Stress intensity factor convergence problem

[plus321]

Hi,
Modelling a crack in a plate in ABAQUS.
Using sweep meshing technique to model the elements around the crack tip (collapsed elements). Using the contour integral method to model the crack. I need SIF output for comparison with theoretical calculation by hand.
The crack length is 0.2mm and the distant force (pressure) condition is 20Mpa on each side of plate.
There are problems with the stress intensity factors I am getting for each contour:

K1 K2
contour 1 -1.7444 0.000253368
contour 2 -1.8815 0.0011523
contour 3 4.94398 0.000474001

this is with the most refined mesh.
The K2 values are converging nicely but are those K1 values appropriate?

My question is what should I be expecting - all K1 values to be near each other for each contour?

Using a coarser mesh I got the following values as well:
K1 K2
contour 1 9.22192 -9.16E-05
contour 2 4.08834 -4.80E-05
contour 3 6.65992 -5.51E-05

Again these K1 values aren't the same.

any help would be massively appreciated

 

[ABQDwarf]

Hi plus321,

the solution on the first contours are affected by the stress field ahead of the crack tip. To get the right one you must consider the SIF exclude the first contours.
Usually I consider the SIF's values from the fifth contour, where I get usually an error of about 1% or less compared to analytical solutions.

What kind of model you're consider?

regards
ABQDwarf