code_saturne User's Forum

Hello, everyone,

I am going to build a two-temperature model of an electric arc plasma torch. It implies separate temperatures (or enthalpies) for heavy particles (molecules, atoms and ions) and electrons. The electric arc is non-transferred and can burn in re-strike mode.
I need two-temperature model because high temperature electrons increase electric conductivity behind the anode arc attachment and make the re-strike easier.

Does anyone know if it is possible in Code_Saturne to have two energy conservation equations separately for heavy particles and electrons?
And if so, what should I add to the model and how can I assign properties of the electron gas?

Thank you.
Best regards,
Rodion

Hello,

2 scalars behaving as temperature or enthalpy are certainly possible in some cases at least. The code assumes a "main" thermal scalar, which may be accessed directly (iscalt in Fortran, thermal_model in C), but if you are careful about this (i.e. ensuring the second thermal scal is independent), you should be OK.

Things get more trick if you need 2 separate total energies (including kinetic aspects) such as for the compressible model, fo I guess this is not your case.

Also, if physical properties depend on those 2 scalars, they'll also depend on the concentration of each type of electron, so you'll probably need scalars representing the concentration of each electron gas (in which case I'm not sure 2 enthalpies are needed; 1 common enthalpy accounting for the mix is simpler, if at least one of the two types of particle have constant or near-constant temperature (so enthalpy depends only on density). I'm not sure I understand whether "non-transferred" aspect means "ignore heat exchange" or not, so this suggestion might be off...

Also, using a Lagrangian model for one or both of the electron jets might be an alternative. This would require seeding the flow with enough electrons so that you can compute an electron density for a given mesh cell to compute Eulerian quantities, and defining associated properties (mainly a "relaxation time" for equations relative to each type of particle I believe) Though you would probably need to code the cs_user_lagr_ef user function to handle the effects of the electric field on particles (as body forces).

So there are a variety of options, but I am not familiar enough with the electric arcs model to know which suggestion is the best (or worse)...

Regards,

Yvan

Thank you for your answer, Yvan!

As far as I remember, the main thermal variable for in electric arc simulations in Code_Saturne is enthalpy. So I need two enthalpies for both heavy species and electrons. (or enthalpy for heavy species and temperature for electrons — it is usually referred to as "2T model") Anyway, both change a lot, so I cannot replace them with just one common enthalpy.

I think currently I cannot consider compressible flow in Code_Saturne because compressible module does not work with electric module at the same time.

Along with separate thermal variables I need separate concentrations for heavy species and electrons. The physical properties are defined by all this parameters. Concentrations of heavy species and electrons can be different due to non-constant degree of ionization. Degree of ionization depends on temperature. Due to electric neutrality, number density of electrones corresponds to number density of ions with respect to ion charge. But neutral atoms number density can be higher or lower than number density of ions.

So the question is how can I have two thermal variables and concentrations for both heavy species and electrones and how can I introduce the properties of electron gas to the model separately from heavy species properties?
I need Code_Saturne to solve both thermal variables (gradient calculation, balance calculation, etc.).

Non-transferred arc is the arc like in the attached picture.

Best regards,
Rodion

Hello,

You cannot have 2 variables marked as thermal variables, but can can definitely have 2 scalars behaving as an enthalpy and define user properties based on whatever values you have in these variables.

Most scalars behave like enthalpy rather than temperature by default (I.e. no multiply/divide by Cp) so that part should be straightforward (though you may need to compute the enthalpies matching the electron concentrations needed for boundary conditions).

Best regards,

Yvan

Thank you for your answer, Yvan!

Okay, so how can I add another scalar which will be solved by CS, its conductivity and the concentrations of electrons and heavy particles?

Also I will need to recompute all the properties with respect to the concentrations. Does CS do it by default I should do it by myself?

Best regards,
Rodion

Hello,

You can simply add 2 user-defined scalars (which is even possible using the graphical interface).

For the associated properties based on concentration fractions, this will not be automatic, as by default the scalars are "passive" (i.e. do not influence the other fields, can be seen as tracers): you will need to program user-defined property values (either using the interpreter in the graphical interface or user subroutines; the interpreter solution may be nice for testing, but is slower).

Regards,

Yvan

Thank you for your quick response, Yvan.

Okay, I see. But how can I make Code_Saturne solve the transport equation my user-defined scalar — electron temperature?

Best regards,
Rodion

Hello,

By default, any user-defined scalar is considered to be both diffused and transported (you can disable either in user subroutines using the iconv or idiff keywords/members of the "var_scal_opt" option settings).

So that part is automatic. The tricky part might be defining the correct variable user properties such as diffusivity.

Regards,

Yvan