CBOU for user elements

[luc]

Hi all,
I'm trying to run a simulation using PETSc and parFEAP but I have some troubles to apply the boundary conditions.
Essentially it seems to me that FEAP knows that I am using the CBOUndary command but just can't find the nodes where to apply the load.

Here how I specify the CBOUndary:

CBOUndary
  LINEar 1 0 0 0 0 0 0 0
    1,0.0,dj
    2,0.0,di

CBOUndary
  LINEar 1 0 0 0 0 0 0 0
    1,di,0.0
    2,di,dj
                              !Blank Termination Record
CSURface
  DISP,1
  LINEar
    1,0.0,dj,1
    2,0.0,di,1
                              !Blank Termination Record
CPRO
  LINEar 1 0 0 0 0 0 0 0
    1,0.0,dj
    2,0.0,di
                              !Blank Termination Record

and here is what I get back in the O file:

 FEAP * * plate_imperf                                                         

   C o o r d i n a t e    S u r f a c e   C o n d i t i o n
   x_1 coord.   y_1 coord.   x_2 coord.   y_2 coord.   x_3 coord.   y_3 coord.
   0.0000E+00   1.0000E-05   0.0000E+00   2.0000E-05   0.0000E+00   1.5000E-05
  1 B.C.  2 B.C.  3 B.C.  4 B.C.  5 B.C.  6 B.C.  7 B.C.  8 B.C.
       1       0       0       0       0       0       0       0

       N o d a l    B o u n d a r y    C o d e s

      Node    1-B.C.    2-B.C.    3-B.C.    4-B.C.    5-B.C.    6-B.C.
              7-B.C.    8-B.C.

 FEAP * * plate_imperf                                                         

   C o o r d i n a t e    S u r f a c e   C o n d i t i o n
   x_1 coord.   y_1 coord.   x_2 coord.   y_2 coord.   x_3 coord.   y_3 coord.
   2.0000E-05   0.0000E+00   2.0000E-05   1.0000E-05   2.0000E-05   5.0000E-06
  1 B.C.  2 B.C.  3 B.C.  4 B.C.  5 B.C.  6 B.C.  7 B.C.  8 B.C.
       1       0       0       0       0       0       0       0

       N o d a l    B o u n d a r y    C o d e s

      Node    1-B.C.    2-B.C.    3-B.C.    4-B.C.    5-B.C.    6-B.C.
              7-B.C.    8-B.C.

 FEAP * * plate_imperf                                                         

   C o o r d i n a t e    S u r f a c e   C o n d i t i o n
   x_1 coord.   y_1 coord.   x_2 coord.   y_2 coord.   x_3 coord.   y_3 coord.
   0.0000E+00   1.0000E-05   0.0000E+00   2.0000E-05   0.0000E+00   1.5000E-05
  1 Prop  2 Prop  3 Prop  4 Prop  5 Prop  6 Prop  7 Prop  8 Prop
       1       0       0       0       0       0       0       0

       N o d a l    P r o p o r t i o n a l    L o a d s

      Node

 FEAP * * plate_imperf                                                         

   C o o r d i n a t e    S u r f a c e   D i s p l.
   x_1 coord.   y_1 coord.   x_2 coord.   y_2 coord.   x_3 coord.   y_3 coord.
   0.0000E+00   1.0000E-05   0.0000E+00   2.0000E-05   0.0000E+00   1.5000E-05
   d_1 displ                 d_2 displ                 d_3 displ
   1.0000E+00                1.0000E+00                1.0000E+00

       N o d a l    D i s p l a c e m e n t s

       a_node    a_dof   d_a displ

          Material    Element Tag   Element Type  History Terms  Element Terms
              1              1             12              0              1

If you have any idea I'm open to suggestions.

Thanks!

[FEAP_Admin]

It has not found any nodes which means that there are no nodes that satisfy the specified search.  (your usage looks fine)

Are you sure that there are nodes at these values? 

If this is in a partitioned file, then it is easy to
imagine that sometimes the search will come up empty but that would not be an error. 

If this is in the original file then you need to make sure you have nodes at these locations.

[luc]

I think that there are nodes to be found in the segments described in the CBOU commands. So for info here are the values of my parameters:

  di=2.0e-5
  dj=1.0e-5

and here are the coordinates of half the nodes on the mesh:

(actually I removed the list of nodes, because I exceeded the 20000 characters limit, the nodes coordinates are in a file attached to this message).

these coordinates appear in the parallel input file generated by parfeap, which I then use to run the simulation in parallel.
The above list contains only half the nodes in the mesh, the other half of the list contains exactly the same nodes. This is due to the fact that I am doing a mixed formulation and I need to create elements for each fields.
I am worried that feap is filtering out the nodes that are redundant (which is all the nodes in my simulation!).

Let me know what you think about this. I can send you the input files and O file by mail if you want to get a closer look at them (they are to big to be attached here).

Finally, I also provided a plot of the nodes on my mesh (in blue) and of the boundary conditions as specified in CBOU (in red).

Thank for your help,
Luc

[FEAP_Admin]

I tried to follow what you are trying to do but it is too hard to figure out.
Please make a small version version of your problem and post the initial Ifile, as well as the partitioned Ifile_xxxxs.

To debug problems, it is useless to work with large problems.  Always reduce your problem to the smallest possible
case that shows the pathology.

[luc]

Sorry for the long delay between posts.
Since I needed a quick turn around to continue to make progress in my work I just programmed a little script outside FEAP which finds the nodes on the boundary, list these nodes in the input file and use CBOU with the NODE argument.
This means that I do not fix the problem but just work around it... yet it was fast to implement and works perfectly.
If times allow I'll look back at this problem and try to figure out what it is that FEAP does differently when I use CBOU on an edged segment!

[FEAP_Admin]

fyi, you can also write a user mesh command (umeshXX.f) to implement short fixes like this.

[luc]

Yes indeed that would be the cleanest way of working but I already have a NURBS mesh generator in MATLAB so I would need to transfer it to FEAP plus I would need to run feap twice, one to create the INPUT file and a second time to read it for actual simulation.

Also we are running low on unused umesh and umacr so I'm trying to externalize things if possible or to branch the said umesh with if statement in order to perform more operations with one file...

I'm hopeful to get more time in the next 3months to rewrite a little my functions and subroutine and package it more nicely. If so I'll try your approach and post about it.

Best,
Luc