CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Fractional steps method (http://www.cfd-online.com/Forums/main/98278-fractional-steps-method.html)

challenger149 March 7, 2012 04:35

Fractional steps method
 
Can anybody give information about implementing chorin's fractional steps method? I have problems specially about solving poisson equation

FMDenaro March 8, 2012 13:44

Quote:

Originally Posted by challenger149 (Post 348083)
Can anybody give information about implementing chorin's fractional steps method? I have problems specially about solving poisson equation

once computed the intermediate non-solenoidal field you use the Hodge decomposition

V* = Vn+1 + Grad (phi)

Taking divergence of both sides you get

Div (Grad (phi)) = Div V*

with the BCs:

d phi/dn = n . (V* - Vn+1)

The latter condition says that you hav to modify Div V* on the boundary using the normal component of Vn+1 and use homogeneous Neumann BC d phi/dn=0.

This way the compatibility equation is verified and you get a unique solution apart a constant value.

o_mars_2010 April 12, 2013 01:01

Dear FMDenaro
Could you please explain more simply how to modify Div.V* on the boundary, using the normal components of Vn+1??

FMDenaro April 12, 2013 03:20

Quote:

Originally Posted by o_mars_2010 (Post 420025)
Dear FMDenaro
Could you please explain more simply how to modify Div.V* on the boundary, using the normal components of Vn+1??

The equation is:

Div (Grad (phi)) = Div V* (1)

with the BCs:

d phi/dn = n . (V* - Vn+1) -> n . V* = d phi/dn - n . Vn+1

then substitute in RHS of (1) and simplify the terms

o_mars_2010 April 12, 2013 03:33

But how do we know the value of Vn+1
if we have a free stream inlet velocity in the inlet, So, Vn+1 will be known in the inlet.
what about the other boundaries i.e walls and outlet.
Vn+1 is not known at these boundaries ???
Sorry if my understanding is wrong

FMDenaro April 12, 2013 04:17

on a wall n.Vn+1 must be prescribed ... some difficulties can appear for an outlet where you have to prescribe some condition derived by physical assumption

o_mars_2010 April 12, 2013 04:36

Quote:

Originally Posted by FMDenaro (Post 420069)
on a wall n.Vn+1 must be prescribed ... some difficulties can appear for an outlet where you have to prescribe some condition derived by physical assumption

Thanks a lot for your help. Do you have any research paper explains how to deal with the outlet ???


All times are GMT -4. The time now is 23:53.