CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Cell Reynolds Number (http://www.cfd-online.com/Forums/main/134-cell-reynolds-number.html)

laliong August 26, 1998 09:41

Cell Reynolds Number
 
Hi, I'm a beginner of CFD. In some paper, I read this words "cell Reynolds number". But I'm not sure what exactly this words mean. Any explanation will be very appreciated. Thanks.

Fabien Coppens August 26, 1998 10:06

Re: Cell Reynolds Number
 
Hi. It probably refers to the local Reynolds number based on the velocity in the calculation cell.

andy August 26, 1998 10:11

Re: Cell Reynolds Number
 
...and the length scale of the cell.

laliong August 26, 1998 10:28

Re: Cell Reynolds Number
 
Thanks for your responses. But what is the constraint of cell Reynolds number for flow with high Reynolds number ?

Oliver August 26, 1998 11:40

Re: Cell Reynolds Number
 
The cell Reynolds number can be taught as The Peclet Number after Roache. The Reynolds Number is the ratio of Inertial to Viscous forces. While the Peclet Number is the Ratio of Convective to Diffusive Fluxes across the Cell. Therefore at low Peclet number, diffusion of a property is of higher importance than convection of that property across the cell.

In order for a discretization scheme to behave appropriately for different Peclet numbers it must be transportiveness, which is one of the fundamental properties of a discretization scheme. All this says is, that for low Peclet numbers the discretization schemes must behave as thought diffusion is the dominant characteristic, while for high Peclet number the discretization scheme must behave as thought convection of the property is the dominant characteristic.

Hope this is helpful

Fabien Coppens August 26, 1998 11:46

Re: Cell Reynolds Number
 
There are various ways that local Reynolds numbers can be useful, including as an idea of the local stability characteristics of a discretization scheme. Some schemes have critical Re above which you get oscillations, etc... Another useful local variable is the CFL number defined as V*dt/dx where dt and dx are (local) time step and grid scale respectively. Every scheme has a theoretical maximum CFL number.

John C. Chien August 26, 1998 12:51

Re: Cell Reynolds Number
 
Assuming that you are writing the finite-difference equations of the Navier-Stokes equations in Cartesian coordinates, ( use central difference or one-sided upwinded difference scheme), you can group the coefficients of the final finite-difference equations in such way that some will appear as Re,x , Re,y , Re,z . where Re,x is defined as (rho * U * delt-x)/mu, Re,y is defined as (rho * V * delta-y)/mu, and Re,z is defined as (rho * W * delta-z)/mu.These are teh local cell's Reynolds number because U,V,an W are local nodal point velocity components.(U=U(i,j,k)), V=V(i,j,k), and W=w(i,j,k) ). And the delta-x, delta-y, and delta-z are the local cell sizes in x-, y-, and z- directions. ( delta-x= X(i+1,j,k)- X(i-1,j,k), delta-y = Y(i,j+1,k)-Y(i,j-1,k), delta-z = Z(i,j,k+1) - Z(i,j,k-1) ). There is nothing magic about the definition of these cell Reynolds number at this stage because you will get these in the process of deriving the finite-difference equations. The problem is when you try to solve this set of finite-difference equations, you have to use some kind of finite size mesh. Based on the mesh size you use, you can change the values of the local cell Reynolds numbers, because they are directly related to the delta-x, delta-y, and delta-z. When you increase the cell size ( using coarse mesh), you automatically increase the cell Reynolds numbers ( the coefficients of your finite-difference equations ). In your equations, you have terms coming from the left-hand side of the momentum equation ( convection term) , you also have terms coming from the right -hand side of the momentum equation. As a result, in your final finite-difference equations, some terms will have positive coefficient, and some terms will have negative coefficient. Because not all terms have local cell Reynolds number attached to them, when you use a coarse mesh, some terms will become larger because the cell Reynolds number will increase with cell size. Sometimes, this will change the nature of the finite-difference equation. ( negative terms can be enlarged to an extent that they become larger than the rest of the terms simply because you are using coarse mesh. ). Numerically, when you try to solve the equation, you can see the solution wiggles because you have just changed the nature of the equations. In early days, when people use central difference for both the first-oder convection term and the second-order diffusion terms, there were positive and negatice coefficients in the final equation. As the cell Reynolds number becomes greater then two, the nature of the finite difference equation was changed because of the cell size used. The result was instability in the solution ( sometimes the solution diverged). By using the "upwind" one-sided difference , you can obtain the finite difference equation with all the coefficients remain the same sign regardless of the cell size you select. At least using this method, you can avoid the stability problem related to the cell size. In other words, when you use certain finite-difference methods, you can easily change the nature of the equation. ( you don't have this problem with the original partial differential equations because the cell size is approaching zero.) This is a big step in the CFD development. Once you can handle the stability problem, you would like to get more accurate solutions by using more accurate schemes ( first order upwind scheme tends to smear the solution profiles. more diffusive. ) The best way to get the feeling is to use the 1-D model equation with convection and diffusion terms, and write a simple program with various difference schemes to check the sign of the coefficients and the final solutions. ( you don't get this part by just running a commercial code.)

Rasputin September 8, 1998 07:01

Re: Cell Reynolds Number
 
John,

I thought the cell Peclet number was used, not the cell Reynolds number. What's the difference? Is there a difference? I suppose one looks at fluxes the other considers absolutes.

Rasputin September 8, 1998 07:47

Re: Cell Reynolds Number
 
Oops, of course Pe = RePr. Doh...

Suchit March 19, 2015 12:39

About Matleb Programe for reynold's equation solution by finite diffrence method
 
you have any matleb programme in which Reynold's equation is solved by finite difference method???????
i am doing work on this bus not getting any reference work for complete convergent and divergent shape for parabolic shape....
My mail id : shahsuchit007@gmail.com

Suchit March 19, 2015 12:40

you have any matleb programme in which Reynold's equation is solved by finite difference method???????
i am doing work on this bus not getting any reference work for complete convergent and divergent shape for parabolic shape....
My mail id : shahsuchit007@gmail.com

Suchit March 19, 2015 12:42

About Matleb Programe for reynold's equation solution by finite diffrence method
 
you have any matleb programme in which Reynold's equation is solved by finite difference method???????
i am doing work on this bus not getting any reference work for complete convergent and divergent shape for parabolic shape....
My mail id : shahsuchit007@gmail.com


All times are GMT -4. The time now is 06:45.