Subject: Wheels (3/20/1997)
[9/26/88] Restoration - Westours (David Beagle) and Tour Alaska (Robert Stone) both presented video tours of their respective restored and rebuilt dome cars, which are used regularly on the Alaska Railroad. The amount and extent of work which they have done to these cars was astounding. The amenities provided, and the crew requirements make their dome cars miniature cruise ships.
One interesting phenomenon was the rapid wear out of the wheel sets. Because of the numerous curves and grades, the average wheel set only lasts 70,000 miles. An experiment was made which put flange lubricators on the wheels to reduce wear. Instead, it resulted in increased wear, because the volcanic dust acted as a grinding compound on the wheels.
[10/2/90] Wheel Defects - I have received a bulletin from CSX advising that during 1989-1990 it found an inordinately high number of derailments caused by broken wheels. Upon examination, they found failures were due to cracks and chips in the outer rim area of the wheel tread. Therefore, they have begun a much more rigorous inspection of wheels, and particularly the outer tread area, to try to prevent derailments.
Chips, and broken spots in the outer rim, or in the rolled over section of a rim, are not necessarily an immediate hazard. However, when these defects extend more than 5/8" in-board of the original rim face, or more than 1/8" into the depth of the tread, then they are dangerous, and will be condemned by CSX on the spot. Defective wheels should be removed from service, and turned, or scrapped.
[6/21/94] We have received a 6 page staff report from the NTSB. It is the recommendation of the staff that all passenger car wheels meeting the following criteria be immediately removed from service: The combination of (1) tread brake, (2) rim stamped, and (3) straight plate. No other wheels are affected.
We cannot respond to this proposal until we learn from you the number of such wheels in existence, and the economic impact. If you have wheels of the type described, please fill out the attached survey and return it to Diane Elliott before July 14, along with your comments in writing.
An outline of the report follows. The complete report is available from Larry Haines (224 Orr Drive Somerville, NJ 08876) by sending him $1.00 to cover the cost of copying and postage. The report is well written, persuasive, and thorough.
The Ringling Bros. wheel broke on January 13, 1994, without warning, into three pieces, resulting in 2 deaths and considerable property damage. The train was a mix of freight and passenger equipment. A thermal or fatigue crack developed at a stamped serial number in the rim. The wheel showed no evidence of physical abuse. However, at some time it had been severely overheated.
The wheel had been manufactured in 1972, and had new bearings applied in 1990. It was a class B wheel designed for high speed passenger service with severe braking conditions and heavier loads. A thermally damaged wheel will not fail unless a crack is present. Initial cracks may be invisible. Tensile forces from overheating may open up a crack beyond the critical size. Straight plate wheels are less tolerant to overheating than are curved plate wheels. Straight plate passenger wheels are very common. Effective Jan. 1, 1994, straight plate freight car wheels must be replaced with curve plate wheels when worn out. Rim stamped freight car wheels have not been made since 1978.
Wheel overheating in freight cars has been largely eliminated because of better single car tests, dynamic braking, and the adoption of curved plates. Detection of thermally damaged wheels is a practical impossibility. The wheel in question may have been thermally damaged several years ago, and may have been re-contoured after that time.
The next prior accident caused by a failed passenger car wheel was the Auto Train derailment in 1976, which likewise involved a mix of freight and passenger car equipment. There is no indication in the NTSB proposal as to the type of brake valve involved, nor whether it was in direct or graduated. See associated diagram to distinguish straight and curved plate wheels.
National Transportation Safety Board
Railroad Division (ST-30)
490 L'Enfant Plaza East, S.W.
Washington, D.C. 20594
Paul L. DeVerter II June 14, 1994
Chairman of the Board of Directors
American Association of Private Railroad Car Owners, Inc.
1301 McKinney, Suite 5100
Houston, Texas 77010-3095
Dear Paul,
As a result of the derailment of the Ringling Bros. and Barnum & Bailey Circus blue train near Lakeland, Florida, on January 13, 1994, (DCA-94-MROO3), the NTSB staff is considering supporting a recommendation to the Safety Board that would recommend the immediate removal of all straight-plate rim-stamped tread-braked railroad wheels from service. This would include excursion, tourist, and operating museum railroads, or any other commercial or private passenger car operations regardless of mileage or connection to the general railroad system.
Straight-plate rim-stamped tread-braked railroad wheels are more susceptible to instantaneous catastrophic failure if they have been thermally damaged (overheated) than either hub-stamped straight-plate or curve-plate wheels. This is particularly dangerous since there is currently no reliable and practical way to determine if a wheel has been thermally damaged, especially if the wheel has developed surface corrosion or grime over any telltale discoloration.
