ABAQUS contact help- pin inside hole of different radius

[bjoman]

I'm having some trouble modeling a contact situation using ABAQUS 6.3.5/CAE. The problem is (3) wood 2x4s spliced together by 2 pins..as shown (poorly) below..
_
_| |_
| | | |
| | | |
| |_| |
|_| |_|

(3) wood 2x4s
(pic shows up even worse when posting than when sketching it here, but I think you can get the gist of how the 2x4s are arrranged with 2 pins going through all three blocks)

the diameter of the pins is 1/4" and the the holes for the pins have a diameter 1/16" bigger than the diameter of the pins. I originally created the model with the pins centered in the holes, then I imposed a downward displacment on the center wood piece to make contact with the pins, then imposed downward displacement on the two pins and the center piece to make contact with the pin holes in the two outter wood pieces. Then I load the model by placing a pressure on the top of the center piece The problems I am encountering are incorrect normals on the pins and contact surfaces I've checked the viewer and they appear to be facing the correct direction)and D.O.F./solver problems on the pins where the two faces of the blocks meet.

So my questions are:
Is there a better way this should be modeled?
Is there anyway to make sure the normals match up when a larger radius cylinder (the pin hole) is contacting a smaller radius body (the pin) which is inside of itself?

Any other suggestions or thoughts would also gladly be appreciated.

 

[corus]

It seems ok what you have done but with the pin being smaller than the hole then you'll probably only have point contact at the pin/joint interface. I presume that you also have contact surfaces between the flat surfaces of the splice so perhaps the problem arises where your contact is either between the hole edge or the flat surface and hence the normals don't match. If there is no easy answer then perhaps you could replace the point contact you have with the pin by a tying the nodes together and just leaving the contact between the flat surfaces to remain.

corus

 

[bjoman]

Right now I have the contact surfaces set up so that the whole pin is a contact surface for each of the 3 pin-hole sections. Would it be better to partition the pin at each splice and then just designate that contact surface as the master/slave surface?

Also, how I've set up the master/slave surfaces is: for the outer splices the pin is the master and the pin holes are the slave, while for the middle splice the pin hole is the master and the pin is the slave. I did this because it seemed logical based on how the the model is loaded. Is this the right way of going about this?

 

[corus]

In terms of run-time I'd only make that part of the pin surface likely to be in contact as the contact surface. It'd be inefficient to make the whole of the pin surface a contact surface.

The choice of master/slave surface is normally made on mesh density and stiffness rather than loading. I think you can actually define the contact twice so that a surface is both master and slave if you're concerned about nodes passing through the other surface. Generally though if you match the two meshes up to have the same mesh density then you'd get better results with little effect of nodes passign through surfaces, rather than having one as a coarse mesh and the other a fine mesh.



corus