanical shelling. No martensite is formed during thermal
mechanical shelling and the cracks initiate at the surface. Often, these cracks are found
on wheels with heat checks and the cracks appear to be closely related. Note that in the
case of spalling, the crack network is either perpendicular or parallel to the surface (Ref.8).

Wheel spalling occurs in service after the wheel slides on the rail and patches of
martensite are formed on the tread. The wheel slide generates very high temperatures at
the contact patch and the steel is austenitized. However, the large heat sink of the
remaining cold wheel quickly quenches the small tread patch and untempered, hard, brittle
martensite is then formed. The combined effects of the localized contact patch heating
during the slide (compressive upset plastic deformation) and the 4% volume increase due
to the martensitic transformation result in a tensile stress field around the martensite patch.
Subsequent in-service loading leads to eventual cracking and fracture of pieces from the
wheel tread. Due to the brittle nature of martensite and the stress concentrations formed
when pieces come off the tread, wheel spalling can result in severe progressive damage to
the wheel. The appearance of shelling, spalling and thermal mechanical shelling can be
very similar, particularly if the spalling extends around the wheel tread from repeated
sliding and skidding, but spalling can often be corroborated by slip marks/patches on the
wheel tread. A number of excellent papers have been written on the subject of wheel
spalling and some are listed as references here (Ref. 7-13).

For built-up tread (why made code 76), the AAR (Ref. 3) specifies that A wheel is
condemnable whenever the tread has built-up metal 1/8 inch or higher on the wheel tread.
This defect is the responsibility of the car owner. Slid flat wheels (why made code 78) are
the responsibility of the handling line and are condemnable if the flat spot is two or more
inches in length or there are two or more adjoining spots each 1-1/2 inch or over in
length. Clearly, these defects can cause high impact loads and damage to wheels, rails,
lading and the freight car. As for out-of-round wheels (why made code 67), Canadian
National found that out-of-round wheels with healed shells could produce impact loads up
to 199 kips (Ref. 14). To be condemnable under AAR rules, out-of-round wheels must
register at least 90,000 pounds on a wheel impact load detector and have a verified out-of-
round runout of 0.070 inches (Ref. 3).

AAR BRAKE SHOE FORCE TESTS

AAR requirements for freight car air brakes are outlined in S-401-99 (Ref. 15). Cars that
are new, rebuilt or converted from cast iron brake shoes to composition brake shoes must
have braking ratios in accordance with this document. The relevant net braking ratios for
cars with high friction composition brake shoes, non-lever hand brakes, and a 30 psi brake
pipe reduction from 90 psi brake pipe pressure are shown in Table 1.

Car Type Loaded Net Braking Ratio Maximum Empty Minimum Hand
Minimum Maximum Braking Ratio Braking Ratio
TOFC/COFC 11% 13% 38% 10%
All Other 8.5% 13%
Table 1. Net braking ratios required for new and rebuilt freight cars (Ref. 15).

Braking ratio is defined as the ratio of braking force against a cars wheels to the cars
weight (Ref. 16). The empty braking ratio is defined as the braking force divided by the
cars light weight. Wheel sliding occurs when a car is empty and this phenomenon is
responsible for creation of tread defects such as spalling. For the loaded braking ratio and
hand-braking ratio, the gross rail load (GRL) of the freight car is used in calculations.

PREVIOUS BRAKING RATIO INVESTIGATIONS

A number of studies have been conducted to determine the root cause of brake related
wheel defects. Although certain past efforts have primarily focused upon details of the air
brake system itself (Ref. 17, 18), or have discussed improper use of handbrakes
(Ref. 19, 20) causing brake related wheel defects, this paper deals primarily with the issue
of braking ratios on equipment currently in service.

Bartley of Canadian Pacific (Ref. 9) reported that the amount of wheel spalling is inversely
proportional to car weight. Janshego and Lonsdale (Ref. 17) reported that two coal cars
with wheel shelling/spalling problems had a current empty braking percentage greater than
the level required by AAR for newly constructed or overhauled cars. Further, Butler,
Lonsdale and Luke (Ref. 18) found that five out of seven steel Conrail coal cars had a
current empty braking ratio greater than the 38% level required by AAR for new cars.
These cars had lost considerable weight from their as constructed weight. Weight
changes can occur due to corrosion, a change in wheel weight, or from a modification to
the car such as different bottom-dump gates. The two of seven cars that were below the
38% empty braking ratio level were within three percent and four percent of the 38% limit,
respectively. Also, five of the seven cars did not have a sufficient loaded braking ratio.

More recent testing found that although four older steel coal cars had an acceptable
current empty braking ratio (below 38%), two of the four cars were within three percent of
the 38% AAR specified maximum, and all four were above 30% (Ref. 21). Also, all four of
the cars had a loaded braking ratio below the minimum 8.5% specified for new/overhauled
cars, and one car had an insufficient hand-braking ratio. Although the latter two braking
ratios do not contribute to wheel damage, they are clearly important from a safety point of
view.