The derailment of the circus train illustrates the unpredictability and danger of thermally damaged straight-plate rim-stamped wheels, and the fact that such wheels cannot be reliably detected. I have enclosed a synopsis of the accident and some basic technical wheel information regarding this issue.
I am requesting that you inform your members of a proposed recommendation to remove all straight-plate rim-stamped tread-braked railroad wheels from service; the basis of this proposal, and to write to me individually or collectively of any adverse effect such a recommendation might have to their operation. This feedback would help the NTSB staff gauge the recommendation's possible impact to the industry in light of a safety benefit.
Sincerely, /s/ Russell G. Quimby, Safety Engineer-Rail
(202) 382-6644
FAX (202) 382-6884
[the following synopsis and wheel information was attached to the above letter]
On January 13, 1994, a northbound(note 1) Ringling Bros. Barnum & Bailey (Blue) Circus Train derailed at 9:08 a.m. eastern standard time while passing through Lakeland, Florida on CSXT railroad en route to Orlando, Florida. The derailment occurred at mile post (MP) A849.7 on the CSX Transportation (CSXT) system of the Fertilizer Business Unit(note 2) (FBU), Lakeland Subdivision, on single-line main track in Train Control System (TCS) territory under the control of a CSXT dispatcher in Jacksonville, Florida. The train was operated under contract by a CSXT operating crew and pulled by three CSXT locomotive units. The 53 car train was a mix of passenger and freight cars with 34 passenger cars on the head end followed by 18 89-foot loaded TTX-type flat cars with containers, cages, or truck trailers, and a bi-level auto rack car on the end. About 9:00 a.m., an industrial worker was watching the circus train go by and observed two pieces of a wheel fly off a passenger car and land in nearby woods. The circus train continued 2.8 miles across five grade crossings with the broken wheel and derailed truck of the 23rd car, a passenger car, before 16 circus cars derailed at the Park Spur turnout starting with the 22nd car. Of the 16 derailed cars, five were laid on their side, the rest remained upright. The derailment resulted in two circus employee fatalities. Both employees were found crushed under unsecured equipment that fell on them when the general derailment occurred. Thirteen other circus employees received minor injuries.
According to the CSXT engineer, the train was moving 47 mph at the time of the accident. None of the locomotive units were equipped with speed or event recorders. The circus train was operated by a three man CSXT crew; engineer, conductor, and brakeman. After the accident, the crew was tox tested according to the FRA rules. The brakeman's test was positive for cocaine and several other drugs.
After the derailment, the two wheel pieces that completed the broken L-4 wheel from car 89 were recovered and taken along with the rest of the wheel set to the local CSXT car shop for inspection. The wheel had broken into three pieces; one piece had remained attached to the axle. The wheel parts displayed the classic characteristics of an overheated and subsequently thermally damaged wheel. A subsurface anomaly, "oyster shell," was noted by investigators where the wheel had broken below the deeply stamped "4" of the serial number and along the vertical straight line of the number. The oyster shell pattern is typical of a progressive failure from a fatigue or thermal crack. This is usually due to overheating from a stuck tread brake as a result of a malfunctioning brake valve, a hand brake left applied, or a misapplied composition rather than a cast iron brake shoe, but may also result from heating torch burn.(note 3)
Further inspection revealed that both the accident and mate wheel had good contour and showed no evidence of flat spots or physical abuse such as strike marks. Investigators noted that the wheel set was equipped with Timkin AP Class EE 6 inch Amtrak type roller bearings with locking plates stamped and etched "ASX 12-90 BRX-R," indicating that the bearings were mounted in December 1990 by the American Railway Wheel Services Shop in Sanford, Florida. The roller bearings had remained intact and showed no sign of failure. The axle had a stainless steel band stamped "RB585," and was stenciled in white paint "RB136." These were identification numbers used by Ringling to keep track of wheel sets and work performed by contractors. The complete wheel set was then shipped to the NTSB metallurgical lab in Washington, D.C. for examination and failure analysis.
The back face of the rim of the broken wheel was stamped indicating that the wheel was manufactured in September, 1972, by the Standard Steel Division, serial number 4617, and was a controlled cooled passenger car straight-plate class B wheel that was 36 inches in diameter. The mate wheel was identical to the accident wheel except that its serial number was 4611. According to the AAR Manual of Standards and Recommended Practices, Section A, Part Ill, Passenger Car Requirements, class B wheels are designed for passenger cars in "high speed service with severe braking conditions and heavier wheel loads." The Vice-President of Quality and Technology for Standard Steel said that both the accident and mate wheels are from the same serial set which were part of an original order of "probably 100 or 150 wheels for the Southern Railway." He further stated that he believed that the two wheels had stayed together since being first mounted on the axle, and had never been removed off the axle.
