Dear All

If you first opens attached file, then you can understand my question:

http://www.4shared.com/file/8VArlSA_ba/Images.html

This is a double mounting ear (or clevis) that is welded to steel plate. Applied force from a hydraulic cylinder on the ear is about 27 tons. As you can see in the images, steel plate nearly behind the ear has been torn off. I actually what is the technical name for this type of failure, maybe it is converse PUNCHING failure.

So I as mechanical engineer tend to redesign this structure. To do that I need to know why this has failed and hence I have to calculate stresses in important points. As I prefer not to use FEM, I wonder if anyone can help me in calculating by hand, even a rough calculation is good.

Thanks.
regards.
Hadi

 

Hadi

I cannot open the images, can you insert the images in a post?

 

Peter

Sure; You can see them in following:

 

Hadi

You have not indicated the thickness of the mounting plate, assuming it's to scale, it appears too thin.

Looking at the amount of weld, it has been repaired many times before and therefore was not a sudden failure.

 

Peter;
Thanks, you are right, it has been welded one time before (without considering pre heating and etc).the reason for weldment was they might have needed to stiffen this! So one reason might be that mounting plate (thickness about 45 mm) got brittle. Although it was sudden.

Any way I need to know what is the basic step to design such structure. For example how do you say it is thin? What if I say load is 2 tons? I need analytical calculation to justify for supervisor.

Thanks again for fast reply.

 

Hadi

If the 'ears' are 45 mm thick then the mounting plate does not look that thick, possibly only 25 - 30 mm.

Look at this data sheet http://www.atos.com/english/technical_tables/english/B500.pdf C134 for standard sizes

 

Peter;

Thanks, it is a good source. I thought that I am a good searcher in the web, but I found out you are better than me!!

I wonder if you know the type of this failure, a technical term that I can use for googling.

Thanks.

 

Hadi

I would guess, because the mounting plate is too thin, it has been flexing causing a fatigue failure.

 

because of the length of the lugs they put a moment into the plate. you can estimate the stress that this moment causes by assuming that it is reacted by the material in the square that just contains the lugs. In other words, if you cut out a square just big enough to hold the lugs, and wanted to glue it back in place, how strong would the glue need to be in shear. (of course we are not talking about actual glue - just an analytical approach) if you cut almost all the way around the square so it looked like this [], with the load going left/right, then the piece would obviously tear out. so the resistance is in proportion to distance. Set it up like the bending stiffness derivation in your text and you can get an equation for the stress as a function of load, lug size, and plate thickness. That will give you the shear stress and then us that in combination with some conservative estimate of tension/ compression to get a good guess at the true state of stress. (use Mohr's circle)

I think the design fix is to put some section (I or C channel) under the lug, running parallel to the cylinder and run that until it can be tied into some major framework. Depth (the I in mc/I) is your friend.

 

Hello!. As I see it, it's the amount of weld that cause the failure. The base plate has metallurgical structure damaged by the excess heat put in. The crack occurred in the so called "heat affected area" of the weld, a transitional area of a few millimeters (more or less, depending of the welding procedure) between the weld and the unaffected base plate. You may try to reinforce that area, lower the distance between pin axis and the plate (if that's possible) in order to minimize the momentum given by the force.