Discussion Groups

Simulation Mechanical and Multiphysics

*Expert Elite*
Posts: 588
Registered: ‎01-19-2009

weird drop test

171 Views, 4 Replies
04-23-2012 01:16 AM

Hello,

I have a part which is falling from 0.5m to the impact plane. The first part of the impact is fine but then the part returns to the original height. Something here is not right. i am attaching the movie maybe someone can eplain why it does that.

Thanks,

Dan

Employee
Posts: 492
Registered: ‎03-25-2010

Re: weird drop test

04-23-2012 08:00 AM in reply to: Dan.M

Hi Dan,

I want to confirm what you wrote: "the part returns to the original height". Is it correct that the part returns to the original height?

In the ideal world, there is no energy loss, so the part should bounce indefinitely without losing any height. So the fact that the simulation predicts the same is a good thing!

The file you attached to the post did not include the movie. What version of the software are you using?

Sincerely,
John Holtz, P.E.
Senior User Experience Designer, Simulation
Autodesk, Inc.

Current version of Mechanical & Multiphysics: 2013 SP1 (2013.01.00.0012 28-Jun-2012)
*Expert Elite*
Posts: 588
Registered: ‎01-19-2009

Re: weird drop test

04-23-2012 12:45 PM in reply to: INACTIVE_AstroJohn

Hi,

I am using 2013 version of Multiphysics . If you would run the simulation for 5 or 6 seconds you will see the part behaivour. I agree that in  ideal world there is no energy loss and it can bounce forever. The question is how to control that? Can I use a different stifness on impact planes so they will mimic real world?

Thanks,

Dan

Employee
Posts: 398
Registered: ‎05-06-2010

Re: weird drop test

04-24-2012 07:58 AM in reply to: Dan.M

The definition for load curve 1 is only for period 0~1 second, but your simulation is up to 10 seconds. This implies that there is no gravity from 1 to 10 second, and this is why you see the part flies away.

BTW:

the load curve for gravity should be a constant, instead of ramping from 0 to 1.

The new load curve I suggest looks like this. Please have a try.

----------------------------------------------------------------------------------------------