Re: [sparta-users] Using SPARTA grids for a nonhomogeneous temperature surface

[Steve P. <sj...@gm...>]

If you have 10,000 surface elements, how are you thinking you want to
assign a temperature to each one?

Steve

On Thu, Jun 1, 2023 at 8:45 AM Sarah Peterson <sar...@vk...>
wrote:

> Hi Steve,
>
> The goal is to have static temperature values along a single DSMC run
> assigned at the beginning of the simulation.
>
> Thank you for your help.
> Sincerely,
> Sarah Peterson
> Von Karman Institute
> ------------------------------
> *From:* Steve Plimpton <sj...@gm...>
> *Sent:* Thursday, June 1, 2023 4:53 PM
> *To:* Sarah Peterson <sar...@vk...>
> *Cc:* geo...@dl... <geo...@dl...>;
> spa...@li... <spa...@li...>
> *Subject:* Re: [sparta-users] Using SPARTA grids for a nonhomogeneous
> temperature surface
>
> You don't often get email from sj...@gm.... Learn why this is
> important <https://aka.ms/LearnAboutSenderIdentification>
> Warning: This is an email sent from outside the von Karman Institute. Do
> not open links or attachments unless you were expecting them and have
> verified they come from a safe source. Contact IT in case of doubts.
> Sorry for the confusion.  You are both correct that surf-style
> variables are not yet implemented.
> It is something we have thought about, but not yet done.  They would
> probably work like
> grid-style variables, where the center point of a triangle could appear in
> the formula so
> that the value of the variable could be a function of the location of the
> triangle as well as
> of time, some per-surf compute value, etc.  If that would be useful for
> your model, let me know.
>
> However, per-surf custom vectors are supported, which is what is used by
> fix surf/temp
> and is supported by the various surf_collide and surf_react models.  They
> are more general
> than a surf-style variable would be, since you could write any code you
> wish (e.g. a new or
> modified fix) to calculate a per-triangle temperature.
>
> You haven't explained how conceptually you want to set per-triangle
> temperatures, and whether
> they are static or dynamic values.  If you explain, I could likely advise
> the easiest way to do it.
>
> Steve
>
> On Thu, Jun 1, 2023 at 7:05 AM Sarah Peterson <sar...@vk...>
> wrote:
>
> Dear Steve,
>
> Thank you for the thorough explanation. It was very helpful.
>
> I'm not sure what I understand by a surf style variable. Looking through
> the documentation the only method I found to assign a temperature to an
> element involves using "compute surf" to get the total energy and using
> "fix surf/temp" to set some sort of temperature. That command uses group
> surf IDs to set the temperature of each section. So it can't assign
> temperatures on a single element basis, it can only handle groups of them.
> Is that what you meant by that? How does that method work to assign
> temperatures to individual elements?
>
> I'll admit this is my first time working with SPARTA so I am not
> completely familiar with every aspect of the program.
>
> Sincerely,
> Sarah Peterson
> Von Karman Institute
> ------------------------------
> *From:* geo...@dl... <geo...@dl...>
> *Sent:* Thursday, June 1, 2023 9:47 AM
> *To:* sj...@gm... <sj...@gm...>; Sarah Peterson <
> sar...@vk...>
> *Cc:* spa...@li... <
> spa...@li...>
> *Subject:* Re: [sparta-users] Using SPARTA grids for a nonhomogeneous
> temperature surface
>
> You don't often get email from geo...@dl.... Learn why this
> is important <https://aka.ms/LearnAboutSenderIdentification>
> Warning: This is an email sent from outside the von Karman Institute. Do
> not open links or attachments unless you were expecting them and have
> verified they come from a safe source. Contact IT in case of doubts.
>
> Dear Steve,
>
>
> thanks for the explanation, getting the terminology clear was helpful for
> me as well.
>
>
> RE: using a surf-style variable formula and the variable command: as far
> as I know, surf-style variables are still not implemented in the code? (at
> least looking at the master branch; lines 445-446 in variable.cpp)
>
>
> Mit freundlichen Grüßen
> Dr. Georgii Oblapenko
> Postdoctoral Researcher
> Alexander von Humboldt Stiftung Fellow
> Institut für Aerodynamik und Strömungstechnik
>
> Deutsches Zentrum für Luft- und Raumfahrt (DLR)
> ------------------------------
> *From:* Steve Plimpton <sj...@gm...>
> *Sent:* 31 May 2023 22:33:12
> *To:* Sarah Peterson
> *Cc:* Oblapenko, Georgii; spa...@li...
> *Subject:* Re: [sparta-users] Using SPARTA grids for a nonhomogeneous
