immersed boundary method with unstructured mesh

x.zhang

I have a question about immersed boundary method. Hope someone could give me a hint.

I am trying to use immersed boundary method in an unstructured N-S solver. I don't know how to choose the distribution function. This continuous function is designed to mimic the delta Dirac function.

For uniformed Cartisian mesh, it is just the multipling of two 1D distribution functions ( along x and y direction) which are easy to creat.

For unstructured mesh, I think this function is distance based and after the distribution,the element area weighted summation must equal to the original varible value on the front.

Thanks a lot.