Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations waross on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

[ABAQUS] How to confirm that ABAQUS introduced geometric imperfections to the model

Status
Not open for further replies.

g.alshamsi

Civil/Environmental
Sep 29, 2020
53
Hey all,

I'm modeling an assembly of a steel shear wall using S4R elements. I'm using a elastic-perfect plastic material model and a relatively fine mesh (10 mm global seed).

I want to include geometric imperfections in my displacement controlled general static step; from my research and what I understand through reading the manual, these are the steps that I took:

1) Make a copy of the original model (called it BUCKLE)
2) Perform a linear perturbation step to find the critical eigenmodes (requested 1st 5).
3) Edited the keyword file by adding
*NODE FILE
U
4) After the analysis is complete, I went back to the original model (lets call it Pushover) and added the following keywords:
*IMPERFECTION, FILE=BUCKLE, STEP=1
1,0.4
2,0.2
3,0.08

Where 1,2,3 refer to the eigen modes and 0.4,0.2,0.08 refer to the scaling factor (percentage of the wall thickness); I then ran the analysis and plotted the force-deformation shape of the new analysis against a previous one that doesn't include imperfections.

My issue is that, the models are exactly the same; in terms of deformed shape and F-D plot. I rechecked the .DAT file of the 2nd analysis and it reports that geometric imperfections were indeed considered:

*imperfection, file=BUCKLE, step=1
OPENED RESULTS FILE BUCKLE
RESULTS FILE WRITTEN BY Abaqus RELEASE 6.19-1 27-Sep-2020 01:58:15

READING RESULTS FILE BUCKLE

***WARNING: THE PART INSTANCE FOR SOME NODES IN THE OLD MODEL COULD BE FOUND
IN THE CURRENT MODEL. THE NAMES OF REQUIRED PART INSTANCES MUST BE
CONSISTENT.

***WARNING: THE PREVIOUS ANALYSIS CONTAINED 61067 NODES, BUT ONLY 60571 NODES
WERE FOUND ON THE RESULTS FILE
IMPERFECTION DEFINED BY LINEAR MODAL SUPERPOSITION
MODE NUMBER SCALE FACTOR
1 0.400000
2 0.200000
3 8.000000E-02

Can anyone shed some light on this issue? My gut feeling is that the software is not considering imperfections in the 2nd analysis however I can't confirm that. Is there another way to confirm that ABAQUS did indeed introduce the imperfections to the geometry? I looked at the deformed shape of the assembly at t=0 in the post processing mode with a high scale factor, however there seems to be no change. I tried increasing the scaling factor of the eigenmodes as well (multiplied by by 10) and there was no difference.

Any input is greatly appreciated, thanks in advance.


G-.
 
Replies continue below

Recommended for you

Have you turned NLGEOM on ? Try with Riks step type for the second (postbuckling) analysis.
 
Yes NLGEOM is on, should have mentioned that in the original post.

I tried RIKS but I'm running into convergence issues, I'm fairly a beginner when it comes to RIKS. I don't think the analysis procedure affects whether the imperfections were included or not would they?
 
Different types of steps should give similar results but postbuckling simulations are usually performed using Riks procedure.

Various structures have different imperfection sensitivity and there are several ways of defining imperfections. I'm saying this because I recommend you perform some simple tests (for example using cylindrical shell sensitive to imperfections) to make sure that you use these procedures correctly.
 
Thanks for the input, I'm still trying several several scaling factors based on previous studies, I'll update if I make some progress.
 
Request field output COORD and check the nodal positions after the datacheck.
 
Yup, for some reason the nodal positions are the same. I'll recheck the definitions again and try it on a simpler model for the time being.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor