CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CD-adapco (http://www.cfd-online.com/Forums/cd-adapco/)
-   -   grid size (http://www.cfd-online.com/Forums/cd-adapco/56785-grid-size.html)

jay September 4, 2008 02:10

grid size
 
Hi...

I made a spray model.

The conditions are:

injector hole size : 3mm mass flow rate : 0.209611kg/s (injection velocity : 25m/s) injection fluid : water (@300K) (into the air @300K, 1bar) nozzle : single hole injector model : Huh's model, Reitz model The solution domain size is not fixed. Injection time & run time are not fixed, either.

I just want the spray plume.

I made a number of models to get the spray plume. However the injector hole size is much bigger and injection velocity is slower than tutorial.

So I couldn't define appropriate grid size, time step and # of parcels.

Does anyone have idea? Plz help me...

I always appreciate all.


usker September 4, 2008 04:09

Re: grid size
 
Regarding the grid size and time step try to determine max.courant number.

Courant Number=time step*Velocity magnitude/cell size.

Max.Courant Number must be less than 1, for some cases such as combustion Max.Courant number must be less that 0.5.

jay September 4, 2008 06:30

Re: grid size
 
Thanks usker..

So, how about the injector exit hole size?

You mean.. in the spray analysis, injector hole size doesn't affect to difine the grid size or time step..

Is that right?

jay September 4, 2008 06:37

Re: grid size
 
And...how do I define "number of parcels/injector" ?

usker September 4, 2008 07:57

Re: grid size
 
For spray calculations you can use 2 methods: 1.Lagrangian Multiphase, 2.Eulerian Multiphase.

If you use Lagrangian approach, the purpose of the Lagrangian approach is to remove the necessity for resolving the nozzle. If your nozzle orifice diameter is in microns then you cannot create grid with such a small cell size. Well, may be you even can do it but computation time will take forever. If the cell size around the nozzle is large, then you will not be able to achieve accurate results, the grid resolution should at least be fine enough to resolve the physical scales of nozzle geometry. The largest errors are caused by an inadequate mesh size near the nozzle, where the velocity and species density gradients are strong and can only be resolved by extremely fine meshes.

However, in your case your nozzle orifice seems large enough, so you may use Eulerian approach by creating the grid around the nozzle with cell size smaller than your orifice diameter. Also, Eulerian approach has an advantage of being computationally more efficient in situations where the phases are widely dispersed and when dispersed phase volume fraction is high. I think this is what you want to get from your calculations, right?

In order to determine number of parcels/injector you need to specify Maximum Number of Parcels, then you devide this number by your injection period. The parcels are counted as cells so the larger the number of parcels, the longer it will take to compute the case. Just try to keep the number of parcels/injector equal to 50000000 and determine MaxNumber of Parcels.

Regards


jay September 4, 2008 10:00

Re: grid size
 
Thanks for your kind answer, usker..

your infomation is alway very helpful for me..

Actually I want to use Star-cd built-in breakup model(ex,Reitz..), so Eulerian Multiphase is not good for my case, i think..

I have one more question (very sorry bothering you :) )

In my case, I think, the injection velocity is too slow. Does that make those of problem? How about your opinion?

Thanks...always..


usker September 5, 2008 05:00

Re: grid size
 
you first try to calculate the case...


All times are GMT -4. The time now is 21:14.