Roller Bearing Capacity

[l8tr]

I'm trying to set up an experiment on a beam in 4-point bending. I would love some advice on the proper method to design the roller (expansion) bearing.

I can't find any recent references explaining the design of these (aside from Hertzian contact modeling, which as I understand it, only give me the contact stresses and not the actual roller capacity). Other references I've found include Roark's formulas (which references some old railroad documents) and give me roller sizes that I think are much larger than what I would expect.

I will have a load of 220kips, bearing length of 10in, and am flexible on steel strength and diameter of the roller.

 

[MiketheEngineer]

sounds like something you have on a highway bridge?? Check that out

 

[gwynn]

Expansion bearings I see on highway bridges are most often of the sliding variety, not rolling. Eg. a teflon/stainless steel interface. I've don't remember if CSA-S6 or AASHTO address roller bearings very well since I have never seen them used. A bridge bearing could work though. An elastomeric pad with two steel plates on top. The plate directly on top of the steel would have teflon bonded to the top and the one above that would have polished stainless bonded to the underside. You would just need to size the elastormeric pad to be able to take the rotations expected. With the railroad reference in the OP, Arema specs may have actual rollers covered better than highway codes.

A better resource for information specific to rollers might be a mechanical engineer or someone like Hilman. The easiest thing might be to just buy a 125 ton NT roller from Hilman (if needed design/fabricate the rotational aspect of the connection youself though, that part is definaely cheaper to do yourself).

 

[ishvaaag]

Old AASHTO at least did ... have not looked for it in current code but it may subsist.

 

[dhengr]

“I'm trying to set up an experiment on a beam in 4-point bending. I would love some advice on the proper method to design the roller (expansion) bearing.” Are you a student?

What’s a beam in 4-point bending? What is the beam made of, its length and size, loads, deflection, curvature and slope at the reactions, how many reaction points, etc. etc? All of these must come into play as you set up your experiment. What are you trying to prove or show? We can’t see any of these things from here, so a few sketches and more explanation might be helpful. Is it reasonable to assume that you will need at least one knife edged reaction (pinned reaction, free to rotate, but not free to translate) and one roller (more free to translate) to comply with our normal approach to beam analysis? Both of these reaction types involve some attention to Hertz bearing stresses. Do you have a “load of 220kips” or a reaction of 220k?

I’m not real familiar with the bridge design codes, but I’ll bet they use rockers instead of simple rollers because they can get a much greater contact radius in a limited height, for the high loads and movements they deal with. And, that will probably be your problem too, but maybe not the thermal movement. There is no way of getting around it, Hertz stress theories pretty much rule on this type problem. The railroads have done a lot of work on this bearing stress problem as relates to wear and long life of steel wheels on steel rails. The roller bearing people have pretty much beat this horse, almost to death, at a different (smaller radius) scale. It always boils down to contact radius, hardness and Fy or Fu of the various materials in contact. AASHTO may have some nice tables and simple formulations for their typical applications, otherwise, you just gotta slog through the problem, depending upon your needs and materials. Hilman rollers distribute the large load and do roll, but not without considerable resistance to translation. So they may not be what you want. They will translate, but will not allow free rotation. They work because they distribute the load to 6 or 8 small radius rollers, but any beam rotation will overload the inner most rollers. And, if this is a repeated application over a small distance, they will likely cause deformations in the plate on which they roll. They will not be appreciably better than one slightly larger roller for your application.