[Mourousias]

Hello everyone,
I want to make a simulation of helium cold plasma jet in ambient air. For the modelling of turbulence I use the LES model with dynamic one equation subgrid model. I also use a periodic force on the flow for the plasma simulation. I want to make a classical grid convergence study using the Richardson extrapolation based methods. Unfortunately I know that the LES model makes extremely difficult the classical convergence study. In addition the periodic nature of force I apply on the flow, makes also difficult to evaluate for the grid convergence the mean values of the fields, a technique that is proposed when dealing with turbulence.

So my question is, if it's wrong for the validity of my simulation if I test the grid convergence in the simple case of laminar helium jet in ambient air (without the turbulence model and without the periodic force). I wil also validate my plasma simulation later with experiments.

Thank you in advance for your replies

[FMDenaro]

Originally Posted by Mourousias

Hello everyone,
I want to make a simulation of helium cold plasma jet in ambient air. For the modelling of turbulence I use the LES model with dynamic one equation subgrid model. I also use a periodic force on the flow for the plasma simulation. I want to make a classical grid convergence study using the Richardson extrapolation based methods. Unfortunately I know that the LES model makes extremely difficult the classical convergence study. In addition the periodic nature of force I apply on the flow, makes also difficult to evaluate for the grid convergence the mean values of the fields, a technique that is proposed when dealing with turbulence.

So my question is, if it's wrong for the validity of my simulation if I test the grid convergence in the simple case of laminar helium jet in ambient air (without the turbulence model and without the periodic force). I wil also validate my plasma simulation later with experiments.

Thank you in advance for your replies

If you are using an implicit filtering (that is the filter is let to the discretization) and a dynamic SGS modelling, you should be aware that the LES solutions over successively refined grids will tend towards the DNS solution, you cannot reach a converged LES solution.

Testing a laminar solution is useful if you have some exact solution to compare with. That helps in validating the code but if you want to validate your LES code you should use some well assessed test-case for LES:

http://torroja.dmt.upm.es/turbdata/a...ARD-AR-345.pdf

[Mourousias]

Thank you very much for your reply.

In the same manner I also have to validate the final code which includes the plasma. Also as you said a converged solution using LES implicit filtering is not possible.I am a little bit confused but aren't validation and grid convergence two separate things? So far i have understand that the first is a test for the whole solution ( physical modelling and grid) given that the grid convergence is achieved ( meaning that the discretization error is close to truncation error). I also think that the necessity of grid convergence is increased especially if adjustable parameters are involved but i am not sure. So my critical question is if it is meaningless the grid convergence study in a simualtion much simpler than its final form ( stripped out of the LES and the periodic force)