> temperature surface
>
> Hi - the terminology in this thread is somewhat confusing.  There is
> only a single volumetric grid in SPARTA,
> though it can adapt over time.  There is also a collection of static
> surface elements.  Grid cells can be assigned
> to grid groups.  Surface elements can be assigned to surface groups.  In
> both cases a "group" is simply a
> set of one or more surface elements (or grid cells).  There is a hard
> maximum of 32 grid groups and 32 surface
> groups.  We don't know of any use cases where more groups are needed.
>
> What you are asking about (I think) is assigning temperatures to
> individual surface elements.  That has nothing to do with
> groups.  A surface temperature is only currently used by surface collision
> and reaction models (surf_collide and surf_react commands).
> All the existing models only define it as the resolution of a single
> surface elements.  If you want the temperature to vary
> across a single triangle, then you should simply break the big triangle
> into smaller triangles (which you input to SPARTA), and define a
> temperature for
> each small triangle.  You can define per-triangle temperatures in 2 ways.
> Using a surf-style variable formula and the variable command.
> Or using the fix surf/temp command which computes termperatures (per
> triangle) via a formula using energy deposited
> onto the triangle as tallied by a compute or fix.  If you have some other
> model in mind for computing the temperature
> of a triangle, then you could implement a new fix similar to fix
> surf/temp.  Or posslbly you could add it as option to
> the existing fix surf/temp command.  E.g. use a different formula.
>
> Steve
>
> On Wed, May 31, 2023 at 7:18 AM Sarah Peterson <sar...@vk...>
> wrote:
>
> Thank you for the link. That is unfortunate, I was hoping it was an
> arbitrary constraint due to concerns about computational power and memory
> usage.
>
> I found this message (Thread: [sparta-users] Modify SPARTA to allow more
> than 32 surface groups? | SPARTA (sourceforge.net)
> <https://sourceforge.net/p/sparta/mailman/sparta-users/thread/CAMkvQcT2u7Lst3r3WxnEPpnM7%2B2%2B0_nrmPU4Svg5G4bj-HDMRw%40mail.gmail.com/#msg36445633>)that
> had a similar issue to me though I'm not sure I understand the proposed
> solution. I thought surfaces were also subject to the 32-group limit as
> well.
>
> Thank you for your help.
> Sincerely,
> Sarah Peterson
> Von Karman Insititute
> ------------------------------
> *From:* geo...@dl... <geo...@dl...>
> *Sent:* Wednesday, May 31, 2023 2:46 PM
> *To:* Sarah Peterson <sar...@vk...>;
> spa...@li... <spa...@li...>
> *Subject:* Re: Using SPARTA grids for a nonhomogeneous temperature surface
>
> You don't often get email from geo...@dl.... Learn why this
> is important <https://aka.ms/LearnAboutSenderIdentification>
> Warning: This is an email sent from outside the von Karman Institute. Do
> not open links or attachments unless you were expecting them and have
> verified they come from a safe source. Contact IT in case of doubts.
>
> Dear Sarah,
>
>
> this is a relevant discussion:
>
> https://github.com/sparta/sparta/issues/260
>
>
> So it seems there is no way around that except by writing your own code to
> assign surface temperatures.
>
>
> Mit freundlichen Grüßen
> Dr. Georgii Oblapenko
> Postdoctoral Researcher
> Alexander von Humboldt Stiftung Fellow
> Institut für Aerodynamik und Strömungstechnik
>
> Deutsches Zentrum für Luft- und Raumfahrt (DLR)
> ------------------------------
> *From:* Sarah Peterson <sar...@vk...>
> *Sent:* 31 May 2023 12:02:25
> *To:* spa...@li...
> *Subject:* [sparta-users] Using SPARTA grids for a nonhomogeneous
> temperature surface
>
> Hello,
>
> I am a student researcher working with your program to try and couple it
> with a material solver to model systems with a nonhomogeneous temperature
> distribution. However, I am having issues with the number of grids that
> SPARTA allows. My initial plan was to use the "group surf" command to
> generate grids on the surface and assign them different temperature values
> with "fix surf/temp", but SPARTA only allows for 32 grids to be active at
> one time. On large objects this would be rather imprecise. Is there a
> method to bypass this limit or an alternative method to reach my desired
> goal?
>
> Thank you for your help.
> Sincerely,
> Sarah Peterson
> Von Karman Insititute
> _______________________________________________
> sparta-users mailing list
> spa...@li...
> https://lists.sourceforge.net/lists/listinfo/sparta-users
>
>