yade-users team mailing list archive

[Václav S(milauer <eudoxos@xxxxxxxx>]

>> Your spheres have some microscopic young's moduli, but due to the
>> disctribution of interactions you get macroscopic modulus that can be
>> different; and I want to compensate that). I want to make sure that
>> given a plane, sum of "surfaces" of all interactions (cylinders between
>> spheres with the radius of the smaller sphere, right?) is equal to
>> nominal, macroscopic surface of specimen. It depends on sphere radii
>> discribution, for sure; perhaps it can be calculated analytically for
>> regular arrangements. For other cases, the simulated rigidity may be
>> artificially higher/lower. I haven't tried to quantify that yes, though.
>>     
>
>
> If I understand what you mean, you want to compute analytically the 
> macroscopic young modulus as a function of the microscopic young 
> modulus. I am not sure there is currently a general way to do that for 
> random assembly, it is not easy at all. There were formula like that in 
> SDEC: the input was a macroscopic young modulus and SDEC computed 
> automatically the microscopic young modulus attributed to each 
> particles. But in reality it didn't work, the macroscopic young modulus 
> wanted was not reached.
> The formula used were based on a paper from Cambou, but the formula, I 
> think, were valid only for a given fabric corresponding to the ones 
> studied by Cambou.
> You can find more details in the PhD thesis of Sebastien Hentz who used 
> all that to simulate concrete with SDEC (in which there was a radius of 
> interaction greater than 1). In addition maybe you will find help about 
> interaction laws to use etc... in this thesis, unless you have already 
> read it.
>
>   
Yes, I have read it, thanks. I don't want to derive that analytically, I
will plug intput and output into a neural network and train it to give
me right interaction radius to have the same microscopic and macroscopic
modulus (maybe, it is in collaboration) for different distributions - in
some very primitive way at the beginning, though.

_______________________________________________
Yade-users mailing list
Yade-users@xxxxxxxxxxxxxxxx
https://lists.berlios.de/mailman/listinfo/yade-users