Learning Forum

[marrom15]

Hi,

i am trying to design a torsion testing machine modeled in Static Structural that looks as follows

The green brackets are bonded to the blue cylinder which rotates around the red base. Remote force in Y direction is applied to the smaller cylinder bonded to the brackets.

At first i used body to body revolute joint between the base and the blue cylinder, but the solving time was over 30 min. Equivalent stress was also high on the brackets (around 110 MPa) but i didn t know what value to expect.

Then i copied the system and tried frictionless contact between the 2 bodies and added an 0 mm displacement on the blue cylinder in Z direction just to be safe. Equivalent stress went down to around 60 MPa to the brackets along with cutting solving time to half.

All bodies except the little cylinder are flexible and large deflection in on (non-linear problem).

I thought that the 2 models were equal, and expected similar stress. Does anyone know the reason this happened? Which is the right way to model this problem?

Thanks in advance.

[peteroznewman]

A revolute joint consists of a spider of elements from one face, to a central coordinate system, and another spider of elements from that coordinate system out to another face.

A contact consists of nodes checking for penetration of element faces.

They are going to give very different stress results because they are very different connections.

[marrom15]

Thanks for your reply Peter,

may i ask you which connection in your opinion describes the model more realistically?

I think i understand the difference but does that justify double the difference in stress? 

Although both connections do the job of revolution, from the images and your explanation above it seems that the contact is closer to reality.

Do you agree?

 

[peteroznewman]

The joint can pull on the whole circle in a way that can increase stress. The frictionless contact can only push on one side and is more realistic.

[marrom15]

Thank you for your help