Simulation af a shaft 1

[coasterchris]

I'm trying to simulate a driveshaft from a conveyer system with COSMOSMotion 2008.
The shaft is driven on one end and restrained in two self aligning bearings on both sides.

I've tried different ways of simulating this situation, but I just can't get it right.
Using the standard restraints I can’t get the shaft bending in the way this would happen in real life. (the way COSMOS 2010 has the possibility to choose from self aligning bearings as restraint)
Using “tool-components” for making the natural deformation of the shaft possible results in deformation of those parts as well. This does show believable outcome for the values of stress but does not show a coloured plot of only the axle because of immensely high stresses in the “tool-components”.

Which is the best way of restraining the shaft to get a right outcome for the stress and have a plot which shows the stress distribution in only the axle?

 

[kellnerp]

What is the problem you are trying to solve?

Can you post a screen shot of how you are trying to set this up including the restraint arrows?

TOP
CSWP, BSSE

http://www.engtran.com

http://www.niswug.org

"Node news is good news."

 

[coasterchris]

The problem is the restraining of the DOF's of the driveshaft.
I'm trying to set up a standard COSMOSWorks template document for doing studies on driveshafts.
These are used very often as the company I work for manufactures machines for the dough product processing industry which contain a lot of conveyer systems.

The goal is one document to use for different shafts.
In every situation the restraining and loading of the shaft will be in an identical way, so that only dimensions of shafts and magnitudes of the loads have to be modified.

http://files.engineering.com/getfil...4a0c-bba8-c70b28948d04&file=vervorming_as.bmp

This is the way of bending I'm trying to simulate

The shafts are fixed in two Systemplast bearings which have the possibility of self-aligning. This is where the trouble starts.
I would like to have the possibility (like COSMOSWorks 2010 has) to choose a kind of restraint which simulates the self-aligning functionality of the bearings.
I tried to simulate this using “clamp-components”, but this does not model the situation in a correct manner seeing as it induces a moment in these components.

Hopefully this clears up what I’m trying to achieve.

 

[kellnerp]

Again, what is the problem you are trying to solve with the FEA model? I am not asking what problems you are having building the model.

This model is so simple that you could probably use beam elements and get a decent solution.

TOP
CSWP, BSSE

http://www.engtran.com

http://www.niswug.org

"Node news is good news."

 

[PeteDB]

What I think you want to use is a REMOTE LOAD - DISPLACEMENT for each bearing. Create a split line on the shaft at each bearing to attach a REMOTE LOAD - DISPLACEMENT that is on the shaft axis at the bearing location.
If you have access to Solidworks customer portal there is an explanation of this restraint located in their Technical Support - Learning Resources - On Demand Videos page.

 

[coasterchris]

I interpreted “what is the problem” in the wrong way. =)
Maybe the uploaded picture showing the displacement isn’t the clearest, it was the only screenshot I had at the moment of typing my post.
What I am trying to solve is to find where the highest stress in the driveshaft will occur, to be able to make a judgement on the strength.
I don’t think a simple beam-mesh will do the trick for this.

The shaft is loaded with a double +x^2 nonuniform distributed load which simulates the pressure on the driveshaft induced by the tension of the conveyer belt.
Then there is the motor torque on one side which is induced at the start-up of the line by acceleration and the resistance to rotation because of the tension in the belt.
The thing is the bearings te be able to move otherwise the shaft does noet bend in the right way.

Hopefully this cleares stuff up a little?
Thanks for the reply’s so far, hopefully you guys could help me find an answer.

 

[kellnerp]

If you were doing this with a hand calc you would use Roark & Young to find the moments and stresses and maybe Petersen for stress concentrations.

If you were doing this with beam elements you would use Petersen to find the stress concentration where the diameters change.

Many machine design books show methods to combine torsion, bending and axial stress loads. I would be surprised if Machinery's Handbook doesn't give some guidance. Before doing FEA it is a good idea to review these manual methods so you understand what is going on in the bar. For example the bending portion of the stress is going to do a full reversal with each revolution and therefore will excite fatigue as a failure. The torsion portion of the stress is a steady load when compared to the load reversals caused by bending although a portion of the torsion will be related to the tension in the belting.

If you are doing FEA it might be a good idea to run separate load cases to obtain the stresses due to the various load components. While Simulation won't allow you to combine results, geostar, which you should have, will allow combining results of linear analyses.

The one thing you can do in your full FEA model is apply a line load on the top and bottom of the roller to apply the belt tractions. Whether you need this sort of accuracy on a conveyor analysis is another thing.



TOP
CSWP, BSSE

http://www.engtran.com

http://www.niswug.org

"Node news is good news."