As is known,turbulence in the atmospheric surface layer is anisotropic.Now I want to use FLUENT 6.0 to simulate the small scale pollutant diffusion ,how can I define the (swirl) diffusion rate at vertical direction and horizontal direction in the diffusion transport equation?I didn't find any description about the definition in the Chapter "Modeling Turbulence "and UDFs.help me!

how big is the computational domain? and how big is ur machine ?

the computational domain is about 100m(Length)*100m(Width)*50m(Height),I don't understand what you mean saying "how big is ur machine".Species transport in the atmospheric boundary layer is decided by Turbulence diffusion .the diffusion rate is not same at three directions,can you find out where we can define the diffusion rate(not molecule diffusion rate) at different directions?

thank you for help!

sorry to bother you again, how is ur mesh size ? i.e., how many cells you have in the computational domain?

pyramid grid is used in my definition,there are Domain Extents:

x-coordinate: min (m) = -4.000000e+01, max (m) = 4.000000e+01

y-coordinate: min (m) = -7.216450e-16, max (m) = 4.000000e+01

z-coordinate: min (m) = -4.000000e+01, max (m) = 7.498304e-34 Volume statistics:

minimum volume (m3): 3.876812e-04

maximum volume (m3): 2.258250e+00

total volume (m3): 1.259514e+05 Face area statistics:

minimum face area (m2): 9.069004e-03

maximum face area (m2): 3.937478e+00 but i think the mesh size has little affect on the this simulation,what do you think?

Hi The 125,000 cells is the starting, I believe if u do successive refinement or CGI, you will end up with a 125 millions cells. So you start your simulations, your next generation will check the results !!

This is a very common problem of most atmospheric modeling. I performed several environmental modelling research including vegetation in a small lake, radiation effect on water, generation of electricity by wave. All needed massive parallel computing.

You may want to use a combination of CFD and correlation, perdioc BC in two dimensions etc..

In case you don't understand what I am talking about, its long way to go babe !!!

tuki72@hotmail.com

hi guy,your e-mail is not true! how can I send some doc to you ?

or else you can login in my e-mail: wind_flow2@hotmail.com password 19801212 to get the doc,thank you !

Hi,

If you are using a gradient-diffusion model for the scalar transport the diffusion rate is the same in all directions. By the way, how large is the source or initial pollution cloud. If it is smaller than the turbulent length scale you probably cannot use Fluent for the calculation of dispersion.

Tom

try again..It was overpopulated by crazy cfd emails.

hi,Tom,thank you for your attention.

the source can be seen as a chimney ,0.5m(Radius) *5m(Height).can I ask a silly question- how much is turbulent length scale in atmospheric boundry layer? the diffusion rate is same in all directions in scalar transport,can we define it as a vetor in fluent?

by the way,can you give me your e-mail?

The eddy viscosity may well be assumed to be the same in all dirs in gradient diffusion models but the resulting diffusive fluxes may be different as it will be likely the vel grads will be non-isotropic (i think!).

Any how, I'd be more concerned with the assumption of steady state ABL BC. Gusting should not be ignored. As for grid; get it fine in the near field, close to stack emmision to get the bending correct. Use empirical rise height eq of Briggs to judge accuracy of down stream rise height. Dont use k-e - its too diffusive.

Good luck!