Hollow Shaft/tubing design with thrust bearing, Fillet size/design

[Philtech]

I have a 10,000 to 100,000 lb of weight (thrust) that is subject to changing tension loads on the tubing string (maybe 27000 lb going back and forth to 33000 lb every 6 sec to create over 1 million cycles every 2 months). The, tubing string is attached to a larger diameter hollow shaft or tubing (Same Internal Diameter) such that the larger diameter shaft provides the ledge that sits on a thrust bearing (Vertical thrust). The smaller tubing is 3.14 inches in diameter and the larger tubing/ledge is 4.5 inches to fit over the bearing.
QUESTION: For never fail fatigue design, Is it better to have a radius fillet to connect the two tubing shafts and then use a bearing to fit over the horizontal part of the larger shaft (bearing ID greater than the radius fillet OD), or to use a gradual radius to a 30 degree ledge on the larger shaft and then use a matching 30 degree ledge washer to transfer the weight of the tubing to the horizontal cyclindrical thrust bearing. SInce the force is a thrust force but caused by tubing weight, does the thickness of the ledge matter above the bearing?
THANKS for any input.

Picture: (probably no help, but picture on left has larger upper hollow shaft directly on horizontal bearing race B which puts bending forces farther out from the OD of the lower shaft)(Picture on right uses a 30 deg wedge taper in lieu of just a radius to connect the larger upper shaft (2.44 in I, 4.5 OD) to the lower hollow shaft (2.44 in ID. 3.14 OD), which then sits on a special washer (W) that is tapered to fit into the fillet and 30 degree taper connecting the upper and lower shaft). This would move the average bending forces due to weight being exerted on the bearing closer to the OD of the lower shaft and thereby possibly help reduce fatigue failures)????



! ! or ! !
! ! ! !
! ! ! !
--. .-- . .
bb . . bb w . .w
. . bb . .bb

 

[ccw]

Philtech,
I think you have the right idea with the taper design. You want to keep the stress lines from bunching around sharp changes in cross section as much as possible to avoid fatigue problems. With 100,000 lb on the small shaft you are already at 32,600 psi peak stress, no stress concentration applied. Your proposed shape is practically the same shape shown as an example in some of the mech. design texts, Spotts, Shigley, etc. I am not sure how you are generating the joint between the large and small tubes, machining a transition section, or welding. If you are welding, you should get the joint(s) radiographed, specify full penetration, melt through welds, minimum suck back.

 

[Philtech]

Thanks to ccw. I am not welding but maching a solid piece to allow better strength and less chance for corrosion attacking the weld area. I am looking into the tapered ledge, but have to be careful to have a large enough washer that will not split due to the horizontal forces (from the wedge shape versus a horizontal washer). I am also adjusting with stress concentration factors, which is hard to do since the Peterson Book has a hollow cyclinder but with only pure axial forces (even though the ledge in my design will have some bending stresses also). I also am looking at adapting the stress concentration factors for a t-head design to simulate a tubing split open and put together for calculation purposes.

Thanks for your input.