The NTSB laboratory factual report confirmed that the accident wheel was overheated resulting in reverse stress thermal damage, a phenomenon explained later. Although the mate wheel was intact after the accident, it too was found to have been overheated and thermal damaged. The NTSB metallurgist found separation of the accident wheel was the result of fatigue cracking that originated in the bottom of numeral character "4" of the serial number on the gauge face of the rim. Based on the presence of multiple crack origination sites and a relatively small fatigue fracture zone (as compared to the rest of the fracture), both the stress level and stress intensity factor below the "4" were high. Examination of the intact (mate) wheel also revealed a fatigue crack located at the bottom of the long leg in its stamped number "4." Both wheels had evidence of severe overheating of the tread surfaces. Based on the micro structure, some areas exceeded a transformation temperature of about 1330 degrees F. Both the micro structure and hardness of the base material were normal for the specified material.(note 4) However, the material underlying the stamped markings in both wheels was decarburized and had a hardness significantly lower than the hardness range specified by the manufacturer. This reduced the tensile strength at the fracture initiation area in the accident wheel by 35 percent.
Tread-braked railroad wheels act much like automobile brake drums and therefore are subject to heat. During braking, the brake shoe is pressed against the tread of the wheel, which stops or slows the railroad vehicle through friction. This friction between the wheel tread and the brake shoe creates heat. Some of the heat is transferred to the brake shoe, but most of the heat is absorbed by the steel railroad wheel. The wheel in turn gives some of the heat to the rail and some heat is dissipated to air flowing by, but most of the heat is retained by the wheel which acts as a heat sink. If braking continues for very long, the temperature of the wheel tread and rim increase, but the plate and hub of the wheel will remain at a relatively lower temperature. Of coarse, the temperature differential between the rim and the hub is dependent on the duration and and intensity of the brake shoe force. It is this temperature differential between the rim and hub that is significant and creates the destructive stresses that eventually cause the wheel to fail.
The heated rim wants to expand, but is restrained by the cooler plate and hub. Under normal operating conditions this is not a problem and the wheel returns to its manufactured condition of circumferential rim compression which holds the wheel together like a barrel stave. However, under unusually heavy and/or prolonged braking conditions such as a stuck brake, excessive heat is built up (overheating) which causes the wheel rim to expand and deform beyond the steel's elastic limit. This results in permanent deformation and leaves the rim in a state of tension stresses that no longer want to hold the wheel together, but want to pull away from the wheel plate and hub. The mechanical rolling of the wheel and the forces of weight involved contribute to the deformation and the creation of residual tensile stresses. The presence of circumferential residual tensile stresses in the rim defines a thermally damaged wheel.
A thermally damaged wheel will not fail unless a crack is present. Cracks occur in the wheel tread and rim in regular service from a variety of causes including normal braking applications. The tensile forces which result from overheating may open up any radial cracks that may be present in the wheel. The larger the crack, the less residual tensile stress is needed to cause brittle fracture of the wheel; the smaller the crack, the greater the residual stress that is needed to cause brittle fracture of the wheel. Initial radial cracks in a wheel may be so fine as to be invisible to the naked eye and/or ultraviolet magnaflux inspection if corrosion is present. Once the crack grows to a certain point called the critical flaw size, the whole wheel will break instantaneously.
In thermally damaged railroad wheels, the stresses are greatest on the back face of the wheel rim. AAR research confirmed that stamping the back of the wheel rim provides a weak point for a crack to start in a thermally damaged wheel. Consequently, the AAR prohibited the rim stamping of railroad wheels used on interchange-freight cars in 1978. According to the AAR, few if any rim stamped wheels are still in use on interchange-freight cars due to the relatively high attrition rate of freight car wheels. However, rim stamping can still be found on some locomotive wheels, transit cars, and private passenger car wheels like those of the circus train. Generally these type railroad vehicles are as not subject to thermal damage in normal operation as freight car wheels, and are not regularly interchanged like freight cars.
