yade-dev team mailing list archive

[Jerome Duriez <Jerome.Duriez@xxxxxxxxxxx>]

Hi, I'm coming back on capillary files. So we faced last time situations where considered Laplace solutions were "wrong": because a zero volume meniscus was computed for contacting spheres. Which is finally no-sense.

This error might be easy to spot, especially with the LOG_ERROR you, Bruno, wrote in the code.

But... Now I wonder if other cases might lead to non-sense results ? This last error concerned contacting spheres: case D =0 (or adimensionned D* = 0) and is easy to spot for this reason.
Could such problem appear for non-contacting spheres ? (yes, probably ?) How to spot there is an error in this case ?

Maybe there is no solution in the current state of the code / data files, but any remark is welcome.

Jerome




________________________________
From: Yade-dev [yade-dev-bounces+jerome.duriez=ucalgary.ca@xxxxxxxxxxxxxxxxxxx] on behalf of Bruno Chareyre [bruno.chareyre@xxxxxxxxxxxxxxx]
Sent: October-30-14 8:54 AM
To: yade-dev@xxxxxxxxxxxxxxxxxxx
Subject: Re: [Yade-dev] Some cleaning in capillary law

On 29/10/14 22:37, Jerome Duriez wrote:
Ah, this is a clarification... I knew this was succion value dependant, but did not catch that the zero volume is inconsistent... Then, would it possible to know a validity range of the capillary data files ?

Only trial and error can tell at this point. Keep in mind that the max value is defined in terms of non-dimensional pressure pR/gamma (p: succion, R: radius, gamma: surface tension), not in terms of p.
I must say I'm not in a hurry to fix those things since we are in the process to introduce another capillary model which will deprecate this one (the work of Caroline Chalak).

A second thing is that I was thinking to commit, with the reverted fCap, the use of alpha * -currentContactGeometry->penetrationDepth for D value (instead of computation of l. 112, (*) ).
It would be equivalent to current computation, and maybe useful for periodic simulations. I'm currently using it (mainly for non-periodic cases, yet)

Yes, it makes no sense to compute penetration again, as suggested in the code comment.
You are welcome to change this.

B