Dry rolling contact life prediction 3

[BobBartlett]

I have a metal on metal dry rolling contact application. An example of what I am talking about is a railroad wheel on a steel rail.

We have built several of these machines over the years, and I have some historical data from them, but I want to know if I am being overly conservative.

I am looking for DEMONSTRATED CRITERIA for design related to fatigue life or wear rate based on (for example)such properties as yield or tensile strength; radius of each member; width of each member, etc...?

I am familiar with the Hertian stress equations, but they give an equivalent stress that I desire to correlate to predicted life.

I suspect that if some sort of fatigue life curve were constructed that I would have "wear" on one end that transitions to "spalling" on the other end.

 

[dhengr]

Bob:
You’re not asking for much, are you? The railroad wheel on rail is a good example and I’ve been watching that progress for forty years now. I certainly don’t claim to be an expert on that one topic, just used a lot of wheels and rail and seen a lot of failures of both. There are so many variables in this type of problem that I doubt that you will find any general curves, or such like, that can then be applied to your specific problem. History, testing, intelligent record keeping and investigation, and good engineering judgement and experience, and materials and their treatment knowledge are really all you have to go by. Your question is not unlike the development of an S-N curve for any material or detail. You start testing, watching and plotting, and in 50 years you’ll have some historic basis for a better decision. Otherwise, you have some history on your own product and that’s a place to start. You can start shaving pennies off its cost, until it fails to perform as you want, then you back up a bit. Generally, for small quantities nobody can damn you for being slightly conservative. They won’t remember that they asked you to save a nickle, but will blame you for the first customer complaint.

 

[BobBartlett]

Do you have any data on railroad wheels regarding number of miles that a wheel is expected to operate? I can go measure a wheel and back out some data points.

A fellow I knew once said, "One data point is worth a thousand consultants."

 

[dhengr]

Bob:
Unless you are actually dealing with a railcar wheels on class 1 railroad rails and traffic, this info. will only be of general educational value for your own problem. The various railroads all have their favorite specs. and suppliers, and are sometimes vary opinionated, without very thorough technical knowledge over the whole range of issues, but pretty good record keeping and price/cost info. The AAR and the suppliers have the better industry wide testing, technical knowledge and history; and of course, the supplier data must at least be looked at for any company, methods or materials, bias. They all have warehouses full of data on this subject, there for the gleaning. It’s a very important and expensive issue for all of them, with the AAR pretty much acting as the neutral (least biased), but entire industry wide arbiter, through its Mech. Dept. committees, labs and testing programs. The AAR has reams of specs. and testing on this wheel/rail wear and life problem. But, again, unless your problem is specifically railroad wheel on rail, the data may not interpolate very well to another problem. It is very good for general knowledge on the subject, but then, so might be, the general study of metallurgy, heat treating, wear, fatigue, fracture and the like. All the parties have spent years playing with material mechanical properties, hardenability vs. brittleness and toughness, tread and rail head shape and contour, best manufacturing methods (cast vs. wrought), territory and geometry on which they are used, type of railcar and load conditions under use, and on and on. You don’t have to go measure any wheels, the wheel rim shape and thickness limits, wheel diameter limits, allowable load limits for a given wheel dia. are all in the AAR specs., the real trick is all the other variables.

The study of ball bearings, vs. cylindrical or spherical bearings and race design and their life and wear may be of general instructive value too. As might be any literature or design specs. on any type of rolling mill rollers. The Hertz stress problem, both stress level and load patch shape and size, are not insignificant issues in these types of problems, but they are only one of a dozen important variables. So, while a good indicator of fatigue life and durability, they may not translate in one simple step to a solution. A better explanation of your actual problem might be helpful in furthering the discussion. Again, if you are only making a few of these per year, and it ain’t broke, you may not want to try to fix it.

 

[BobBartlett]

DH;

Do you have access to the "warehouses full of data on this subject, there for the gleaning?"

(I am hoping that this is a yes/no)

 

[dhengr]

No immediate access. And, not very current, I would really have to dig for what I do have on the matter. You would have to go to the wheel suppliers or the AAR Mech. Dept. and see what you could get from them. You certainly have to be more explicit and descriptive about what you are doing and what you need, and I’m sure some of the people doing that work could be helpful. They could at least tell you how applicable they thought their work might be to what you need.

 

[GregLocock]

You mentioned in passing roller bearings, which are also a model for your failure modes. Pick up a Timken engineering catalog for their numbers, admittedly on rather special steels/processes.

Assuming you do actually have some failures have you the beginnings of a weibull chart?

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[MintJulep]

Do you have brake shoes riding on your wheels? Railroad wheels do.

Do you have the thermal cycling associated with brake shoes riding on wheels?

Do you have 1 or 2% creep in your rolling contact like railroad wheels do when braking, or when they are powered as on EMUs?

Do you have impact loads like railroad wheels do at track joints, frogs etc.?

If the answer to any of these questions is "no", then data based on railroad wheels is probably not relevant.

 

[BobBartlett]

MintJulep;

Thanks so much for your reply.

