Hey all,
I have issue extracting results from the output Plaxis 2D:
trying to extract vertical stresses under a foundation, I have chosen stress points very close to the base of the foundation.
using the command:
g_o.getcurveresults(g_o.Stresspoints[0], step, g_o.ResultTypes.Soil.SigyyE)
The problem is that the value extracted isn't the one I can see in the viewer, while inspecting the overall stress filed with the hint box. phase / step number / coordinates are correct!
The difference is quite significant and cannot result from interpolation / intergration between the points. The result is also quite off from what is expected, so I would assume the stress extracted from the stresspoint is wrong.
I have attached a pdf with screenshots and would appreciate a lot if you could tell why I am getting wrong results.
Cheers,
Nataly
PDF
Dear Nataly,
It is possible that Result smoothing is not considered in the Output results.
Both the plot and the command support result smoothing which can make the results differ.
Note that for the command, an extra parameter to consider result smoothing is a boolean, so True or False, and the default value is True.
Dear Stefanos,
thanks for your prompt response.
I checked the output viewer, and the result smoothing is/ was on.
So, if the default value of the command is True, then the setting is the same and the diffrerence in shown an dextracted value should not come from the smoothing.
I tried to include True/ False into the command but it seems not to have this option. Checked the command manual - nothing about result smoothening, boolean true or false.
See files attached.
Hello again,
Apologies, I missed which command you were using. Indeed, the getcurveresults does not have smoothing as it retrieves the results straight from the data from the calculation kernel.
Seeing the Hint box, however, I can see that you have not placed your mouse exactly on the stress point, from which the stress data comes from.The hint box does some processing to be able to show the results on any location that highly relies on the data around that location.
To the result processing: a stress result is computed on stress points. A selected curve point for stresses shows the results from a stress point.
The current plot you are looking at is not based on the stress points but based on the nodes, which is why you see nice colour shadings. That means that the results from the stress points are extrapolated first to the nodes and then the hint box processes this data to give you a value where your mouse is.
Try hovering with your mouse at the exact location of the selected stress point and see the hint box result.To show the stress points you can select to view the eye icon on the explorer (left side).
I am not sure why you thought my mouse is not at the exact point. In my first pdf attached, I highlighted the coordinates of the stresspoint and the coordinates within the hint box. they are exactly the same. meaning the hint box is placed at the location of that stresspoint.
I would not be surpriced seeing a slight difference between the values (due to smoothing, interpolation or anything else)
But here the diffrence is HUGE! -17.6 kPa vs. -60.8 kPa.
Moving around the stresspoint with the hint box, I don't see any values that are even nearly close to the result of the stresspoint.
Maybe worth mentioning, I am running the same type of analysis with different soil profiles. The cooredinate of the stresspoints and the commandos used to extract the results, are exactly the same. from about 15 models, 2-3 had this issue.
I am then wondering what is a better (more reliable and stable) way to extract the stresses from the soil elements? Previously I saw a recommendation to use stresspoints for such data. Now it seems like it is not giving the results we need.
To clear things out, first of all I know that you were not exactly on the stress point as the Hint box looks like this when you are exactly on a stress point (more information is given):
And not like this when you are in between:
I understand that the difference is large but this highly depends on the neighbouring elements (mesh density) and the values. From the screenshot, I can see that between two neighbouring stress points the stress results are actually quite different:
Again using this plot with the Hint box will not give you the exact results but the interpolated (backend shape functions running) value based on what you have nearby. This highly depends on the mesh density so if you refine that area you will get more accurate values hovering the mouse.
The results on stresses are primarily computed on stress points, so if you are interested in stress results, then selecting a curve point and either using the tabular results or the command line or the Curves manager is the way to go.
I hope this helps.