CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   LES: mean velocity and turbulent kinetic energy (http://www.cfd-online.com/Forums/fluent/43048-les-mean-velocity-turbulent-kinetic-energy.html)

 MET December 6, 2006 16:15

LES: mean velocity and turbulent kinetic energy

Hi guys, I'm simulating an unsteady turbulent flow, using LES with a Smagorinsky-Lilly subgrid model (the Dynamic one). I have a udf to get velocity and velocity gradients in some specific cell of the computational domain.

I have some questions:

1)Since the LES applies a spacial filter to resolve the governig equation of the flow (u= [u]+u'), the velocity values that I get from the flow field are the istantaneous value (u) or the mean velocity value ([u])?

If they are the istantaneous one, how I can get the mean value?

2)How I can find the value of the subgrid turbulent kinetic energy? I need that to calculate the Reynold stress.

Thanks Met

 Emmanuel December 6, 2006 19:13

Re: LES: mean velocity and turbulent kinetic energ

Hi,

I have a similar problem. I have a simple unsteady flow in a nozzle ejector. I am using a sinusoidal input for my total pressure and a constant static pressure at the input. At the output I have the same constant pressure as in the input.

My question is how do I obtain the mean average velocity of both the input and the output, since the Report-Surface Integral, only yields the instantaneous velocity.

One method I am considering is to run a cycle and plot these velocities via Solve-Monitors-Surface, and export this plot and integrate to find the average. Note the exit velocity which I am more interested in is not a pure sinusoidal output.

 suvash December 7, 2006 01:42

Re: LES: mean velocity and turbulent kinetic energ

I can give you some idea. Go to SURFACE->LINE?RAKE. Then choose the end points to draw a line along where you want to get the average velocity. Then go to SOLVE->MONITOR->SURFACE and select your line and choose x/y velocity and vertex average/area weighted average value.

Suvash

 Emmanuel December 7, 2006 05:39

Re: LES: mean velocity and turbulent kinetic energ

Thank you, I want the velocity at the exit. So the problem is when I use SOLVE->MONITOR->SURFACE and select the veocity at the exit, I will obtain a plot. How will I best find the average of that plot? Like I suggested before, I think exporting the plot to EXCEL or MATHLAB, and then using these programs to integrate and average might do it. But, is there another way to find the average of this plot without exporting the graph?

 suvash December 7, 2006 05:47

Re: LES: mean velocity and turbulent kinetic energ

Going to SOLVE->MONITOR->SURFACE select the exit surface then select the velocity/x,y velocity from "report of" and then from "report type" select vertex average. then you will get the average velocity of that surface at every time step.

This is what I know!!!

 Emmanuel December 7, 2006 05:59

Re: LES: mean velocity and turbulent kinetic energ

hey, thanks, let me try that. I have another concern. When I check for conservation of mass using REPORT->FLUXES->MASS FLOW RATE and selecting all inlets and outlet for this my unsteady case, I notice that I do not have conservation. but, it was conserved for the steady case. Is this normal for periodic flows?

 David December 7, 2006 07:20

Re: LES: mean velocity and turbulent kinetic energ

If you initialize the data sampling before iterating then you can plot the unsteady statistics such as mean velocity and rms velocity on a particular area of the model using the line/rake function.

 MET December 7, 2006 11:27

Re: LES: mean velocity and turbulent kinetic energ

Probably my problem is more general than Emmanuel's. I need to now the gradients of the mean velocity in all the cells of the domain.

Anyway, there is one thing that I don't understand. In the LES, the Reynolds decomposition is based on a spacial filter. Using data sampling I obtain an avarage in time and not in space. What I really would like to get is the mean velocity used to resolve the governing equation. For Large Eddy Simulation this value is avaraged in space. (Fluent use as space filter the volume of the computational cell).