Greetings!

Rochak Badyal this is a very common method used to create the fluid domain while analysing an airfoil. This is called C-mesh or C-domain. The reason behind this kind of domain is to distribute the cell (mesh) density along the critical areas. The frontal and rear portion of the mesh is not that much critical compare two to the floor disturbance happening in the upper and lower part of the aerofoil.

I have attached few snaps, the first one shows the pressure contour over aerofoil while second and third one shows the velocity contour at 0 degree and 10 degree angle of attack. You can see how the velocity and pressure distortion is happening in the upper, lower and rearer part of the domain.

This is the reason why we create C-mesh, to capture the flow variation without making dense meshing on complete domain. This will also reduce the computation time.

Checkout my paper, Article Comparative Investigation of Aerodynamic Effects on Geometri...

Hope it may help you.

Regards

Moaz Hussain!

Thanks Moaz Hussain .I would like to further ask what if I used smaller rectangular domain before the leading edge of the aerofoil, as I can see before the leading edge there is not much difference in pressure or velocity but in the C- mesh there are a lot of cells before the leading edge(more thickness in the middle) which will unnecessarily increase the computation cost.

Regards

Rochak

Dear Rochak Badyal , yes you can do it and it's all dependent on the need, like how precise flow analysis you want to carry out.

Now, for a moment consider in your posted snap that you have to know about flow variation at very far point. As you can see the domain thickness is about 10 to 15 times of the chord length of your airfoil. So, now you are making rectangular fluid domain with same thickness and also you need to consider the stagnation pressure at LE. If the meshing is uniform you will end up having a veru6big mesh data as compared to C-mesh because it uses smaller nodes in critical area and larger one in less critical.

You can create a small rectangular one but as I said it's all dependent on requirements.

Dear Rochak Badyal

There is no such hard and fast rule that you have to follow for grid generation. All that you need to see is whether your grid setup is numerically adequate or not and if the simulations are time-taking or not. For an airfoil, it is important to generate a grid looking at its curvature. The semi-circular followed by a rectangular domain is easier to generate. However, had this been a confined channel, you would have been working with a different grid altogether. If computational time is something that is worrying you then search for appropriate numerical methodologies (with focus on time-related constraints) instead of relying solely on grid setup. For detailed insights, go through the articles from Roache and co-workers from University in Albuquerque. However, the explanations from Moaz Hussain are also correct.

Dear Rochak Badyal

The main reason for selecting a C-type grid is to make the curvature of the grid match the leading edge of the aerofoil. This ensures better flow separation. Remember, changing angle of attack of the aerofoil will need a fresh mesh.

Apart from this, as per your another concern, you can certainly use a rectangular domain for the same problem. You have already got a couple of good answers from Moaz Hussain and Aniruddha Sanyal. It is true that there is no hard and fast rule for selecting a computational domain, which is well explained by Aniruddha Sanyal . Several researchers are using a rectangular domain with unstructured grid.

However, you can also use a multi-block structured mesh, a close-up of which is shown towards the end of the paper:

Mittal, S., & Tezduyar, T. E. (1992). A finite element study of incompressible flows past oscillating cylinders and aerofoils. International Journal for Numerical Methods in Fluids, 15(9), 1073–1118.doi:10.1002/fld.1650150911.

This will resolve 2 important issues: (1) the issue related to curvature of mesh as discussed earlier, and (2) the number of nodes and elements will also decrease.

Thanks Moaz Hussain for explaining it very clearly !!

Thanks Aniruddha Sanyal for explaining . Can you please share the link to the articles from University of Albuquerque

Regards

Dear Rochak Badyal,

Try to get hold of this book by P.J. Roache,

" Verification and Validation in Computational Science and Engineering" (1998). This is an old book and is very highly cited in the literature.