Min/Max Grain Size

[janzenchoi]

Hello,

The distribution of my equivalent diameters are defined such that the standard deviation is larger than the mean. This introduces several problems, specifically very tiny and very large grains. I tried defining lower and upper bounds for the grain sizes, using the interval keyword, as shown below.

neper -T -n from_morpho -morpho "diameq:interval(0.0174,2.0158):lognormal(0.20766,0.27155),1-sphericity:interval(0.02316,0.57725):lognormal(0.21411,0.09016)" -dim 3 -domain "cube(1,1,1)" -statcell diameq,sphericity

However, the grain sizes were still out of the specified bounds. As such, I was wondering whether it is possible to define a minimum and maximum grain size with the lognormal distribution, and what that command would look like.

Thanks in advance.

[rquey]

diameq:interval(0.0174,2.0158):lognormal(0.20766,0.27155) is not proper syntax. diameq:interval(0.0174,2.0158),diameq:lognormal(0.20766,0.27155) works in terms of syntax, but would not be very good practice.

In your case, you can specify endpoints using from and to: diameq:lognormal(0.20766,0.27155,from=0.0174,to=2.0158), as described in Statistical Distributions.

[janzenchoi]

Thanks for the response, @rquey.

I took your advice, and ran the following command.

neper -T -n from_morpho -morpho "diameq:lognormal(0.20766,0.27155,from=0.0174,to=2.0158),1-sphericity:lognormal(0.21411,0.09016)" -dim 3 -domain "cube(1,1,1)" -statcell diameq,sphericity

However, it gave me 259 grains. Previously, when I ran the command without the bounds, the tessellation would have about 25 grains. Only a few of those grains would be outside of the bounds (i.e., [0.0174, 2.0158]). As such, I feel like there shouldn't be this many grains.

Do you know where I went wrong?

[rquey]

There was indeed a bug. Please use version 4.5.1-2.

Note that, in principle, when using from and to, you should preferably make sure that the resulting truncation does not change the mean value of the distribution (it may be the case with your command, I'm not sure). Otherwise, a correction has to be made and some values may go slightly out of bounds...

Here is what I now get with your command: 25 cells, min = 0.023, max = 1.088, mean = 0.217, standard deviation = 0.237.

[janzenchoi]

Updating Neper to the newest version does help. Thanks again for all the help.