Ansys Rigid Body Motion - Static Structural

[jmart157]

I am and have been trying to do a specific simulation for the last several weeks. It seems like I take one step forward two steps back.

See below:

See the attached files. I'm on Ansys Workbench 13.

www.plastics-consulting.com/rod.stp[/URL]
[PLAIN]www.plastics-consulting.com/tube.stp[/URL]

All I want to do is treat the tube as completely rigid, inside and out. I force the rod down through the tube and see where it comes in contact with the tube ID. You can flush out the top of the rod and the top of the tube section. Essentially the rod should be constrained to the ID of the tube section. That’s it.

The tube should stay rigid, the rod will deflect in order to fit down the curved tube profile. The rod cannot/will not penetrate the tube ID (frictionless). I want to see contact points and stresses.

Looks like explicit dynamics is an option, or static structural.

Both can be treated as steel. I’ve been jumping into Ansys left and right and been doing some very minor valve simulations. This simple concept for this project came up awhile ago. I’ve asked many different people – coworkers, college professors, etc, and no one has been able to point me in the right direction.

Is this over the top difficult?

 

[Mech_Engineer]

Have you implemented some of the suggested settings I mention here: https://www.physicsforums.com/showthread.php?t=433240

Those are pretty good settings for sliding contact conditions; the most important are mesh density, and splitting the problem into a large number of substeps.

What you're looking to do isn't impossible, but it does take someone who is a competent simulation engineer who is familiar with the software being applied.

 

[jmart157]

Yes I have, you just posted in the thread that I revived. It is the exact one I was looking at as I feel this is a similar situation.

I have had no success. The convergence line and criterion never meet. I have revised and tried probably 10 times with changing one thing at a time.

 

[Mech_Engineer]

Yes I have, you just posted in the thread that I revived. It is the exact one I was looking at as I feel this is a similar situation.

I have had no success. The convergence line and criterion never meet. I have revised and tried probably 10 times with changing one thing at a time.

How far is your part moving, and how many substeps are you splitting the solution into?

 

[jmart157]

I'm trying to flush out the top of the rod with the top of the tubing (faces are on the same plane).

Would like to do a substep every 2". If the rod moves 12", then 6 substeps.

JRM

 

[Mech_Engineer]

Would like to do a substep every 2". If the rod moves 12", then 6 substeps.

No way, you'll need literally hundreds of substeps. I would step the rod every .010" or less, with no more surface penetration than .005-.010". Moving it too far means that surafces are penetrating/interfering with each other, and the nonlinear solver can't find a solution.

Hope you don't mind you computer running ANSYS for a couple of weeks ;)

 

[jmart157]

I'm looking at 8600 sub steps then.

I am aware it will be running for awhile - x6 1100t, fx4800, raid 0 across 6 hard drives, 16gb ddr3. Hopefully that will take this from a couple to one.

JRM

 

[jmart157]

Are you saying I have to do substeps with the rod from outside of the tube and push it throughout the entire tube?

 

[Mech_Engineer]

You'll need to start substepping as soon as the rod comes into contact with the tube. No point in doing it before though.