Straight plate wheels are more subject to thermal failure than curved plate wheels. The curved plate wheel design acts much like a thermal expansion joint which tolerates overheating better than the straight plate wheel design and consequently is less prone to thermal failure. This is not to say that curved plate wheels cannot thermally fail, but that they are less subject to thermal failure. Curved plate wheel designs were developed primarily if not exclusively for interchange freight cars which have experienced the greatest failure from thermally damaged wheels, particularly in the 1970's before the large scale adoption of curved plate wheels.(5) As of January 1, 1994, the AAR has prohibited freight car wheel replacement with straight plate wheels; all wheels on freight cars must be replaced with an appropriate curve plate wheel. The interchange freight car fleet of 12 about million wheels is currently about 85 percent curve plate wheels.
Locomotive and passenger car wheels have remained largely a straight plate design because of their relatively low numbers and recurring usage, and because wheel failure due to thermal damage is rare. Also conversion from a straight to curve plate design would, in many cases, interfere with mounting disk brakes or other running gear. Under such circumstances manufacturers are reluctant to invest in the tooling and machinery necessary to manufacture curved plate passenger car wheels when the market is very small and the demand questionable.
The straight plate B36 wheel like that involved in this accident was designed to be used on passenger trains which are; relatively short, have passenger brake valves, and receive frequent attention. These considerations greatly reduce the possibility of overheating a wheel and causing thermal wheel failure. The Safety Board has no record of an accident caused by a thermally failed passenger car wheel since the derailment of the AutoTrain near Jarratt, Virginia, on May 5, 1976.(note 6) Like the circus train in this accident, the AutoTrain was a mix of passenger and freight cars and was much longer than a normal passenger train. However, the wheel involved in that accident was a 33 inch diameter wheel, and there was a mix of passenger and freight brake valves.
Wheel overheating and wheel failure on interchange-freight cars has been largely eliminated due to a number of factors. The new AAR single-car air brake test reduced stuck brakes by more effectively finding defective brake valves which then reduced the number of overheated wheels. The adoption of curved plate wheels which are more heat tolerant reduced the number of thermally failed wheels. The use of disk rather than tread brakes on passenger and transit cars has precluded the possibility of overheating the tread and wheel rim, precluding thermal failure. And finally, the increased use of dynamic braking has also reduced the possibility of overheating wheels by reducing the use of the air brake system.
Detection of thermally damaged wheels is currently a practical impossibility outside a laboratory environment. Even in a laboratory, current non-destructive methods of wheel thermal damage detection are unreliable including magnaflux, magnetic diffusion, acoustic, ultrasonic, and x-ray. Only the destructive cutting of the wheel rim will reveal if the wheel is thermally damaged as any cut will quickly propagate into a crack to the hub if the wheel has thermal stress reversal.
Despite the fact that the Ringling Circus train was inspected at its Tampa origin by CSX mechanical and circus personnel; passed a defect detector 18 miles from the derailment point; passed an observant maintenance-of-way gang who paused to inspect the train; and was stopped and inspected by the train crew 10 miles from the point of derailment, the defective thermally damaged wheel could not be detected and the circus train derailed with two fatalities, many injuries, and great loss of property and equipment.
According to expert witness sworn testimony at the Safety Board depositions regarding the accident, the accident wheel was most probably already damaged when it was installed on the car, maybe having been damaged two or more years previously. It is also likely that the accident wheel was already thermally damaged when it was re-contoured on a truing machine one or more times before it was mounted on the circus car.
(note 1) Northbound by timetable, eastbound by compass direction.
(note 2) This is the same as a conventional railroad division.
(note 3) See figures 3C.9 through 3C.12 and figure 4.2 of the Association of American Railroads Wheel and Axle Manual.
(note 4) Association of American Railroads, Operations and Maintenance Department, Mechanical Division, Specifications M-107-71, Wheels Wrought Carbon Steel.
(note 5) According to the AAR, in 1978 there were 163 derailments from thermally damaged wheels. In 1993 there were 37 derailments, of which 27 involved straight plate wheels.
(note 6) This accident occurred before Amtrak's operation of AutoTrain on the then Seaboard Coast Line Railroad. See NTSB Report number NTSB-RAR-76-11.
[2/6/95] NTSB - I met with the Safety Engineer Russell Quimby at 1:15 and thereafter with the Board Chairman, Jim Hall, at 2 p.m. I explained our plight and the economic burden. Briefly, Pullman ran straight plate, rim stamped, tread braked wheels for almost 100 years without problems. We likewise, have had no problems.