WHEEL SLIDING AND BRAKING RATIO

A wheel will slide on a rail when the retarding force between wheel and brake shoe is
greater than the force between the wheel and rail (adhesive force) (Ref. 16). The retarding
force is a function of brake shoe force and friction between the shoe and wheel tread. If a
car is empty, there is less adhesive force between wheel and rail and the tendency for
wheel slide increases, particularly if the car has lost weight. Also, if the amount of braking
force delivered to the wheel tread increases, wheel slide is more likely. Other factors
affecting the adhesive force include rail condition and lubrication, wheel/rail profiles, etc.

A paper presented at the 1999 Air Brake Association Conference suggested that the
maximum requirement for empty braking ratio should be lowered significantly (Ref. 22).
This paper describes brake retarding force, adhesion demand and available rolling
adhesion. The author states that it is necessary to keep adhesion demand ([coefficient of
friction] X [net brake shoe force/car weight]) less than the available adhesion between
wheel tread and rail to prevent wheel sliding. The author discusses a conservative
approach to minimize wheel sliding where the design net braking ratio is made low enough
to keep the adhesion demand below the available wheel/rail adhesion - even when the rail
is wet. Therefore, the maximum effective (emergency) net braking ratio must be below
28% to prevent wheel sliding. However this means that the empty car design net braking
ratio must not be more than 23.5%. The author recommends that the range of empty car
net braking ratios be between 22% and 30%. Note that the authors suggested 30%
maximum empty net braking ratio is significantly lower than the 38% maximum now
specified by the AAR. The paper also has tables of net braking ratios for different empty
car weights and empty/load proportioning (60%, 50% and 40%) showing safe empty car
net braking ratios to prevent wheel sliding. As empty car weights decrease and loaded car
net braking ratios increase, additional empty/load brake proportioning is needed.
 DE SOTO TESTING RESULTS AND DISCUSSION

The authors recognize that the static dynamometer brake shoe test outlined in AAR S-401
is not designed to be used for equipment currently in service and that there is no AAR or
railroad requirement to perform testing on such cars. However, due to potential changes
in braking systems over time in service, the test was applied to used equipment. The goal
was to determine if current empty braking ratios exceed the 38% maximum empty braking
ratio standard for newly constructed cars, and to relate this to the creation of wheel
defects. Additional information would also be gathered with respect to loaded braking ratio
and hand braking ratio.

A total of nine cars were selected for static dynamometer brake shoe testing (using Jim
Shoe devices) at the Union Pacific De Soto, MO freight car shop on May 30-31, 2000.
Four of the cars were steel hoppers; four were aluminum hoppers; and one car was a coil
steel trough car. All cars had a foundation air brake system except for the coil steel car
that had truck mounted air brakes. Car information is contained in Table 2 and results of
the static dynamometer brake shoe force testing are contained in Table 3.

Car Car Car Car Date Empty/ Gross Rail Listed Empty
Initials Number Type Builder Built Load? Load (lb.) Weight (lb.)
CTRN 373 Steel Hopper Beth. Johnstown 5/79 No 263,000 61,200
CTRN 935 Steel Hopper Pullman Std. 2/75 No 263,000 60,200
CTRN 461 Steel Hopper Pullman Std. 1/76 No 263,000 60,100
CTRN 648 Steel Hopper Pullman Std. 1/76 No 263,000 60,000
MCHX 30237 Alum. Hopper Beth. Johnstown 9/91 Yes 286,000 44,000
MCHX 30528 Alum. Hopper Beth. Johnstown 9/91 Yes 286,000 44,300
MCHX 30004 Alum. Hopper Beth. Johnstown 5/91 Yes 286,000 44,000
MCHX 30483 Alum. Hopper Beth. Johnstown 9/91 Yes 286,000 44,100
CNW 39778 Coil Steel Car ---- 7/68 No 263,000 63,400
Table 2. Car information from De Soto, MO testing.

Car Original Light Current Empty Current Loaded Current Hand

Number Weight (lb.) Brake % Brake % Brake %*
CTRN 373 61,200 44.0 10.2 9.6
CTRN 935 60,200 46.2 10.6 10.4
CTRN 461 60,100 38.6 8.8 13.3
CTRN 648 60,000 42.4 9.7 11.9
MCHX 30237 44,000 33.3 8.0 11.3
MCHX 30528 44,300 31.2 8.0 11.3
MCHX 30004 44,000 32.2 7.7 10.6
MCHX 30483 44,100 31.8 8.5 11.2
CNW 39778 63,400 41.6 10.0 7.9
*Note: Due to a testing error all current hand brake percentages were obtained by extrapolation of applied force.
Table 3. Results of De Soto, MO testing.

