book on Level Set, focus on fluid flow
 

Hi,

Can someone suggest a good book on Level Set Method with

1. emphasis on Fluid flow 2. has example coding 3. apply to un-structured body-fitted mesh

The goal is to understand the details of LS coding to capture liquid/air interface (with surface tension effect the best).

Thanks!

phsieh2005

Re: book on Level Set, focus on fluid flow
 

i don't think there is any book with desired specification. note that your need is state of the art and don't well-documented until now (my opinion).

I knowe two book in this field one dure to Sethian and other due to Fedkiw and Osher. I read the the second book and so don't advice it for you. It is very brief and is summury of Fedkiw and Osher papers (is not deep and for details you must refer to cited papers, most of them published in JCP).

Also implementation of LS is very easy as it is algebrical, the only main issue is application of high order non-diffusive conservative upwind methods and usually high order time integration methods such as RK. I recommend for you to look at CIP (due to Yabe group) method as alternative (instead LS), it seems that it is more rubost.

Re: book on Level Set, focus on fluid flow
 

Hi, rt,

I never heard of CIP before. Is it straight forward to coding? I found the website of Yabe group. I will read some of the papers there. Thanks!

phsieh2005

Re: book on Level Set, focus on fluid flow
 

as you deal with free surface tracking, the mass conservation is crtical and usually main weakness of present methods. the best cure is geometrical scheme such as PLIC, but they are very difficult (and sometimes inefficient and ineffective) for unstructured grids. The algebraic methods such as LSM are easy to implement but they are suffer from mass conservation problem, among algebraic hyperbolic conservation law solvers, CIP type method seems (acclaimed by their authors) to be superior from both accuracy and conservation veiwpoints (althouth they are rarely used on Un-grid)

>>Is it straight forward to coding? as they are algebraic, i think that implementation must be easy (at least in comparison with LSM).