CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Post Processing (http://www.cfd-online.com/Forums/fluent/29898-post-processing.html)

Koorosh MOHAMMADI June 13, 2002 04:36

Post Processing
 
Dear All,

I'm trying to present some 3D graphs of my simulation. Those graphs are about the Absolute Pressure, Total Temperature and Velocities (x,y,z and magnitude) in some plane (SURFACE PELANE) at the middel of my geometry. To control the number of "Level Curves" I prefered to insert "Surface Rakes". Now when I plot for example the values of Z velocity on those rakes, the graphs will be different in respect to whether I switch on the node values or not. Since the node values will present the interpolated data in the faces, the graphs are more smooth and the trends are more clear. But when the node values are swithced off, even though the trend of graphs is the same, but they are not as smooth as previouse one (I have near 1,500,000 3D elements in geometry and they are enough suitable and the simulation seems to be converged.)

Now what you suggest? The graphs should be presented by use of NODE VALUES or with out them. Is it important that if I use NO NODES, I smooth the data by use of another programm?

Best regards

Koorosh

Jonas Larsson June 13, 2002 06:54

Re: Post Processing
 
The uninterpolated cell values are of course the most correct values, these are the cell-averages computed by the solver. This gives more information for you as a CFD engineer - you can easily see if you have local stability/convergence problems in some cells etc.

By interpolating the cell values to the nodes you get a smoother plot with nicer contours. What you do when you interpolate to the nodes is essentially to perform a smoothing of your calculated CFD results. This hides local grid-related problems and if you want to show plots to non CFD people this might be better - things simply look smoother. It looks better than you actual results are though.

I use node-values in publications and presentation material to non-CFD people, but I always use cell values when I look at the results myself or want to discuss them with fellow CFD colleagues.


All times are GMT -4. The time now is 10:48.