All four of the steel hoppers and the coil steel car have a current empty braking ratio that
exceeds the AAR new/overhauled car requirement of 38%. Therefore, wheels on these
cars are more likely to slide in service when the car is empty and brakes are applied. All of
the steel hoppers and the coil steel car have acceptable current loaded braking ratios.
Note that three of the four aluminum hoppers do not have the required minimum loaded
braking ratio (8.5%) for new equipment, while the fourth car just makes the minimum. All
four aluminum hoppers have acceptable empty braking ratios, thus showing that the
empty/load device was functioning properly during the testing.

During testing to find the hand-braking ratio, an error was made and an incorrect force
value was applied for each cars chain. Therefore, the corrected hand-braking ratio was
obtained by extrapolation using the proper chain force value. Note that two of the seven
cars (one steel hopper and the coil car) have a current hand-braking ratio less than the
10% minimum required for new cars in AAR S-401.

A visual examination of wheel treads showed that MCHX 30237 had the worst tread
condition of the aluminum cars. Evidence of skidding and sliding was noted. The CTRN
461 had the worst wheel tread condition of the steel hoppers with many non-condemnable
spalls. If wheels on these nine cars are periodically inspected in the future, progressive
tread damage could be noted.

COMMENTS ON CAR REPAIR RECORDS

We examined UP car repair records for the nine tested UP owned cars to see if the cars
have received a greater than normal number of mechanical component replacements in
the past. A review of repair records for the nine cars showed a strong correlation between
empty braking ratios and changes of brake shoes and wheel sets. Pertinent data are listed
in Table 4. Note that there were more than 2-1/2 times as many brake shoes used and 3
times the brake-related wheel defects on the four CTRN (high empty braking ratio) cars as
compared to the four MCHX cars.

Car Current Brake Wheel Sets Number of Other Notable Data --- (all cars tested with
Number Empty Shoes Shelled or Previous brake cylinder tap prior to Jim Shoe
Brake % Changed Slid-Flat Air Tests testing at DeSoto, MO)
CTRN 373 44.0 39 2 2
CTRN 935 46.2 60 1 2 ABDS Emergency Portion Replaced 6/1/00
CTRN 461 38.6 30 1 2
CTRN 648 42.4 72 5 5
CTRN TOTAL 201 9 11
MCHX 30237 33.3 9 1 1 ABDX Service Portion Replaced 6/1/00
MCHX 30528 31.2 19 0 1 ABDX Service Portion Replaced 6/1/00
MCHX 30004 32.2 18 1 2 ABDX Service Portion Replaced 6/1/00
MCHX 30483 31.8 31 1 2 ABDXR Emergency Portion Replaced 6/1/00
MCHX TOTAL 77 3 6
CNW 39778 41.6 7 0 5
Table 4. Car repair billing data (6/1/95-6/1/00).

It must be noted that the entire build series of UP owned CTRN cars (581 total) and MCHX
cars (550 total) were later compared for the total number of shelled and slid-flat wheel
removals. The MCHX cars were billed for more than twice as many wheel defects (997)
as the CTRN cars were (432 wheel defects).

We also note that five of the nine cars tested at De Soto received a new air brake valve
portion following completion of an air brake test that included a brake cylinder pressure
test. Three of the cars that received new valve portions had each been air tested twice
before while the other two cars that received new valve portions had each been air tested
once before. However, none of these older air brake tests consisted of the brake cylinder
pressure test. This information, and the data shown in Tables 2, 3 and 4, leaves much
room for discussion regarding the variables affecting wheel and brake shoe usage.

Over the last five years, UP series cars have experienced 126,243 single-car and repair
track air brake tests which has resulted in the replacement or repair of 13,559 emergency
and service valve portions (10.7%). Since billing began for air brake testing with the brake
cylinder pressure gauge (July 1998), 936 tests on UP series cars have resulted in 215
service and emergency portion repairs (23.2%). This evidence indicates that the brake
cylinder pressure test finds defective brake valve portions that would not have been
removed without that test. This evidence also suggests that leakage into the brake
cylinder is a significant contributing factor to wheel problems.

CONCLUDING REMARKS

Static dynamometer brake force testing using the Jim Shoe device is useful for
determining the current braking ratios for equipment in service. Such testing can find
equipment that has an excessive empty braking ratio, a low loaded braking ratio or a hand
braking ratio that is too low. Further testing of additional car types and a review of
mechanical repair records will allow railroads to find freight cars with braking problems and
will allow for savings of maintenance dollars.

Development of a new static dynamometer brake force test device that is easier to use
(particularly one where brake shoes do not need to be removed) will make field testing
easier and will allow for problem cars to be found. We recommend that efforts continue
to develop such a device.

