thread825-381363
I'm trying to model a bicycle wheel with the spokes modelled as pre-tensioned cables. I'm using NX Nastran and I'm following a compression-only behaviour approach. I'm using CROD elements with a non-linear elastic stress-strain curve that doesn't (almost) take compression.
I've seen the post above and it looks like someone made it, but I can't manage to get SOL106 to converge. I guess that when the cables (spokes) lose the pretension the solver detects some kind of instability and keeps bisecting my load without converging.
Here is a snippet from the f06 file:
*** USER INFORMATION MESSAGE 6137 (DFMN)
INPUT MATRIX TO DECOMPOSITION HAS 6 SINGULARITIES.
SEE FOLLOWING MESSAGES FOR DETAILS.
*** USER INFORMATION MESSAGE 4158 (DFMSA)
---- STATISTICS FOR SPARSE DECOMPOSITION OF DATA BLOCK KLLRH FOLLOW
NUMBER OF NEGATIVE TERMS ON FACTOR DIAGONAL = 5527
MAXIMUM RATIO OF MATRIX DIAGONAL TO FACTOR DIAGONAL = 5.6E+19 AT ROW NUMBER 667817
*** USER WARNING MESSAGE 4698 (DCMPD)
STATISTICS FOR DECOMPOSITION OF MATRIX KLLRH .
THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER THAN
1.00000E+07 OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL.
USER INFORMATION:
THIS MESSAGE MAY BE IGNORED IF NO GRID POINT IDS OR HIGH RATIO MESSAGES APPEAR IN THE TABLE ON THE NEXT PAGE.
1 SEPTEMBER 28, 2021 SIMCENTER NASTRAN 5/ 9/19 PAGE 12
All the nodes which are listed after this message are laying on the spokes and the solver bisect and fails to converge.
This shouldn't happen because even if a couple of spokes lose their pretension the whole structure is kept in place by the other spokes.
Some tips on how I can get around this?
PS. I have LGDISP turned on (doesn't work also if I turn it off), the pretension is applied with a thermal load, the wheel is loaded radially.
Best,
Massimo
I'm trying to model a bicycle wheel with the spokes modelled as pre-tensioned cables. I'm using NX Nastran and I'm following a compression-only behaviour approach. I'm using CROD elements with a non-linear elastic stress-strain curve that doesn't (almost) take compression.
I've seen the post above and it looks like someone made it, but I can't manage to get SOL106 to converge. I guess that when the cables (spokes) lose the pretension the solver detects some kind of instability and keeps bisecting my load without converging.
Here is a snippet from the f06 file:
*** USER INFORMATION MESSAGE 6137 (DFMN)
INPUT MATRIX TO DECOMPOSITION HAS 6 SINGULARITIES.
SEE FOLLOWING MESSAGES FOR DETAILS.
*** USER INFORMATION MESSAGE 4158 (DFMSA)
---- STATISTICS FOR SPARSE DECOMPOSITION OF DATA BLOCK KLLRH FOLLOW
NUMBER OF NEGATIVE TERMS ON FACTOR DIAGONAL = 5527
MAXIMUM RATIO OF MATRIX DIAGONAL TO FACTOR DIAGONAL = 5.6E+19 AT ROW NUMBER 667817
*** USER WARNING MESSAGE 4698 (DCMPD)
STATISTICS FOR DECOMPOSITION OF MATRIX KLLRH .
THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER THAN
1.00000E+07 OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL.
USER INFORMATION:
THIS MESSAGE MAY BE IGNORED IF NO GRID POINT IDS OR HIGH RATIO MESSAGES APPEAR IN THE TABLE ON THE NEXT PAGE.
1 SEPTEMBER 28, 2021 SIMCENTER NASTRAN 5/ 9/19 PAGE 12
All the nodes which are listed after this message are laying on the spokes and the solver bisect and fails to converge.
This shouldn't happen because even if a couple of spokes lose their pretension the whole structure is kept in place by the other spokes.
Some tips on how I can get around this?
PS. I have LGDISP turned on (doesn't work also if I turn it off), the pretension is applied with a thermal load, the wheel is loaded radially.
Best,
Massimo