Altogether we reported approximately 800 axles in our member's possession, about 350 of which are mounted on cars. The cost in material for the new wheels, turning axles, and rebuilding bearings will be approximately $3000 per axle. This data was supplied by PFL Industries. The cost in material is thus 800 x $3000 or $2,400,000. This does not include the time and cost of jacking the car, and changing out the wheels. In addition, there are no conforming wheels being made, and none expected until 1996.
Apparently, nobody plans to make wheels for 5-1/2" axles or smaller, in any event. This will strand a number of cars, unless they can afford to have new journal boxes made. I also learned today that both the AutoTrain wreck 20 years ago and the Ringling Bros. wreck involved cars which had been modified from cast iron to composition shoes, and had ABD or other freight valves changed from D-22. This may, or may not have been a factor in the overheated wheels - composition shoes do not dissipate nearly as much heat as cast iron. And, freight brake valves do not ordinarily receive the attention that passenger valves do.
I also argued that we have never had an accident or incident on our cars involving an overheated stress cracked wheel, that the real problem is not the wheels but stuck brake valves, and that we take excellent care of our brake valves. I also pointed out that Amtrak has no problem with us using up the metal on wheels already under our cars, but does not agree to mounting spare wheels. (A letter to this effect is on file with the NTSB from John Hutchison, Assistant Chief Mechanical Officer of Amtrak). This is the same action taken by the AAR when the rim stamped, straight plate freight car wheels were discontinued.
The meeting was cordial, and I was assured our comments would be considered. The staff has recommended to the Board that the wheels be banned from Amtrak and from interchange. The decision is now up to the Board.
Being a political animal, Amtrak will likely do whatever the NTSB recommends, even though it is only a recommendation and not a rule or law.
For cars not run on Amtrak, or not interchanged, and run below 30 mph, the staff recommends sanding the rim stamped area and then dye checking that area for stress cracks once a year. We could live with this, except for the speed restriction, although it would be a burden, and we could even coat the wheel plate or rims with a temperature sensitive paint to detect future overheating.
[2/6/95] AMTRAK CAR INSPECTOR CARDS - Everett Fullerton of the CHAPEL HILL discovered that Amtrak had issued laminated pocket cards (dated January 1, 1995) to all its car inspectors. These cards set out the minimum standards for running in Amtrak trains, and also somewhat higher standards at initial terminals. I am attempting to get a card for each active member, and if I cannot, then I will reproduce them next month.
The change which strikes the hardest is that the wheel tread must be 1-1/8" minimum at the initial terminal. This is 1/8" greater than our governing SMP 28603, which makes the AAR standard applicable. The upshot would be to condemn wheels which, according to the SMP and AAR standard, have more than 20,000 miles remaining. Our average member probably runs less than 4000 miles/year.
Since the Amtrak standards tend to become the de facto FRA standards, I have asked John Hutchison to get us an exception. This Amtrak rule (for Amtrak cars) is a real waste to us. The AAR standard for passenger car wheels is 1", and thus we will have to continually explain that 1-1/8" is only the rule for Amtrak's own equipment, and 1" applies on freight trains and non-Amtrak cars. I have suggested a letter for inspectors to be carried in the electric locker. Mr. Hutchison is conferring with the FRA, and will advise.
[2/15/95] Diane Elliott attended the NTSB hearing on Feb.14 where the staff made its formal report, and the Board adopted its recommendations. This is a summary of the bad news, and she will detail the long version in another document:
1. To the FRA - prohibit replacement of rim stamped, straight plate, tread braked wheels.
2. To AAR - develop inspection procedures, and in meantime prohibit interchange of these wheels.
3. To Amtrak - prohibit use of these wheels.
4. To AAPRCO, TRAIN, NRHS, ASLRA, RPCA, Museums, etc. - inform your members of these recommendations.
5. To same - remove corrosion, and periodically inspect wheels.
6. To Ringling Bros. - remove wheels, secure interior furnishings, install emergency brake valves, install intercom.
[3/20/97] Wheel Life - Many car owners have asked for information on what they should expect in terms of miles for the life of a set of passenger car wheels. The answer is - it depends. If the car is run on good track, without excessive braking, and not too many curves, then you should expect over 300,000 miles.
A recent article in Railway Age explored the problem in terms of the Los Angeles County Metropolitan Transportation Authority's Red Line. The anticipated life, in transit service was 300,000 miles. However, because of worn rail profiles, out of tram trucks, poor tracking characteristics of the trucks, and the like, their new wheels were only averaging 6000 miles between flange truing, and 19,000 miles total life. If you notice rapid wear, then you know you have a problem, and you need to get to work on the problem.
Paul L. DeVerter II
Comments