REFERENCES

1. Association of American Railroads, 1999 Car Repair Billing Data, AAR, Pueblo, CO.

2. D. H. Stone, K. Sawley, D. Kelly and W. Schust, Wheel/Rail Materials and Interaction:

North American Heavy Haul Practices, Proceedings of the International Heavy Haul
Association STS-Conference, Moscow, Russia, 1999, pp. 155-168.

3. AAR 2000 Field Manual of the Interchange Rules, Association of American Railroads,
Washington, D. C., 2000.

4. Association of American Railroads, Circular Letter c-9014, June 3, 1999.

5. J. Kalousek, J. H. Hornaday, Jr., and W. N. Caldwell, "Wheel Shelling Problems on the

Canadian National Railways' British Columbia Northline," Proceedings of the 9th
International Wheelset Congress, Montreal, PQ, Canada, 1988, Paper 5-2.

6. Eric Magel and Joe Kalousek, "Identifying and Interpreting Railway Wheel Defects,"
Proceedings of the International Conference on Freight Car Trucks/Bogies, Montreal,
PQ, Canada, June 9-12, 1996, pp. 5.7-5.20.
7. E. Magel and J. Kalousek, "Martensite and Contact Fatigue Initiated Wheel Defects,"
Proceedings of the 12th International Wheelset Congress, Qingdao, China, September
21-25, 1998, pp. 100-111.

8. D. H. Stone, G. J. Moyar, and T. S. Guins, "An Interpretive Review of Railway Wheel

Spalling and Shelling," ASME RTD Vol. 5, 1992, pp. 97-103.

9. G. W. Bartley, "A Practical View of Wheel Tread Shelling," Proceedings of the 9th
International Wheelset Congress, Montreal, PQ, Canada, 1988, Paper 5-1.

10. M. T. Gallagher, M. A. Polzin, H. R. Wetenkamp, Increased Loading of 36 Wheels

Thermal and Mechanical Considerations, Presentation to the Car Department Officers
Association Annual Meeting, Chicago, IL, September 17, 1991.

11. H. C. Iwand, D. H. Stone, and G. J. Moyar, "A Thermal and Metallurgical Analysis of
Martensite Formation and Tread Spalling During Wheel Skid," ASME RTD Vol. 5, 1992,
pp. 105-116.

12. Kevin J. Sawley, "Railway Wheel Slide Damage," Proceedings of the Engineering
Against Fatigue Conference, University of Sheffield, England, March, 1997.

13. Jian Sun et al., "Progress in the Reduction of Wheel Spalling," Proceedings of the 12th
International Wheelset Congress, Qingdao, China, September 21-25, 1998, pp. 18-29.

14. E. Clegg and W. G. Blevins, Wheel Impact Load Detector Experience on CN,
American Railway Engineering Association Bulletin, October 1996, pp. 499-523.

15. Association of American Railroads, Standard S-401 Freight Car Brake Design
Requirements, Effective February 1, 1999.

16. The Air Brake Association, Management of Train Operation and Train Handling,
Chicago, IL, The Air Brake Association, 1972.

17. S. W. Janshego, II and C. P. Lonsdale, "Conrail Use of Over-the-road Testing to

Investigate the Cause of Wheel Shelling on Cars in Coal Service," Proceedings of the
89th Annual Convention and Technical Conference of the Air Brake Association,
Chicago, IL, September 16, 1997, pp. 187-200.

18. S. L. Butler, C. P. Lonsdale and S. T. Luke, "Conrail Use of Brake Cylinder Pressure
Testing to Investigate Brake Related Wheel Defects," Proceedings of the 90th Annual
Convention and Technical Conference of the Air Brake Association, Chicago, IL,
September 16, 1998, pp. 64-78.

19. W. G. Blevins, "CN Wheel Spalling and Shelling," Proceedings of the 90th Annual
Convention and Technical Conference of the Air Brake Association, Chicago, IL,
September 16, 1998, pp. 98-118.
20. H. Zakaib and K. Arrey, Disproportionate Changeout of Wheelsets at the B End of
Cars Equipped With Truck-Mounted Direct-Acting Brakes, Proceedings of the 69th
Annual Convention and Technical Conference of the Air Brake Association, September
20, 1977, pp. 158-175.

21. C. P. Lonsdale, Mark Lowe, S. T. Luke and J. G. Rees, Field Determination of Braking
Ratios for Cars in Coal Service, Presented at the 91st Annual Convention and
Technical Conference of the Air Brake Association, Chicago, IL, September 21, 1999.

22. Edward W. Gaughan, Optimum Net Braking Ratios for Light Weight, High Capacity
Cars, Presentation to the Air Brake Association Annual Technical Conference,
Chicago, IL September 21, 1999.