how these expected line (orange colors) being plotted in graph? it seems few of additional parameter being assumed and fixed inside MFront.
It is TEFL reference graph associated with the parameters. I just digitalize the graph to be able to plot it together with my results for comparision.
Engauge-digitizer
Do you mean parameters to define the behaviour or numerical parameters used by the integration algorithm?
Wait, I think you are asking something different. How to automate that plot given the parameters.
You have a line: 2 Parameters.
Cap is an ellipse in the p q plane: 3 parameters. (Major axis, Eccentricity, and center.)
Then itâs imposed that both geometries connect (at the Ellipse uppermost point) which eliminates one parameter from the equation.
TEFL model prefers to eliminate the Eccentricity.
Yield Surface reduces to 2+3-1 = 4 parameters. , , , .
probably needed further to look up the code, since it have relation to transition surface with some parameter adjusted.
documentataion in DruckerPragerCap material not mention it, but result plotted shown.
I donât think so. âat the Ellipse uppermost pointâ is implicit.
There is another condition to completely constrain the geometry. Ellipse axis aligned to p q. I assumed that was obvious.
Strictly 2+4-2 = 4
iâm not in detail and further investigate, but plotted result (blue lines) shown transition surface.
Yep. In fact that could explain that small offset. Seems the upermost condition is not enforced.
I wonder why he calls R the Eccentricity of the ellipse and not the Aspect Ratio ??
According to this equation the Elipse is centered at pa:
a=abs(pb-pa)=pa-pb (Minor Axis)
b=d-tan(beta)*pa (Major Axis)
Up to here seems ok âŚbut then
Eccentricity of the elipse is = c/b = sqrt(b2-a2)/b and he is imposing Eccentricity = a/b (Aspect Ratio)
I donât understand this point.
EDITED: And yes, this material behaviour needs more parameters if one wants to get an smooth transition with continuity in the first derivative. Abaqus explain that point. Actual model seems to enforce connection between Fs with Fc directly at the uppermost point of the elipse (Major axis) without transition Ft.
itâs relationship from geometry of the graph
I have reduced the maximum time step again from 0.01 to 0.002 (just for some of the worst points cause itâs time consuming) and agreement is excellent now.
I guess connection at the uppermost point of Ellipse with the line is right and It is my fault I donât fully understand the math steps behind (details are not provided for the R-formula).
Both âRâ & âalphaâ factor in Abaqus is ratio needed to adjust by user, since only âp_aâ can be known. It seems these factor is taken from experimental calibration, usually for concrete âRâ factor is in ranges from 0.1 to 0.25
Hi,
Is there webpage similar to the one for Drucker Prager material model but for Mazars and Unilateral Mazars.?
On page where the model and parameters are described? Iâm lost in many pages inside TEFL, but none seems to address to the collection of material behaviors and their description.
Tanks
maybe, cause of Mazars material model is well-known by paper and book. Also, it has been adopted by many FE solvers.
confirmed for MohrCoulombAbboSloan material models, result of CalculiX & MFront given as expected results (SF=1.283), below OpenGeoSys (SF=1.301) and above Plaxis (SF=1.265)
1 Like
hi, it seems MFront material only possible using single material in a model. Below example case, CalculiX material Mohr-Coilomb have no problem since it allowed to rename for each layer.
Is this limitation can possible to improves (i.e ignore suffix) for using the same multiple material of MFront libraries? thank you.
*Material, name=@CALCULIXBEHAVIOUR_ELASTICITY@1
** The material properties are given as if we used parameters to explicitly
** display their names. Users shall replace those declaration by
** theirs values
*User Material, constants=2
150000,0.3
*Material, name=@CALCULIXBEHAVIOUR_ELASTICITY@2
*User Material, constants=2
200000,0.3
1 Like
another additional confirmation for material models known working properly: DruckerPrager, DruckerPragerCap, Mazars and FichantLaBorderie(attached)
peak ultimate loads are well agree and fit experimental bounds, maximum deflection is a little bit lower (probably due to my mesh and element type selection).
btw, thereâs interesting material of ductile damage (GursonâTvergaardâNeedleman) available from outside sources, hopefully this model can be included in next release of MFront/CalculiX distribution. Thank you.
1 Like
another interesting ones (Hoek-Brown) is under development by OpenGeoSys - link
*edited: seems it already done for HoekBrownC2.mfront and hopefully will be available in next CalculiX MFront integration, thank you.
@Author Mehran Ghasabeh, Dmitri Naumov, Thomas Nagel;
@Date 15 / 11 / 2023;
@Description{
Implementation of non - associated flow rule of the Brown yield criterion,
which is derived based on the work of Hoek and Brown,
https : // doi.org/10.1016/j.jrmge.2018.08.001.
2 Likes