Capillary contact law

Even i can understand the results i got from my simulation using the this capillary contact model, and can justify the reasoning for not capturing the yield stress and capillary pressure relationship based on the degree of saturation (given the model works only in the pendular regime). But want to make sure i am thinking in the right direction.

Thanks...!!

Hi,

It seems to me you did not correctly reflect the behavior of YADE's capillary model (Law2_ScGeom_CapillaryPhys_Capillarity) in your question. For strictly overlapping particles, capillary forces still exist (where did you see the contrary ?).
Maybe you were confused by the idea that this overlap is not taken into account during computation of capillary forces through solving Laplace-Young equation:

- in reality two liquid bridges between strictly overlapping or just touching particles would be different (even with the same suction, surface tension, contact angle and particles radii)
- nevertheless, YADE capillary model will give you the exact same bridges / capillary forces for these two cases

Jerome

Sorry if i understood wrongly, at (delta)n = 0, fn=fcap (right...), then for (delta)n > 0; fn increases with delta. The problem i faced was when i compared the capillary stresses (i used capillary get tensor) with contact stresses (get stress tensor), the capillary stresses were very small and remained almost same for the entire loading path. The stresses at boundary matched with my contact stresses. So based on what i got i can say that there was not much contribution due the capillary for isotropic compression.

Please correct me, if am thinking in wrong direction.

Thanks

If you observe a negligible influence of unsaturated conditions (with negligible capillary stresses in front of other stress contributions), the most probable reason is you have too big particles. In reality and in YADE, only fine soils with small particles are sensitive to unsaturated conditions. What diameter (and confining pressure, also) are you using ?

I am using particle radius ( monosized) as 0.02 mm (E = 50 MPa). I generated my packing with sigmaiso= 15 kPa for a porosity 0.40 and then introduced the capillary pressure and further loaded my packing isotropically to achieve a target pressure of 1500 kPa.

I have done otherwise too, at particle assembly itself i introduced liquid bridges and isotropically compressed to a target porosity of 0.40 at sigmaiso = 15 kPa. and then loaded to 1500 kPa. In both the cases i got same response independent of my capillary pressure.

Thanks

This 1500 kPa is only one case; even for smaller sigmaiso < 60 kPa, void ratio are fairly constant for each condition with capillarity. But when i am doing shearing, the model works perfect.

From my experience in doing experiments i can say that, in pendular regime where degree of saturation is very low, yield stress is independent of matric suction (even for different void ratios) but, over there the imposed matric suctions are very high in order of several kPa.

So, what i infer is that the definition of matric suction in soil mechanics in totally different from the capillary pressure (which is an excess pressure due to air and water phase)....i might be wrong also in thinking and deducing such conclusions, but physical meaning of both the terms are different, i reckon. As, i am from soil mechanics back ground my definitions seem contrary (which i agree...) and more appropriate representation of the results should be in term of degree of saturation and water content (as presented in literature for capillary model using YADE).

Since, i am very new to this YADE so want to make sure that my understanding is correct. Any help is appreciable.

Thanks

Hi,
1) BBM with Χ=Sr is irrelevant to the pendular regime. So,
2) no, there is no way to reach BBM with Χ=Sr by simulation of pendular regime.

Your description of the model is confusing.
>" at (delta)n = 0, fn=fcap (right...), then for (delta)n > 0; fn increases with delta."

Instead I would put it this way:
- at (delta)n < 0, fn=0, fcap=something.
- then for (delta)n > 0; fn increases with delta, fcap = something"
where fn is the contact force and fcap is the capillary force, which come from two independent models.

If the capillary effects are negligibly small you can simply increase surface tension.

Bruno

Hi,

I think the best way to understand capillary effect on particle forces is a force-distance-plot. Fortunately I have one in my PhD. You can have a look at page 35 (pic. 3.10).

http://nbn-resolving.de/urn:nbn:de:bsz:14-qucosa-218045

btw, the plot was done with output from Law2_ScGeom_CapillaryPhys_Capillarity ;)

Regards,

Christian

Thanks for your answers....

@Christian, even i got similar results, when my penetrations depths are positive (isotropic compression case) the capillary forces remain constant. So if i compare two scenarios having different capillary pressure, then my void ratio-log p (where p is the mean contact stress) relation is same irrespective of initial capillary pressure, since my kn is constant and moreover Fcap is very small. Unlike your results my Fcap values are much smaller, which might be another reason. Please confirm whether my understanding is right or not.

So, probably i would consider Bruno's suggestion, to increase the surface tension and come back with some data over this problem.

Thanks

Amiya

My mean capillary stress is several order smaller than the mean stress. Is it correct. My capillary pressure input is 10000 Pa, E = 50 MPa. rMean = 0.02 mm.

> the capillary forces remain constant. So if i compare two scenarios having different capillary pressure, then my void ratio-log p (where p is the mean contact stress) relation is same

This is not true.
The capillary force being independent of un does not imply that the capillary force (or overall "the results") is independent of capillary pressure.
However, the fact is that capillary forces are only slighty changing with capillary pressure - and whatever the sign of un. This can explain the similar void ratio-log p.
Anyway, this analysis is pointless as long as capillary effects are negligibly small.

Bruno

@Bruno, Thanks for your insights, when i started this analysis, my hypothesis was i should get different response based on my capillary pressure. Since now it explains why i am getting similar response. Is there any way such that i can create a scenario where i get different response depending on the input capillary pressure.

Thanks

Amiya

>Is there any way such that i can create a scenario where i get different response depending on the input capillary pressure

In the pendular regime, no. It will never be very different. It is a physical feature also observed in experiments (as long as the experiments are in pendular state).

Beyond the pendular state the effect of capillary pressure is more evident, but you need a completely different model.
One possible option is the 2PFV method, which now has an example script here:
https://github.com/yade/trunk/blob/master/examples/FluidCouplingPFV/drainage-2PFV-Yuan_and_Chareyre_2017.py

Cheers

Bruno

@Bruno is this two phase engine "unsaturatedEngine" included in the current yadedaily..??

Best
Amiya

No it is not part of the precompiled package yet.
It is available in the source code and compiled optionally (cmake option -DTWOPHASEFLOW).
Bruno

This question was expired because it remained in the 'Open' state without activity for the last 15 days.