key words in GUI

[Fabio89]

Hello everybody,

I need some information about key words filling in the GUI, in order to complete more correctly my calculation.
Using "Particles and droplets tracking", I'm not sure about key words meaning in chapter 7.8 in the user's guide and key words meaning of boundary condition for particles; above all

in PARTICLES AND DROPLETS TRACKING/STATISTICS:
-what does "number of particles cluster" mean?

in BOUNDARY CONDITIONS/PARTICLES BOUNDARY CONDITIONS:
-what does "number of classes" mean?
-giving X-value to number of classes for inlet, for each class I can set a "cluster number for particle" and a "statistical weight"; what do these key words mean? how do they affect calculation and output files?

Thanks in advance for any help

Best Regards,

Fabio

[Fabio89]

Any help????
I acknowledge these are simple questions but for me very important.

Thank you one more time,

Fabio

[guingo]

Hello Fabio,

sorry for the late (and short !) answer, I am currently out of office.

In the Lagrangian module, a particle "class" is a group of particles that share the same set of physical properties (diameter, density for instance).

- The "statistical weight" is a real number which basically tells how many "physical" particles are represented by a "numerical" particle. This notion was introduced to enable the user to run a calculation with realistic particle mass flow rate, which often involve great numbers of particles hard to simulate. For instance, if we want to simulate a flow rate of 10^7 particle/second, a way to do this is to simulate 10^5 particles/second with a statistical weight of 100.

- The number of particle clusters is related to the possiblity of calculating statistics of "groups" of particles (which may or may not be the same as the classes of particles). Unfortunately, my experience with this kind of post-processing is fairly limited .

Best regards,
Mathieu

[Fabio89]

Hello Mathieu,

don't worry any way for your late, when I posted my topic I was on holiday and next days I came back to home
However Your help is been very important; I cannot immagine the meaning of "statistical weight" and now I'm clearer about "number of particles cluster". I supposed it could mean something like you wrote.

So, Thank you very very much for your useful help here and on the other post. I read it now.

We will hear next time; maybe I'll able to help you

Regards,

Fabio

[Fabio89]

Hi Mathieu,

I looked better at user'e guide and from you reply about "statistical weight", I uderstood that using a n stat_weight for a particle class, I have on a cell (volume output) a total number comes from the sum of statistical weight of all particles on the cell; for example if I set statistical weight = 1.0 for my particle class and in output I have stat_weight = 12 for that cell, it means that I find 12 particles (of my class) in that cell.
Right???

So I tried setting:

-Threshold for statistical weight = 1.0______in order to do not account cells whit no particles

-Boundary conditions -> particles boundary conditions
(wall) particles deposit+memory_________in order to keep particles in the calculation domain
(inlet) number of classes = 3

class 1
statistical weight = 1.0
fliud velocity
diameter = 1e-06
variance = 1e-07

class 2
statistical weight = 1.5
fliud velocity
diameter =2.5e-06
variance = 1e-07

class 3
statistical weight = 2.0
fliud velocity
diameter = 1e-05
variance = 1e-06

but I have this ERROR:

SIGFPE signal (floating point exception) intercepted!

Call stack:
1: 0x7f3ecb6debac <uslabo_+0x18ec> (libsaturne.so.0)
2: 0x7f3ecb6a6173 <lagcel_+0x166f> (libsaturne.so.0)
3: 0x7f3ecb6dad98 <lagune_+0x35f8> (libsaturne.so.0)
4: 0x7f3ecb3c8193 <caltri_+0x59a3> (libsaturne.so.0)
5: 0x7f3ecb3a29e3 <cs_run+0x843> (libsaturne.so.0)
6: 0x7f3ecb3a2cc5 <main+0x1f5> (libsaturne.so.0)
7: 0x7f3ec7b41c4d <__libc_start_main+0xfd> (libc.so.6)
8: 0x40a359 <> (cs_solver)
End of stack

Could you help me!?!?!?!

Thakyou for your more attention

Regards,

Fabio

[guingo]

Hello Fabio,

I think your understanding of the statistical weight is OK

uslabo (for USer LAgrangian BOundary conditions) is the routine managing the particle-boundary interactions; the error is a priori not related to your choice concerning the statistical weight parameters (which seems totally acceptable to me).

I am out of office till late september without access to Code_Saturne, so I can only make hypothesis, but can you please make the following tests and let me know the outcomes :

- switch "deposit+memory" to "deposit" for your wall conditions.

and

- keep "deposit+memory" but unactivate the "boundary statistics" calculations.


Best regards,

Mathieu

[Fabio89]

Good afternoon Mathieu,

Thank you for this new help, however I have just launched my last calculation, regarding your advices.

