Single vs multi-block structured meshes
 

Hello,

I'm learning Fluent by going through the manual and can't figure out exactly what the differences are between multi-block structured meshes and single-block meshes. How do I know when to use what? What are the advantages and disadvantages of both? Thanks for any help.

Check the book on grid generation by Thompson, available online.

ntonkin:

First recognize that we're talking about structured grids where the 3D spatial domain of interest in XYZ coordinates is transformed into a rectangular computational domain of xi-eta-zeta coordinates. In other words, there's a mapping between an arbitrarily shaped spatial region and a parallelepiped (block) in computational space. That mapping may be very difficult or even impossible to achieve for complex shapes. Therefore, the multi-block strategy is used to first decompose the spatial domain of interest into more than one regions, each of which can be more easily mapped into an independent computational block. Just as the blocks are adjacent in physical space, they are also adjacent (connected) in computational space. However, the overall agglomeration of computational blocks may not be rectangular (i.e. they might be L-shaped or even more complex).

Sometimes there are practical reasons for using a multiblock strategy as when you have different materials or fluids that are easily identified by putting them in their own block, boundary condition assignment, or any of several other reasons including something as mundane as the requirements of your flow solver.

Hope this helps.

Hi
I'm student and try to learn grid generation for my course of CFD. but a question today made me confused.I thought that if i want to grid the area around an airfoil I've to make a domain made by a half circle around it and a rectangle in the continue to see the back effects.[I think it would better to know you that I work with gambit].and then make a multiblock domain consist of a half circle with deleted airfoil in it and one or two rectangle at the back so we can see a grid of C type,isn't it?
but today my friend who works with gambit from 1 year ago told me that he can make a single block domain with high quality grid. this made me confused....how?
does he define a special type of points (corner,side,etc)in the domain at the vertices in the domain?
thanks for your guide.