Hello every one. I try to simulate poiseuille flow in the channel. No thing special but the challenge is to coupled this flow with an electrical potentiel and ionic charge. So I creat additionnal variable "Ions" and "Potentiel" for the potentiel I solve Poisson equation. and it works. But for ionic charge I have to solve transport equation with second terme or source terme= div(Ions*E) With E is the elecrical field And I have overflow in the first 10 iterrations. Any suggestions? thx

I would suggest to use a source coefficient. Did you do that?

Gert-Jan

www.bunova.nl

Hello Yes I take the second term div(ion*E) as a source term

I mean a source coefficient...........

Yes I did it just now It's converge but the result is not so good But it's a progress Thx

Dear mac,

How did you implement the div(Ions*EField) as a source term?

Did you do it in integral form?

Opaque

div(IonE)=d/dx(IonEx)+d/dy(ionEy)+.... I creat 2 additional Variables IonEx, IonEy and I get the component of grad(IonEx) and grad(Ion Ey) by Fortran subroutine Do you have other possibility?

Dear mac,

For conservation purposes and consistency with the finite volume method, the div (ionsEField) should be computed as

Sum over faces of a finite volume of (Ions * Efield dot NormalAreaVector)..

There are no derivatives to compute and the quantity will be conserved over the whole domain..

The sum above can only be done via User Fortran. You should contact you help representative and request information on how to compute this type of term.. There used to be a "MHD beta feature in User Fortran" in the 5.5/5.6/5.7.1 days that contained routines to compute the div (J_b) as source for the electric potential.

In any case, I still do not understand the units of your Ions equation..

Hope this helps,

Opaque

Thx a lot The unity of Ions is Coulomb or A/(m^3 s) I use Ions as volumetric variable and it's related with the potentiel by div(grad(V))=ions/eps0

Can you explain me more how to use your proposition? Best regards

Dear mac,

For conservation purposes and consistency with the finite volume method, the div (ionsEField) should be computed as

Sum over faces of a finite volume of (Ions * Efield dot NormalAreaVector)..

There are no derivatives to compute and the quantity will be conserved over the whole domain..

The sum above can only be done via User Fortran. You should contact you help representative and request information on how to compute this type of term.. There used to be a "MHD beta feature in User Fortran" in the 5.5/5.6/5.7.1 days that contained routines to compute the div (J_b) as source for the electric potential.

In any case, I still do not understand the units of your Ions equation..

Hope this helps,

Opaque

Dear mac,

Please contact the CFX help desk about the implementation of additional divergence terms into existing equations..

You will be using User Fortran, in particular USER_GETVAR and USER_GET_MESHDATA to access the electric field, the ions, the normal area vector to control volume faces, control volume sectors.

Then, within an element you will compute the contribution of the sum to each control volume sector.

Opaque.

Dear Opaque I am simulating the effect of electric field in dukt flow. For your quick perusal, the governing equations for electrostatis is written below: 1. E= -Grad (V), and 2. Div(grad V)= (Total space charge density) / (Dielectric const of air) 3.Div(Total space charge density*Velocity)=-Div(Total space charge density*E) where V is the applied electric potential, and E is the electric field The firste equation I resolve it as Algebrical equation method The second by poisson equation with 1 for kinematic diffusivity and the second term is a source The tird equation I resolve it as Transport equation with the second term is a source This last equation causes me problem it's non linear and i get error mesage: "over flow" after few ittrartions I try yesterday to introduce coefficient source trem, it's converge very difficultlly but the result is not correct. I hope you can see now what I like to do I'm waiting for advices

[oghazian]

hi,

i want to investigate the effect of electric field on two phase flow, can it be done by MHD module ( I have no magnetic field) ?

same as you i just want to solve the electrostatic equations.
is it possible in MHD module???