Having problems with getting a job to run; axle with over closure

[nolanb13]

In my job there is an axel which is forced through a bore. The bore has several `noses` or `ribs` which are used to hold the axel in place while reducing friction during assembly. There is an over-closure between the noses and the axel in the model, which creates a large stress in this location. We are attempting to find the force required to push/pull the axel back out of the bore, however each attempted job fails, with less than 10% of the load having been calculated. (It reached 10% on only one of the iterations.)

If you need any input or output files I can provide them, but any help you can provide would be greatly appreciated. Many thanks

 

[Mustaine3]

Have you used the Interference Fit option in the contact pairs to resolve the overclosure?

 

[nolanb13]

Sorry, I guess I wasn't clear. The overclosure is included on purpose. It represents the real parts in the assembly, where the axel is slightly larger than the gap it is pushed into, in order to hold it in place

 

[Mustaine3]

I got that. When these overclosed regions are used in contact, then will the solver try to push them out of each other. And the best method to get this done is with the Interference Fit option. Otherwise the solver will do it in the first increment, which is more difficult.

 

[nolanb13]

Oh right, sorry I didn't know what Interference Fit does exactly. I will give it a try. Thanks

 

[nolanb13]

I cannot find the Interference Fit option to enable. I am using the Abaqus deck within the ANSA software. Do you know what I can do?

 

[Mustaine3]

I'm not familiar with ANSA, so I don't know if it can do it.

If it not supported in the GUI, then you have to add into the input file. Create the normal contact pair and in the step add something like this:
*Contact Interference, shrink
surf-A, surf-B

The Abaqus manuals explain the details.

 

[nolanb13]

Okay, I have set that up but I am getting a warning in the message file:

***WARNING: SOLVER PROBLEM. ZERO PIVOT WHEN PROCESSING A (TIED) CONTACT
CONSTRAINT D.O.F. 1 FOR SLAVE NODE 3064. THIS MAY BE DUE TO THE
UNIT USED FOR THE DISPLACEMENT DIFFERS BY MORE THAN 10 ORDER OF
MAGNITUDE THAN THE UNIT USED FOR THE MATERIAL STIFFNESS. IF THIS
IS THE CASE, PLEASE USE A DIFFERENT UNIT SYSTEM.

OVERCONSTRAINT CHECKS:
An overconstraint was detected at one of the Lagrange multipliers associated
with contact slave node 3064. There are multiple constraints applied directly
or chained constraints that are applied indirectly to this node.
A list of nodes and chained constraints between these nodes that most likely lead
to the detected overconstraint can be printed out to the message file by including
*constraint controls, print=yes in the input file.

This warning repeats for a large amount of nodes, which all seem to lie on the surface of the axle, which is overclosed into the support.

 

[Mustaine3]

Take away all other definitions from the overclosed regions and use a contact property without friction. The friction can be added in the next step, when the overclosure is solved.

 

[nolanb13]

There are no other defitions in the overclosed regions, just the contact definition. I have removed the friction (which did seem to work) and am now running the analysis as Dynamic, Quasi-Static per the advice of a colleague. Im not sure if this was what has caused my new issue, but there is no forced/stress calculated in the simulation. The overclosure should be imposing some stress on the model right?

Edit: scratch that, I am getting zero force everywhere even when I use Static analysis type. Why is there no force being exerted

 

[Mustaine3]

I don't know. You should know your model and check what's going on.

 

[Erik Panos Kostson]

I think that you should get a stress with it. At least this what I see from this video that might be of help: Link