Solved: Re: Issue when analysing assembly with contact+lar...

SylvainA. · ‎May 22, 2013

I'm starting to use CREO simulate 2.0 because of its ability to analyse assemblies with both contacts and large deformation (I'm still designing my parts and assemblies with WF5).

Unfortunately, I'm stuck in a "Decomp. Singular Stiffness Matrix. Recovering..." kind of loop (an analysis can run a couple of hours instead of a few minutes).

The assembly and analysis definitions were made in WF5 and opened/run with creo simulate 2.0.

Does anyone already have this kind of issue and know how I could get rid of it ?

Thanks !

PS: Here is an extract of my *.pas file:

Begin Time Step 1 of 4: 2.50000e-01
Wed May 22, 2013 15:01:24
Contact Analysis detected large displacements at one or more contact regions
This can be caused by a very flexible model or disconnected volumes or components
Temporarily overriding load factor with: 0.025
Iteration Residual norm contact Area
------------- ------------- ---------------
1 1.06435 35.9305 Wed May 22, 2013 15:02:26
2 0.182749 36.4971 Wed May 22, 2013 15:02:47
3 0.103263 37.8594 Wed May 22, 2013 15:03:08
4 0.0193484 34.3545 Wed May 22, 2013 15:03:24
5 0.000323151 34.4427 Wed May 22, 2013 15:03:45
6 0.000297367 34.2831 Wed May 22, 2013 15:04:06
7 0.000204668 34.1412 Wed May 22, 2013 15:04:27
8 0.000163927 34.036 Wed May 22, 2013 15:04:48
9 7.18929e-005 33.7046 Wed May 22, 2013 15:05:09
reset load factor to: 0.25
10 1 37.0183 Wed May 22, 2013 15:05:31
11 0.36423 34.3194 Wed May 22, 2013 15:05:47
*** Decomp. Singular Stiffness Matrix. Recovering...
12 0.00323441 34.0553 Wed May 22, 2013 15:06:21
*** Decomp. Singular Stiffness Matrix. Recovering...
13 0.000716325 33.9881 Wed May 22, 2013 15:06:55
*** Decomp. Singular Stiffness Matrix. Recovering...
14 0.000290376 34.0191 Wed May 22, 2013 15:07:28
*** Decomp. Singular Stiffness Matrix. Recovering...
15 0.0001727 34.0617 Wed May 22, 2013 15:08:13
*** Decomp. Singular Stiffness Matrix. Recovering...
16 0.000121401 34.0466 Wed May 22, 2013 15:08:59
*** Decomp. Singular Stiffness Matrix. Recovering...
17 9.55203e-005 34.0458 Wed May 22, 2013 15:09:44
*** Decomp. Singular Stiffness Matrix. Recovering...
18 7.8602e-005 34.0241 Wed May 22, 2013 15:10:29
*** Decomp. Singular Stiffness Matrix. Recovering...
19 6.67756e-005 34.0632 Wed May 22, 2013 15:11:15
*** Decomp. Singular Stiffness Matrix. Recovering...
20 5.74682e-005 34.0643 Wed May 22, 2013 15:12:00
*** Decomp. Singular Stiffness Matrix. Recovering...
21 5.01283e-005 34.0688 Wed May 22, 2013 15:12:46
*** Decomp. Singular Stiffness Matrix. Recovering...
22 4.41229e-005 34.0759 Wed May 22, 2013 15:13:31
Adjusted gap stiffness to prevent interpenetration
*** Decomp. Singular Stiffness Matrix. Recovering...

This thread is inactive and closed by the PTC Community Management Team. If you would like to provide a reply and re-open this thread, please notify the moderator and reference the thread. You may also use "Start a topic" button to ask a new question. Please be sure to include what version of the PTC product you are using so another community member knowledgeable about your version may be able to assist.

SylvainA. · ‎May 31, 2013

Eureka !

My issue came from a meshing of the conics surfaces which was too coarse, in addition to the rotational DOF of the screw.

I have added a displacement constraint to lock the screw rotation, an autoGEM control to refine the mesh and it worked as a charm.

.............
32 5.20418e-008 31.5623 Fri May 31, 2013 14:57:40
33 4.6305e-009 31.5626 Fri May 31, 2013 14:58:03
***Looser residual tolerance accepted because area and force
converged for all contacts
Begin Contact Pressure Calculations
............

Memory and Disk Usage:
Machine Type: Windows 7 64 Service Pack 1
RAM Allocation for Solver (megabytes): 1000.0
Total Elapsed Time (seconds): 922.35
Total CPU Time (seconds): 1134.17
Maximum Memory Usage (kilobytes): 1340524
Working Directory Disk Usage (kilobytes): 125958

Same analysis without large deformations:

...............

43 5.3210e-12 1.6448e+01 Fri May 31, 2013 15:17:44
44 4.7786e-15 1.6422e+01 Fri May 31, 2013 15:17:50
Begin Contact Postprocessing
................

Memory and Disk Usage:
Machine Type: Windows 7 64 Service Pack 1
RAM Allocation for Solver (megabytes): 1000.0
Total Elapsed Time (seconds): 350.83
Total CPU Time (seconds): 590.56
Maximum Memory Usage (kilobytes): 1200062
Working Directory Disk Usage (kilobytes): 108587

View solution in original post

SylvainA. · ‎May 27, 2013

I think that I have found the answer by myself:

There is a tapered-head screw that is in contact with a tapered hole (the screw and hole are not centered) and this screw was able to rotate. This maybe leads to an unstable assembly and singular stiffness matrix (I have found some litterature about that).

I have added a displacement constraint to remove this rotation freedom and I was able to run an analysis without any "*** Decomp. Singular Stiffness Matrix. Recovering..." message.

But it doesn't converge well (going back and forth between load factors), I stoped it after 1.5 hour of computing.

SylvainA. · ‎May 27, 2013

Unfortunately it didn't work with SPA convergence...

I have to investigate, but there is now a 0 mm² contact area

Begin Time Step 1 of 1: 1.00000e+00
Mon May 27, 2013 12:31:42
Contact Analysis detected large displacements at one or more contact regions
This can be caused by a very flexible model or disconnected volumes or components
Temporarily overriding load factor with: 0.1
Iteration Residual norm contact Area
------------- ------------- ---------------
1 3796.8 41.8902 Mon May 27, 2013 12:33:39
2 0.00208533 1.62218 Mon May 27, 2013 12:34:29
3 1 0 Mon May 27, 2013 12:34:43
*** Decomp. Singular Stiffness Matrix. Recovering...
4 0.998457 0 Mon May 27, 2013 12:35:02
*** Decomp. Singular Stiffness Matrix. Recovering...
5 0.998457 0 Mon May 27, 2013 12:35:20
*** Decomp. Singular Stiffness Matrix. Recovering...
6 0.998457 0 Mon May 27, 2013 12:35:39
*** Decomp. Singular Stiffness Matrix. Recovering...
7 0.998457 0 Mon May 27, 2013 12:35:58
*** Decomp. Singular Stiffness Matrix. Recovering...
8 0.998456 0 Mon May 27, 2013 12:36:17
*** Decomp. Singular Stiffness Matrix. Recovering...

SylvainA. · ‎May 30, 2013

So.

I have tried to replace the load by a displacement constraint (representing the screw tightening) and it worked fine...

But I need to have a load input. Can anyone help or advise me ?

SylvainA. · ‎May 31, 2013