[Archive] RE : InitPot feature

system · 19 October 2004 22:00

Message by julien:
Extracted from old forums.
Previously posted by Thiebault: Wed Oct 20, 2004 1:37 pm Post subject: OK, I think I fixed the problem : in fact I didn’t compile/packaged properly the class… So I can confirm that it works, but only after applying the patch.
Thanks
Benoit

system · 19 October 2004 22:00

Message by julien:
Extracted from old forums.
Previously posted by Mike: Fri Sep 10, 2004 2:18 pm Post subject: initPot feature

Hi,
I was trying to set an initial S/C potential with initPot, but I could not get it running. I tried:

set the initPot directly in UI.
set the SCdirPot values in each phys. surface with initPotFlag=2 as requested in the documentation.
1+2 combined. I always got the S/C pot starting at 0 V.
Could anybody comment on that one?
And btw., how are the different electric nodes handled now (I understand the internal circuitry is not used yet) ?
Thanks in advance for answering this question…
best,
mike
_____________
Michael Klicker

system · 19 October 2004 22:00

Message by julien:
Extracted from old forums.
Previously posted by Roussel: Tue Sep 14, 2004 12:11 pm Post subject:
If you want to define the initial global potential, you have to set initPotFlag to 1, cf. its description:
“flag to define initial pot: 0 => set to 0, 1 => set to global initPot, 2 => set to local potential defined as SC Dirichlet condition”.
It should really be working, Erik used that for Cluster in July here (I implemented that for him then).
It is true you should be able to define it locally through SCDirPot local field by setting initPotFlag to 2, but this feature is not yet implement as you can see on the html documentation “DocSpisNum\HowTo\Controlling NUM from UI.html” (in the right-most column “in use”)
JF
_____________
Jean-François Roussel

system · 19 October 2004 22:00

Message by julien:
Extracted from old forums.
Previously posted by Mike: Mon Sep 20, 2004 5:58 pm Post subject:

Hi,
well I tested that, and I just got random results: Sometimes, the S/C doesnt start at 0 V (I got one case with -0.45V, for whatever reason), sometimes it does (initpot=+10V, initpotlfag=1). Are you sure you included the code into the rc1 release ? Because I could not get this feature to run up to now…
cheers
mike
_____________
Michael Klicker

system · 19 October 2004 22:00

Message by julien:
Extracted from old forums.
Previously posted by Roussel: Tue Sep 21, 2004 5:01 pm Post subject:
Hi Mike, our best alpha tester!
you’re right, it does not always work correctly.
Erik only used it in the case of interest to him, constant potential (electricCircuitIntegrate=0, which is not the default).
So it was not fully tested as I thought.
When time integration of SC circuit is requested things get spoiled, because my inits are too superficial (only surface potentials were initialised not capacitor charges).
Solutions for you are:

work like Erik (electricCircuitIntegrate=0) if it answers to your needs
use the following patch:

add 2 lines to the setPot(...) method of RCCabsSC.java class, which becomes: public void setPot(ScalSurfField potential) { pot = potential

mapSurfToCirc(pot, circ.getPot(), 0)

circ.setPot(circ.getPot())

}
NB: the SPIS/NUM classes are contained in the spis.jar file in SpisNum/build/Lib directory.
You have to:

untar/unjar them: jar xvf spis.jar
make the modification in spis/Top/SC/RCCabsSC.java file
recompile (e.g. simply as "javac spis/\*/\*/\*.java")
tar them back: "jar cvf spis.jar org spis" (org are libraries)

Don't hesitate asking if you're in trouble. After the patch, things should be ok for a conductive SC, and approximately ok for an dielectric-coated SC. In the 2nd case, ground potential is not modified, but coating surface potential is => the ground potential printed in the shell window looks not to be initialised, but if you look at the local potential surface map, it is ok. The screen logs will be improved in the next release when SC circuit capability are improved soon (several grounds, in progress), I don't want to get into that before. NB: your feeling to get random results certainly comes from the fact that the initial potential is not printed. Only the potential after one integration step is printed => it already evolved in a possibly random-looking way. Best, JF \_\_\_\_\_\_\_\_\_\_\_\_\_ Jean-François Roussel

system · 19 October 2004 22:00

Message by julien:
Extracted from old forums.
Previously posted by Mike: Fri Sep 24, 2004 9:54 am Post subject: Hi everybody,
the patch works pretty fine!
Thanks for the help!
cheers
Mike
_____________
Michael Klicker

system · 19 October 2004 22:00

Message by julien:
Extracted from old forums.
Previously posted by Thiebault: Wed Oct 20, 2004 9:56 am Post subject:
Hi everyone,
I tried to run a simulation with the example file Sphere2.geo using a fixed value of the spacecraft potential… and it doesn’t really work :

I set the initPot variable to 20
I set the initPotFlag to 1
and I swicthed of the integration of the S/C circuit (electricCircuitIntegrate = 0)

However :

the potential always starts at 0V (whatever material I use)
the potential is still floating...
I tried with and without the patch suggested in this thread, there is no difference

My version of SPIS is 3.00.01rc1... Thanks for your help Benoit

system · 19 October 2004 22:00

Message by julien:
Extracted from old forums.
Previously posted by Thiebault: Wed Oct 20, 2004 10:00 am Post subject:
… I forgot… I have a special bonus question…
what if I want to have several fixed potentials on the same spacecraft ?
I saw with Julien that there is a way in the plasma editor to set the potential of a group at a certain value, but it is apparently not interfaced yet with SPISNum… what should I change to make it work ?

system · 19 October 2004 22:00

Message by julien:
Extracted from old forums.
Previously posted by Roussel: Wed Oct 20, 2004 10:11 am Post subject:
Hi Benoit,
I am rather surprised by your question.
Your parameters look to be the ones working, tested by myself and Michael.
Could you please check with your neighbour Michael, which are the parameters differences between your case and his case, and tell me.
Thanks
JF
_____________
Jean-François Roussel

system · 19 October 2004 22:00

Message by julien:
Extracted from old forums.
Previously posted by Thiebault:Wed Oct 20, 2004 2:41 pm Post subject:
Now that this works, I tried to use it in different ways :

by changing the initPot value
by using different plasma properties in the Group Editor

Apparently, changing the plasma property of a group doesn't have any influence on the potential taken into account during the simulation. For a S/C with only one potential, it's not important, but if you wish to use several potentials, it becomes a problem... Is there a way to do this kind of simulation with the current version ? If not, what should be modified to make it work ? I would like to point that having two different ways for doing the same thing is confusing for the user. I would suggest to use only the plasma properties to set the initial potential of each group rather than having a global parameter.