Buckling of rod with varying diameter (stepped)

[dubc4]

I would like to know the best method of calculating critical buckling force on a rod of varying diameters. I have a hydraulic cylinder rod that steps down to a smaller diameter (undercut) for the threaded end. Previously we have had rod failures at this undercut.

Currently I'm using the smallest diameter with the overall length of the rod (in Euler's buckling formula) to determine a conservative critical force. I'm not sure if this is the best approach.

Any suggestions would be appreciated.

Thank you

 

[rb1957]

see attached ...

another day in paradise, or is paradise one day closer ?

 

[desertfox]

Hi

Have a look at the file Rb1957 posted, I believe that's where your problems lie, if this is a flange ss rb1957? Suggested then each time the mechanism works then you will encounter a bending moment at the threaded end of the rod, I also believe that if its a pin connection shown in that orientation then again there will be a bending moment on the thread rod.

 

[rb1957]

i thought it was an adjustment fttg (ie the section through a flange of a fttg that sits on the end of the tube)

another day in paradise, or is paradise one day closer ?

 

[dhengr]

Dubc4:
Your’s is an interesting problem. But, I’m always amazed at how little problem/solution info. OP’ers. seem to think is needed to have a meaningful discussion of the problem. Then, I’m also puzzled at how they will let a bunch of smart people who are trying to help analyze the problem waste their time guessing at almost everything about the problem, while the OP’er. is too damn lazy to provide the needed info. We can now see basically how the system works, but your sketch falls way short on dimensions, sizes, weights, forces, etc., etc., all needed to start to understand the problem. Then it would also be nice to see the old (failed) scheme along side your new design, to study the differences; to see photos of the failed pieces; fully dimensioned details of the rod end where the failure has occured. You have these numbers and this info., show it. You’d be surprised what an experienced engineer can glean from a complete set of info. and a better set of sketches.

I think Jboggs pretty well nailed the problem in his 4DEC14, 23:43 post. However, I’d suggest the moment is based on a simple beam span from the lower pin on the lever arm on the left, to the clevis pin on the base of the cylinder, on the right. The beam has two basically uniform loads on it; on the left is the rod weight, and on the right is the cylinder and misc. hardware weight. That’ll get you a simplified moment near rod end threads and machined step. Then there are shock loads, axial loads and fatigue at play here too, and likely some dynamic whipping or flexing (a buckling type deflection) of the long rod with each cycle. His comments are spot on otherwise.

 

[desertfox]

Hi

If the joint at the cylinder Clovis is made a pin joint and orientated the same as the other joints, then there will be no bending because all the members become two force members.

 

[dhengr]

Desertfox:
But, if you put a pin at the threaded rod end of the cylinder, you have a three hinged mechanism, and an unstable axial force transmitting structure. Cylinder extension will as likely just force the axial load system out of a straight line, most likely hanging downward. One leg of the triangle will be the cylinder, the second leg will be the hollow rod, and the third will be a straight line btwn. the lower lever pin on the upper left and the rear clevis pin on the back of the cylinder in the OP’ers. sketch.

 

[desertfox]

hi dhengr

That's a very good point you've made, I think we can get round that by moving the rear cylinder pivot onto the cylinder body which will limit its angular movement, however it would need a loci plot to determine exactly where the pivot needs placing.

 

[berkshire]

Maybe I am looking at that Model wrong, but to me, it appears to operate the shut off valve in tension. But the open position for the valve appears to have a hard stop that will put pressure on the rod in compression. Could this stop be moved to the top of the cylinder to avoid loading up that rod?
B.E.

You are judged not by what you know, but by what you can do.

 

[dubc4]

Thanks again for the responses.

First, I would like to address dhengr's message that states I am "too damn lazy" to provide needed info. I'm providing as much information as I can, and I've asked for help instead of giving all of the numbers and asking someone to do the work for me. There has been meaningful discussion here from very smart people that has prompted me to look at the problem from a different perspective. I do not have access to the old, failed part, nor do I have access to any material (picture/videos/reports) that depict the failure. The old system was similar, but different enough that it could be very heavily loaded in both compression and tension.

I've attached another picture to show the assembly from a different angle just to be clear what is happening.

The "flange" that people are referring to is just a threaded insert that is welded onto the extension rod. So the extension rod threads onto the cylinder rod and is secured into position with a jam nut.

I'm interested in the idea of moving the rear cylinder pivot onto the cylinder body, but I'm a little lost as to what this might look like.

Berkshire, you are correct that the shutoff valve operates in tension. However, when it is opened, the pin opening is assisted by the pressure of the material in the housing. There is still hyd pressure applied in the cylinder to move the rod, and if the lever were to seize mid stroke, the rod would be in compression. The stroke of the cylinder is currently such that the end stop is actually the piston bottoming out in the cylinder (which apparently is also not ideal).

 

[desertfox]

Hi dubc4

Look at pages 20 and 38 on this link you will see trunnions mounted onto the cylinder body

http://www.boschrexroth-us.com/coun...-02-cylinders/a_downloads/RA17038_2013-02.pdf

You will need to layout the linkage mechanism in say 10 degree steps from the valve end and plot the loci of the cylinder and connecting rod.

From your last pic I can see why the cylinder rod end is failing because each time it operates it see's a bending load, if you mount the cylinder pivot toward the front of the cylinder and do the loci plot, you should end up with a four bar linkage mechanism

 

[rb1957]

do you expect this mechanism to carry much load ?

the thing i wondered about was if the link, to the revolver body, was running out of clearance (with the body) and getting overloaded ??

i wonder how relevent the previous history is to this link ?

another day in paradise, or is paradise one day closer ?

 

[Dougt115]

I have been using a similar design. I needed to redesign so that I had a design margin of 1.7 before it would meet the required fatigue life of 2M cycles (this is not infinite life, 6M)for my design.

Look for any plating issues. Chrome plating contains micro cracks and will dramatically cut down the life. Other coatings can have similar effects.

You can also look at rolled threads to get a stronger part.

Make sure your part does not bottom out or reach end of travel as this will greatly increase the loads on the rod.

 

[Jboggs]

I am even more convinced that your failure is one of bending. If previous technicians described it as a buckling failure I can see how they got there. Buckling is a compression failure, but the end result looks like a bending failure. Your initial bending deflection is at the threaded joint. Then your centerline are slightly askew. Then any compressive force along the axis of the rod just amplifies the bending load. Result: fracture. Do something to either prevent that initial deflection (such as a linear slide connection) or support it like a four-bar linkage would. A trunnion mount will help some but will not eliminate the problem, and is awkward to implement.

 

[Dougt115]

Also note that 1 degree misalignment of a nut and bolt cuts the life in half, 2 degrees cuts the life by 95%.

Alignment being defined by the nut face being perpendicular to the thread axis.

So if you are loading this in compression and have any miss-alignment then the moment you are creating can be thought as a misalignment. Seeing as half the shaft will be in compression and half in tension I could easily see the part breaking.

 

[desertfox]

Jboggs

A trunnions mount combined with making the threaded joint at the cylinder rod end into a pin joint as I Suggested earlier in the threads will eliminate the problem completely because all the links will become two force members and pin joints cannot carry a bending moment.

 

[desertfox]

Hi dubc4

Regarding mode of failure, if the failure occurred over a period of time, then it is quite likely to be a bending failure in fatigue, the fatigue would occur because of the tensile stresses generated during cyclic bending, if you have chance and you have another failure of the type you describe, try and get hold of the components and as much information as you can, even better if you can get some photographs of the fractured parts and post them here.