How to model a free multibody in simmechanics?

Question

David St-Onge 2015-3-11

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/182686-how-to-model-a-free-multibody-in-simmechanics

评论： David St-Onge 2015-3-12

Assemblage1.mdl

Hi!

We got a trial version of simMechanics at our departement to compare it to Adams and other simulation software. So no real experience with the software here...

I modeled a simple 2-body 1 revolute joint mechanism which is free to move in the plane (no gravity or gravity perpendicular to the plane). I think the model created by the SolidWork plugin is right, as far as I understand simMechanics (see mdl attached), but the simulations results aren't good. Since both links are identical it should 'flap' the two of them around the joint without moving the mass center. In this case the mechanism starts translating in 'y' quite a lot.

thanks you.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

David St-Onge 2015-3-12

I think I found the solution. It seems my motion input weren't good for simulation. using transfer function to derive the position as in: http://blogs.mathworks.com/seth/2012/02/01/applying-motion-to-simmechanics-models/ and it worked.

Thank you.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Sebastian Castro 2015-3-11

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/182686-how-to-model-a-free-multibody-in-simmechanics#answer_171010

Did you modify the Mechanism Configuration block included?

When you import from CAD, the block is automatically set to have a gravitational acceleration of -9.81 m/s^2 in the y (or z) direction. If you haven't already, you should try changing that vector to [0 0 0] (so, no gravity).

- Sebastian

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

Answer 2

David St-Onge 2015-3-11

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/182686-how-to-model-a-free-multibody-in-simmechanics#answer_171015

Thank you for your answer. I did change the acceleretion to [0,0,0]. Same results.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Sebastian Castro 2015-3-11

It's likely numerical integration errors building up, then.

If you go to Simulation > Model Configuration Parameters and go to the Solver pane , try reducing the Max step size, Relative Tolerance, and Absolute Tolerance parameters.

This will likely help reduce the drift, but may not give perfect results since it's all numerical integration and computers have their limits too. Try it out, though!

- Sebastian

请先登录，再进行评论。