code_saturne User's Forum

Hello everybody,

Recently, we have implemented a modul to generate the electrical double layer (particular physics).
It was planned to start this code in dimensional and adimensional mode. The adimensional mode is obtained by applying Pi-Buckingham theorem.

In the case where there is not the flow - only the particular physic modul is activated :

If we use the code without setting the source term in Navier-stokes equation then the code works in dimensional and adimensional mode. The velocity field equal zero, no iteration during the computation.

When I'm applying the electrical source term (rho*E) in NS equation in dimensional mode, it's ok ! Its impact is quite low to generate a velocity. No iteration on the velocity field.

But, in the adimensionnal mode, a velocity field is calculated during the computation. In this mode, the velocity term is multiplied by 40000 (due to the adimensional mode). We think that to multiply the velocity by this adimensional factor causes a calculation on the velocity (exceeding a threshold value (epzero ?) that defines the zero).

How to avoid that calculation velocity in adimensional mode ?

Thanks

Paul

I am not answering your question, but asking you a question.
How do you run adimensional simulation for Navier-Stokes equation?
The properties of fluid always has some unities when we input it by GUI or user subroutines. It is need to do some mathematical manipulation to find the right Renoylds number for example.

pleblanc86 wrote:Hello everybody,

Recently, we have implemented a modul to generate the electrical double layer (particular physics).
It was planned to start this code in dimensional and adimensional mode. The adimensional mode is obtained by applying Pi-Buckingham theorem.

In the case where there is not the flow - only the particular physic modul is activated :

If we use the code without setting the source term in Navier-stokes equation then the code works in dimensional and adimensional mode. The velocity field equal zero, no iteration during the computation.

When I'm applying the electrical source term (rho*E) in NS equation in dimensional mode, it's ok ! Its impact is quite low to generate a velocity. No iteration on the velocity field.

But, in the adimensionnal mode, a velocity field is calculated during the computation. In this mode, the velocity term is multiplied by 40000 (due to the adimensional mode). We think that to multiply the velocity by this adimensional factor causes a calculation on the velocity (exceeding a threshold value (epzero ?) that defines the zero).

How to avoid that calculation velocity in adimensional mode ?

Thanks

Paul

pleblanc86 wrote: We think that to multiply the velocity by this adimensional factor causes a calculation on the velocity (exceeding a threshold value (epzero ?) that defines the zero).

Paul

Hello,

This is possible. You must check by yourself in the code.
Maybe you could change another factor (geometry size?) to reduce the factor on your source term.
Regards,