Hi guys, I need a structured grid ( C-type is prefered) for the transonic airfoil RAE 2822. it should be well clustered for applications of turbulence modeles. Can anyone help?? thanks Michel

You're in luck.

MeshPilot (http://www.shore-cfd.com) generates C-type structured grids. The default settings seem to be for RANS calculations, although you can change all the spacings etc if you want.

A geometry file for the RAE2822 is included with the package. If you have any trouble, email me and I can email you a grid, but you'd be better off downloading it yourself since it's free.

Good luck

thanks for your help...I got it but I cant cluster the grid somewhere on the upper surface where I expect to have a shock....that's bad...so I have to regenerate it which is not a straight FWD job...any idea!! thanks

Hmmm, you're right that you can't create that kind of localised clustering ... but do you really want to?

Isn't it better to create a series of successively finer grids, but with the same stretch functions/relative spacings, and attempt to get a grid-independent solution? If you can manage to get a grid-independent solution, then your results are more convincing.

I'm pretty sure you could get a sufficiently-resolved shock that way - it's surely a long time since we used to have trouble running a 2D RANS job on a fine grid.

Best regards

hey john, Thanks for u reply, Now I wanna make it clear, are u sayin that if one cluster the grid to get a better resolution of the shock and the shock boundary-layer interaction then one would be getting a grid dependent solution???? Another point is that clustering in the upper reagion would be more expansive to solve and less resolving in the critical areas than clustering at the shock. what do u think?

Hi Michel. To be clear, no I'm not saying that clustering the grid in that way is enough to get a grid independent solution. And yes, having an unnecessarily high grid density in any area is inefficient (however, it's these days usually a small price to pay for getting a grid-independent solution).

There are plenty people out there who would advocate the use of an adaptive grid. Let's assume you don't have adaptive technology. In that case, I'd advocate not trying to cluster the grid where you think the shock should be, and tring to get a grid-independent solution. It's a good way of discovering whether your grid is fine enough. You'd be trying to get a grid-independent solution anyway, right?

That's just my opinion though ...

[harim]

bonjour j'ai un problème quant affiche le coefficient de pression suivent y varie de extrados de 0 verre -1 au lieux de 0 à 1