-Keeping "deposit+memory" but unactivate the "boundary statistics" calculations I had the same ERROR refered in my other post http://code-saturne.org/forum/viewtopic ... 6&start=10;

-Switching "deposit+memory" to "deposit" for my wall conditions I had this ERROR message:

SIGFPE signal (floating point exception) intercepted!

Call stack:
1: 0x7f417a844bac <uslabo_+0x18ec> (libsaturne.so.0)
2: 0x7f417a80c173 <lagcel_+0x166f> (libsaturne.so.0)
3: 0x7f417a840d98 <lagune_+0x35f8> (libsaturne.so.0)
4: 0x7f417a52e193 <caltri_+0x59a3> (libsaturne.so.0)
5: 0x7f417a5089e3 <cs_run+0x843> (libsaturne.so.0)
6: 0x7f417a508cc5 <main+0x1f5> (libsaturne.so.0)
7: 0x7f4176ca7c4d <__libc_start_main+0xfd> (libc.so.6)
8: 0x40a359 <> (cs_solver)
End of stack


To sum up, for my project I need to know how many particles (of my 1000 for example) impact and deposit on a sampling plate, after throwing an inlet tube and a nozzle.
How do you think can I do?
As on the user's guide I understood, without option "memory", particles after interaction with my wall (sampling plate) they disappear from the domain and I don't see them, right???

Regards,

Fabio

[guingo]

Hi Fabio,

Concerning the error, could you please try to inactivate the calculation of the boundary statistics ? It would help to locate the part of uslabo that causes this behaviour. Moreover, if you can run calculations with a version of Code_Saturne compiled in debug mode, it could be useful to run the command addr2line with the error address in uslabo and your executable file (cs_solver) as arguments to find the line of the routine causing the error (Note : some times this command just returns "??" for some reasons..)

For your particular problem, one strategy that may work is to program a counter in uslabo that would increment once a particle deposit on the zone of interest (uslabo is called when a particle crosses a border). For this purpose, if you use "deposit+memory" for your sampling plate, try to add in uslabo something like

!===============================================================================
! 4. Deposition of the particle, which remains in memory
!===============================================================================

else if (iusclb(kzone).eq.idepo2) then

if kzone = <the number of the zone representing your sampling plate> then

< increment my counter>

< print my counter>

endif

Best regards,

Mathieu

[Fabio89]

Hi

I inactiveted Boundary Statistics but I had the ERROR list here:
http://code-saturne.org/forum/viewtopic ... 6&start=10
About to compile CS in a debug mode I ackonwledge I'm not able, so I prefer to not use it, in order to not have other issues.

For the program of a counter in uslabo, I really think I can do it; concerning my counter, it is sufficient to write:
< increment my counter>

< print my counter>

endif
or do I need to inizialize my counter writing someting else?

At last, past days I re-studied "impactors manual design" and the article "CFD Study on Compound Impaction in a Jet-in-Well Impactor" and I noticed that CFD runs, setting Unsteady flow with a Turbolent Model (often k-ԑ), instead I set Unsteady flow with Laminar flow; my velocity profile show absurd values.
I'm sure I made a mistake. What do you think?

Best Regards,

Fabio

[guingo]

Hi Fabio,

inactiveted Boundary Statistics but I had the ERROR list here:
http://code-saturne.org/forum/viewtopic ... 6&start=10

OK , but for this particular error, either activate "displacement mode" with just one particle to track not to slow the calculation (if you use "deposit+memory") or use "deposit" instead of "deposit+memory"

For the program of a counter in uslabo, I really think I can do it; concerning my counter, it is sufficient to write:
< increment my counter>

< print my counter>

endif
or do I need to inizialize my counter writing someting else?

Maybe I was unclear but "< increment my counter> " and "< print my counter>" are not valid instructions, you need to properly declare and initialize an integer variable and makes it increment when the requested condition (i. e. a particle impact) is fulfilled

At last, past days I re-studied "impactors manual design" and the article "CFD Study on Compound Impaction in a Jet-in-Well Impactor" and I noticed that CFD runs, setting Unsteady flow with a Turbolent Model (often k-ԑ), instead I set Unsteady flow with Laminar flow; my velocity profile show absurd values.
I'm sure I made a mistake. What do you think?

(I can give you just very general guidelines) If your question is related to the use or not of a turbulence model, I would recommend to first calculate a Reynolds number relevant to you case and apply (or not) a turbulence model based on this value; for the values of your velocity fields, also make sure your calculation is converged . If you are reasonably sure of your numerical and physical set-up, and if you are not satisfied with your results, maybe it will be interesting to open a new topic describing your problem in detail.