Learning Forum

[elmalaki]

i have problem when activate arc length method with coupled boundary condition 

please reviewing my code and tell me whats wrong

[dinhan0394]

I tried your code but solution have done without error.

What is the problem that you are encountered?

 

[elmalaki]

Thank you for reply
I work in ansys 18.1 and thisis and and i have no solution whith this code.
Which ansys edition you work?

[dinhan0394]

I tried with ANSYS 17.1 and 19.2

Here is result:

[elmalaki]

thank you , i will try to solve it on ansys 19.2

[elmalaki]

i solve the code on ansys 19.2 and the same problem still existing 

the number of substeps in solution not exceed 3 substeps and the deformed shape not matching with the correct result

 

[dinhan0394]

What's error message?

(open the error file to check error message)

[jj77]

The results are unconverged and not to be taken seriously AnPham and elmalaki  - thus the solution does not complete.

If we look in the logfile then we can see that it fails:

 *** WARNING ***                         CP =      70.266   TIME= 092:50
 The auto time stepping (AUTOTS,ON command) does not support arc-length 
 method.  The auto time stepping has been turned off.                   

 *** WARNING ***                         CP =     152.281   TIME= 0934
 A reference moment value times the tolerance is used by the            
 Newton-Raphson method for checking convergence.  The calculated        
 reference MOMENT CONVERGENCE VALUE = 6.620861061E-03 is less than a    
 threshold.  This threshold defaults to 1.0-2 or is specified as MINREF 
 on the CNVTOL command.  Check results carefully.                       

 *** WARNING ***                         CP =     177.578   TIME= 093:47
 The unconverged solution (identified as time 1 substep 999999) is      
 output for analysis debug purposes.  Results should not be used for    
 any other purpose.                                                     

 

[jj77]

 

Instead of force and CP to make that edge rigid (not a good practice), use enforced displacements on that edge (makes it rigid also).

 

Thus change your code for the BC and remove force:

(The buckling load can be calculated by summing up the reactions on the edge at X=L)

 

Have in mind that you need to apply imperfections to trigger certain mode of interest as in experiments - alternatively apply some loads to make it deflect out of plane - so FEA should be similar to the tests performed - also arclen is not the only way to do nl buckling see help manual on nonlinear buckling - finally one might need to play with the arclen parameters in order to get convergence.

 

See this link for more info on NL buckling - also search the internet

 

https://www.youtube.com/watch?v=7ePBKlkGFqE

 

 

[elmalaki]

thank you for your reply

i tried to solve the problem again after changing the code BC ans make a rigid edge, the solution was completed well but the resulted deformed shape was wrong

 

[elmalaki]

this is the required deformed shape

 

[jj77]

Use the upgeom command to generate a plate with imperfections based on this buckled mode shape that you show. Then do nl buckling perhaps push it a bit more also. The imperfection will trigger that buckling mode

There is lots of info about this on the Internet, just search for upgeom and apdl