Modeling latch mechanism

hi, 

Please refer to the tutorial MV-2021: Simulating an Automotive Door Closure Event

You can use a template to define a multi-section simulation. 

Thank you,

I am on it!

Do you have suggestions how to avoid convergence issues when, for instance, swichting off a revolute joint and switching on a fixed joint between the two bodies?

Thank you 

hi,

the convergence issues are caused by different reason from inappropriate setup to non-physical behavior or flexible body out of linear range. Base on what I understand about your problem, I think you need to avoid the non-physical behavior and friction. switching on or off a joint actually does not affect the convergence issues so much.

Thank you Toan for your reply,

In my model I have only rigid bodies with contacts, joined with torsion springs and revolute joints. As the joint reaches an angle, say 90°, to simulate latching, the simulation is stopped with sensors, the revolute joint is deactivated and the fixed joint is activated. My idea is that the discontinuity there makes the DSTIFF not converge.

I sorted it out by using for a short lapse VSTIFF integrator, right after the joint swap, then changing back to DSTIFF.

My intention is now to see if I get stable results with different timesteps and integration tolerances.

Now I have a more pressing question now, if you know how to solve it.

When the last revolute joint the (one between GROUND BODY and the first PANEL) is deactivated, and the FIXED one is activated, the PANEL snaps back to its initial position at time 0!

How can I fix it in the current position?

Hi @silvio

Not sure what you mean back the panel snaps back to an initial position at time zero. Can you post an image?

Do you see any warnings in the log file?

Hi Prakash,

thank you for the reply.

The panel reaches 90deg position, then the fixed joint with Groundbody is supposed to activate. It does, but it is activated in the initial configuration, at 0deg. the panel is back blocked in its initial position. I wanted to fix the panel to the Groundbody in its CURRENT position.

I sorted it out right now, by using a Motion_Joint with rotational displacement equal 90deg which activates with the sensor. It makes the job of the fixed joint.

Hi,

In my opinion, You only need to activate the fixed joint. Deactivate the revolute joint seem inappropriate to me. If you worry about the over-constraint issue, try to use in-plane joint instead.

About the initial position problem,  Glad to know you fixed it. Just for your reference, If you want to restrain the displacement of joints in a certain range, try to use Bistop function in expression. 

Well, deactivating revolute joint will not harm, but the results are questionable.