CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   Reinitialization for level set (

Amir August 30, 2008 00:34

Reinitialization for level set

I want to track the front of gaseous detonation by level set method.

What is the necessity of reinitialization in level set method? why do I use this tecnique?

Is it just for brevity in equations? In other word, is it just for omitting [grad(phi)] in equations? Is it realy necessary?


Ted August 30, 2008 03:48

Re: Reinitialization for level set
due to numerical diffusion during advection of level set field, it will be smeared and so will not be at its actual position. job of reinit is to correct this problem.

since u deal with compressible, u may not diretly figure out this effect (it mainly show itself by mass loss in incompressible ones), but its results should be unphysical.

note that there r some variational based level set method which do not need reinit

Vinesh H Gada August 30, 2008 07:58

Re: Reinitialization for level set

You see, Level Set(LS) function is defined as signed normal distance function (very smooth function) . This is the most important feature of Level set methods used for incompressible/compressible fluid flows.

In course of time, due to front propagation in a particular direction, the LS function field does not remain signed normal distance function field. When this thing happens, the gradient terms calculation becomes inaccurate. Furthermore, this effect also leads to inaccurate calculation of smeared fluid properties (in case of incompressible flows).

The purpose of Reinitialization thus is to 'Repair/RESET' the LS function calculated from advection step to signed normal distance function WITHOUT moving the interface obtained from advection step i.e. phi = 0 contour before and after reinitialization MUST coincide.

Hope this means that in general you need to reinitialize the LS function periodically but not necessarily after every time step.

Amir August 31, 2008 01:23

Re: Reinitialization for level set

thanks a lot for you concern.

Does modeling compressible flow without reinitialization always, show itself by unphysical results?

Is there any specific effect?


Amir August 31, 2008 01:27

Re: Reinitialization for level set

I have several questions about level set metohd.

May I have your e-mail?

My e-mail is



Amir August 31, 2008 01:31

Re: Reinitialization for level set

excuse me!

My e-mail is:


Ted August 31, 2008 02:37

Re: Reinitialization for level set
note that i do not know what happen, i just guess!

simple way is to test, with and without reinit.

if u have implementation problem, you can refer to available open sources in this regard, e.g.,

sorry, i do not more help, in particular by email :(

All times are GMT -4. The time now is 19:04.