Re: [sparta-users] lambda/grid

[Paul T. <bos...@gm...>]

Thanks Steve,

From SPARTA #555 on GirHub, the new doc page with special notes added about lambda/grid is more explicit

After modification of my shell,  mfp and mct were computed without any trouble

The order of magnitude is in accordance with what we expected

Now next step

Jean-Pierre


________________________________
De: Steve Plimpton
Envoyé: Mercredi 19 mars 2025 20:56
À: Paul Trivard
Cc: nna...@en...; spa...@li...; Gallis, Michael A.
Objet: Re: [sparta-users] lambda/grid

Hi Jean-Pierre - there is a new SPARTA PR #555 on GitHub, which has a new doc page for compute lambda/grid,
and also fixes a bug in the code (Knudsen values could change based on whether lambda was included vs excluded as an output keyword).

If you can, please read the new doc page and see if it is clear to you.  Also you can test the source files in the PR, since that
is what we will merge into the public version, when ready.

Thanks,
Steve


On Wed, Mar 19, 2025 at 1:30 AM Paul Trivard <bos...@gm...<mailto:bos...@gm...>> wrote:
Thank you Steve

I will test / check shortly and send to all of you my feedback

Jean-Pierre
________________________________
De : Steve Plimpton <sj...@gm...<mailto:sj...@gm...>>
Envoyé : mardi 18 mars 2025 22:34
À : Paul Trivard <bos...@gm...<mailto:bos...@gm...>>
Cc : nna...@en...<mailto:nna...@en...> <nna...@en...<mailto:nna...@en...>>; spa...@li...<mailto:spa...@li...> <spa...@li...<mailto:spa...@li...>>; Gallis, Michael A. <ma...@sa...<mailto:ma...@sa...>>
Objet : Re: [sparta-users] lambda/grid

Sorry -- the issue here is that the latest compute lambda/grid doc page is very unclear on the correct syntax to use.
I think the confusion was introduced last fall when we started to use the new, more accurate formula.

I believe this is the correct syntax to use 2 computes as the source for nrho and temp:

compute              1 grid all species nrho
compute              1b grid all all temp

compute              2 lambda/grid c_1[*] c_1b[1] lambda tau
dump                 1 grid all 100 tmp.grid id c_2[*]

And this is the syntax to use 2 time-averaged fixes:

compute              1 grid all species nrho
compute              1b grid all all temp

fix                  1 ave/grid all 10 10 100 c_1[*]
fix                  1b ave/grid all 10 10 100 c_1b[1]
compute              2 lambda/grid f_1[*] f_1b lambda tau
dump                 1 grid all 100 tmp.grid id c_2[*]

The key point is that you need one compute to calculate a nrho for *each* species.
And another compute to calculate one thermal temperature for *all* the species.

This is going off of what the code does -- the doc page is completely opaque on this.
For the old formula, you could use one compute for both nrho and temp and access its 2 columns.

The lines above worked in your posted input script (modified to work around the surf data and variables
that weren't included).  So please try it out and see if the MFP and MCT values are what you expect.
Also, are the values you see for MCT in proper SPARTA time units?  The doc page is also unclear on that.

I will work on changing the doc page, and then verify with the folks who added this new formula that
the doc page and code are in sync and are the correct way to calculate the new formulas.

Steve






On Tue, Mar 18, 2025 at 4:29 AM Paul Trivard <bos...@gm...<mailto:bos...@gm...>> wrote:
Thank you Nnaemeka

my shell for "mfp and mct" calculation works if mymixture includes only one specie !

I didn't find any explanation of such behavior

is there any sparta_users having observed such type of thing ?

The example provided with January 2025 SPARTA release considers "Ar" as a mixture (one specie)

I tested the example with 2 species and the job failed

Jean-Pierre

________________________________
De : Nnaemeka Anyamele <nna...@en...<mailto:nna...@en...>>
Envoyé : lundi 17 mars 2025 23:26
À : Paul Trivard <bos...@gm...<mailto:bos...@gm...>>; spa...@li...<mailto:spa...@li...> <spa...@li...<mailto:spa...@li...>>
Objet : Re: [sparta-users] lambda/grid

Hi Jean-Pierre, I ran into a similar issue, earlier this year, here's Stan's response:

:Your "compute grid" command [that calculates nrho] needs to use the "species" keyword. Here are some examples:

https://github.com/sparta/sparta/blob/master/examples/mfp_mct/in.mfp_mct#L35-L41

https://github.com/sparta/sparta/blob/master/examples/surf_react_heatflux/in.surf_react_heatflux#L118-L122"

Hope this helps, I also found that there may be an issue with how you're indexing the fix vectors/arrays - you may or may not need an asterisk somewhere... but try adding a compute that specifically has the species key word in it, and use that compute/fix to calculate lambda/grid values.

Best wishes,
N.
________________________________
From: Paul Trivard <bos...@gm...<mailto:bos...@gm...>>
Sent: Monday, March 17, 2025 7:57:54 PM
To: spa...@li...<mailto:spa...@li...> <spa...@li...<mailto:spa...@li...>>
Subject: [sparta-users] lambda/grid

Dear_users,

"mct" calculation based on examples provided with the SPARTA release January 2025 was tested in the input file attached.

Special care to select "species" keyword for compute grid as recommanded by Stan.

However "ERROR remaining  at "lambda/grid" calculation with the following message (see partial log.sparta attached)
"Number of species does not match size of compute vector or array"

Any idea about the reason of such behavior

Many thanks by advance

Jean-Pierre

_______________________________________________
sparta-users mailing list
spa...@li...<mailto:spa...@li...>
https://lists.sourceforge.net/lists/listinfo/sparta-users