If I...
don't have brake shoes
have no thermal cycling
have no braking/traction forces
have no impact loads (like going through turn-outs)

Then would you expect that railroad wheel data applied to my project would be conservative?

 

[Medeski]

BobBartlett,

I'm not sure how much this will help, but you may want to look at pages 179-182 of the Demag catalog. There is some information on service life and loading conditions for similar applications.

http://www.demagcranes.us/Pool/Common_documents/20335244_070330.pdf

Best,
Medeski

 

[dhengr]

I would say it’s conservative, and reasonable, if what you are dealing with is a std. 36" AAR wheel on a std. AAR rail, and under normal AAR running conditions. Otherwise, maybe not. We can’t see what your project is from here, and you won’t tell us, despite being asked. And, since you are being so secretive about every detail of what you are trying to doing, but keep looking for some magic numbers to run with, it is probably not too conservative. I can only assume you are studying a std. AAR wheel/rail situation. If so, say so, please splain. A 12" wheel will be different than an 18" wheel, as will a two flanged wheel on a lighter rail. Two 12" dia. by 36" long cylinders running against each other will be even differenter. All of the things MintJ lists are covered in the AAR testing and studies. But, geometry and other variables are such important considerations that it may be dangerous to call it an AAR wheel if it ain’t a AAR wheel.

 

[BobBartlett]

Dear Mr. Medeski,

Thanks. I will have to review that for its valuable information, but the rollers on a Demag crane base are a similar application. I should be able to do a little reverse engineering.

Mr. DH,

I am sorry that I cannot give you more details, but in an effort to assuage your curiosity I can say that it is a flat wheel on a flat tyre. The wheels are different sizes and loads for different units, and then we occasionally have customers that ask if they can increase the loading on the unit that they already have.

Having said that, I followed your advice and went to the AAR website. If they did any wheel testing it would be valuable to me. Have you ever actually seen that data to be able to tell me what it contains, because I am going to have to pay AAR to access their data (based on everything else about their site).

Or might you be able to direct me to a friend of yours with a company that designs/makes railroad wheels?

 

[dhengr]

Bob:
With the little detail you have given, ‘flat rimmed wheels, with flat tyres,’ and ‘similar to crane wheels,’ there will be a lot of reverse engineering btwn. your application and railroad wheels. How are the tyres mounted to the wheels? That’s a whole new wrinkle to the problem. Railroads gave up on that many years ago. Some of the AAR and supplier testing may still be of general interest for its background, but railroad wheels sounds like a much different set of design conditions and running environment than you have.

Most of the people I knew at the railroads, the AAR and the suppliers are managing/running rocking chairs these days, if not pushin up daisies. Those mentors, of mine, are long retired. And, at the rate railroads and suppliers have consolidated I don’t even know what names some of the companies might be operating under now. The AAR will have stayed fairly constant in its activities and charter, but they went through a significant reorganization some years ago too, and again, the people will have changed. I wouldn’t use the internet, I would call their offices in Wash., D.C. You want to talk to the Mech. Div., Dir. of Tech. Committees, maybe the Chairman of the Wheel & Axle Committee or some such, for some direction as to who to talk to. Much of the general info. is in the public domain, so try your local library or the Mech. Dept. of your serving railroad. For starters, you want Section G, in several volumes, the Wheel and Axle Manual for gen. info. on wheels, and go from there. You do have to be a member or an affiliate member of the AAR to gain access to much of the better, more detailed, material. Again, maybe some of the Mech. people at your serving railroad could get some of it for you. ASME used to have a Rail Transportation Div. which people from many of the suppliers and railroads belonged to. As I recall much of the early research and thinking on the subject came out of that ASME group, from the Univ. of Ill., Ill. Institute of Tech., in connection with the AAR Test Center in Chicago; and later from the Pueblo Test Center. Try your own wheel supplier or crane wheel suppliers in general, and have them do some hunting for you. As for railroad wheel suppliers: for Cast wheels; Abex Corp., Griffin Wheel, Garrett Railroad Car Parts, Unity Railway Supply Co.: for Wrought wheels; Bethlehem Steel, Standard Steel, Canadian Steel Wheel, and others, and maybe under new names, but all those plants, and the demand, didn’t go away, so they are still out there.

I think what you’ll find is that at the calculation level, the Hertz stress is almost everything; so that load (a narrow load range) and wheel diameter for a given rail and wheel geometry can be reasonably related. More specifically, the AAR determined/decided on a fairly narrow range of max. principal stress differences, which leads to a max. shear stress range, which is the determining factor in the failure of the metal. Then they could improve/raise that stress range, by playing with material chemistry, metallurgy, manufacturing methods, heat treating the rims, etc. But, the down side was the running environment in all its variations, and the ever changing geometry of the rail head and the wheel rim and flange. It sounds like your speed is quite different, and your flat wheel and flat tyre are vastly different. You may want to contour your tyre or rail in some way to define the contact area and to confine that load area so as to limit high edge stress problems.

You are on a real fishing expedition here, where we don’t know what kinda fish you want or what you plan to do with them. At this stage I’d have to do the same digging and searching you’ll have to do for meaningful and truly applicable info. I don’t have any specific names up my sleeve.
Good Luck