Different Postview output for 1.4.2 vs 1.4.4 with same .xplt file
issueid=633 08-16-2013 08:34 AM
Junior Member
Different Postview output for 1.4.2 vs 1.4.4 with same .xplt file
Getting different values for 3rd principal strain from Postview 1.4.2 vs 1.4.4 with same .xplt file.

I seem to be getting different values for 3rd principal strain depending on which version of Postview I'm using, e.g. version 1.4.4 vs. 1.4.2. To illustrate I can send screenshots from two different .xplt files, or send the .xplt files themselves. For example, with "model B", the minimum 3rd principal strain is 14.7% (1.4.4) vs. 13.5% (1.4.2). For "model A", it is 14.4% (1.4.4) vs. 9.5% (1.4.2).

Brian
Issue Details
Issue Number 633
Project PostView
Category Unknown
Status Not a Bug
Priority Unknown
Affected Version Unknown
Fixed Version (none)
Users able to reproduce bug 0
Users unable to reproduce bug 0
Assigned Users (none)
Tags (none)




09-28-2013 12:56 AM
Junior Member
I am using 1.44 while i have some kind of similar problem. It seems the result shown by the legend is not accurate and the error could be very large. in my simulation, the maximum strain shown by the legend is 0.108 while the summary function gives 0.14. can you check if the summary function from "Post" Tab gives you the same result in both versions?

regards
Liang

10-02-2013 08:48 AM
Lead Code Developer
Hi Brian and Liang,

Just wanted to let you know we are looking into this and will report back when we found the source of the discrepancy.

Cheers,

Steve

10-14-2013 09:31 PM
Lead Code Developer
Hi Brian,

I finally figured out the source of the discrepancy. At some point between 1.4.2 and 1.4.4 I modified the algorithm to evaluate the Lagrange strains. In 1.4.2 it was evaluated at the center of the element and then for each node, the nodal value was calculated by taking the average of the surrounding elements. In 1.4.4 the strains are evaluated at the nodes for each element and then for each node, the average is taken from the corresponding element values. I hope this makes sense. In areas where the strain is uniform the difference should be negligible. However, in areas where the strain changes rapidly, the 1.4.4. values will generally be higher compared to 1.4.2. This is indeed what you are seeing. So that explains the differences, but the question now is, which one is more accurate. Well, probably neither to be honest. If your strain changes rapidly over a distance of a few elements, you will not get an accurate prediction of strain, regardless of what strain recovery method you use. You may need to refine your mesh to get a more accurate prediction of strain (and consequently stress). So, bottom line, I don't think there is a bug in PostView. The observed difference is simply an effect of a slightly different strain-recovery method that was implemented in 1.4.4. If for some reason you need to able to reproduce the results from 1.4.2. what I can do is built in an option so that you can choose which strain recovery method you want to use. Just let me know.

Cheers,

Steve

+ Reply