yade-users team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxx>]

nejib hadda a écrit :
> Hi Bruno, 
>
>  I've been looking into random distributions available on YADE, and I
> think that the  boost::uniform_real<> is OK for my work, actually it is
> not the fact of having an uniform distribution that is worrying me,
> rather it is how to fix a range of radii [rmin, rmax] (Do you agree that
> my distribution is still uniform even if rmax is 10 time (in my study
> case) greater than rmin: it is the percentage of each r  (rmin<r<rmax)
> that makes my distribution uniform or not. is it that what it is meant
> by "uniform" on YADE or not?).
>
>   
Uniform means all radii in the range [rMean-rRelFuzz; rMean+rRelFuzz]  
are generated with equal probability. In TT interface, rRelFuzz is named 
stdDev (what does "Fuzz" mean btw?). Be carefull, it is not uniform in 
terms of mass distribution, and some specialists of granulometries would 
not call this uniform.

> Also I've met a problem that I could not explain when changing values
> into boost::uniform_real<0,1>, depending on the values entered YADE
> returns spheres coming out of the box or can't generate the requested
> number of spheres. Tell me if I should not change these values or if I
> need to do something else when changing them.
>
>   
Currently, positions are generated using the same uniform PDF (Prob. 
Density Function) as sizes. If random numbers are between 0 and 2 
instead of 0-1, positions will be between 0 and 2*sizeOfTheBox. You need 
to uncouple size PDF and positions PDF (the second should always be 
uniform I think).
Read the current makeCloud carefully, or perhaps try writing your own 
algorithm. It might help you to understand what is going on there.

Bruno