Contact problem

[Xuannamdo-danang]

Dear Professor Taylor,
Dear all,

Currently, I would like to create a mesh like in attached file. I read the Manual user and Manual example for contact problem but I do not understand how to declare contacts for this case. Could you please tell me how to do?

Thank you so much

Yours sincerely,
DO XUAN NAM

[Prof. R.L. Taylor]

Here is a model  problem.  This shows you how to set up a problem

[Xuannamdo-danang]

Thank you so much for your reply, Professor  R. L. Taylor

[Xuannamdo-danang]

Dear Professor Taylor,

Based on your input file, I did another example (attached file) to comprehend contact prolem. However, result is not like expected

Please tell me where errors are

Thank you so much

Yours sincerely,
DO XUAN NAM

[Prof. R.L. Taylor]

Before solving a problem you should check the mesh that the program processed all the data correctly.  Often looking at the log file helps.  In your case the program did not find one of the surfaces because you had an "a"  instead of an "h" for the location.  Then one of the surfaces is going the wrong way so you need to reverse the pair to get the correct solution.  You almost had things correct which is good.

Contact surfaces should traverse boundaries so that the computed normal by "right-hand-rule" points outward, that is toward the other surface.  In 2-d this is required for the master surface always, but the slave surface need not since only nodes are used.

Node you could have noted both surfaces were not found by using the PLOT PAIR command

[Xuannamdo-danang]

Thank you so much Professor Taylor. My program has just run properly.

[Xuannamdo-danang]

Dear Professor Taylor,

After the first example to understand contact problem, I come back my main problem in which I  need to create a mesh like in attached file. However, result obtained is still not like expected.

Could you please tell me how to solve this problem

Thank you so much

Yours sincerely,
DO XUAN NAM

[Prof. R.L. Taylor]

You need to understand how to separate the mesh input from the contact input.  There is no END for the mesh before the contact data. 

You need to use graphics to see what you successfully input

For example use PLOT BOUN to see the restrained dof's;

Use PLOT PAIR n to see the contact pair n

Use CHECk to make sure the mesh is correct, it also tells you how many dof's are fixed in each direction to ensure you will not have rigid body modes in the solution -- rigid body modes in a static analysis  will cause the solver to produce erroneous solutions.

[Xuannamdo-danang]

Dear Professor Taylor,

Firstly, I would like to thank for your quick reply.
Professor, I did this example once again but results obtained are still not like expectecd.
Could you please tell me where errors are and why?

Thank you so much

Yours sincerely,
DO XUAN NAM

[Prof. R.L. Taylor]

Look at the attached file with a PLOT PAIR command.

Also you have no boundary conditions for the displacements to act on.

You need to use the PLOT commands to look at your mesh before you try to solve the problem.  Make sure you have all set as you want and then try to solve.  Remember, you cannot have rigid body modes anywhere in the problem unless you do a transient solution (and contact is not fully coded for transient solutions).

[Xuannamdo-danang]

Dear Professor Taylor,

I think that "contact is also called changing boundary condition". It is the reason why I did not put the boundary condition in my input file. Now, I try to put the boundary condition, however, results do not change, i.e, they are not like expected.

I do not why and how to find a suitable solution

Please tell me

Thank you so much

Yours sincerely,
DO XUAN NAM

[Prof. R.L. Taylor]

Do not try to do the solution until you plot everything while in INTEractive mode of solution.

1. Plot the mesh  PLOT MESH

2. PLOT BOUN to see where the restrained boundary conditions are located.

3. Use PLOT WIPE to clear the screen and plot the mesh again

4. PLOT  PAIR to see where you contact surfaces are located.

Does everything LOOK LIKE YOU WANT IT?  From what I saw before it DOES NOT

[Xuannamdo-danang]

Dear Professor Taylor,

Thank you so much for your quick reply.

According to your guide, I did this example once again but with a coarse mesh. The results obtained shows boundary conditions and pair (slave and master surfaces). However, during process of running  there is no contact between these surfaces.

Please tell me why and how to solve this problem

Thank you so much

Yours sincerely,
DO XUAN NAM

[Prof. R.L. Taylor]

If you increase the normal penalty value by about 4 orders of magnitude it seems to work quite well (1.d8 is a good thing to try)

[Xuannamdo-danang]

Thank you so so so much Professor Taylor, it ran properly

Yours sincerely,
DO XUAN NAM