CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   SIMPLE algorithm, P and u at inlet (https://www.cfd-online.com/Forums/main/66630-simple-algorithm-p-u-inlet.html)

mauz July 20, 2009 12:06

SIMPLE algorithm, P and u at inlet
 
Hello everybody,

I'm solving the N-S equations using the SIMPLE algorithm.
It converge fine and quite fast when I set dirichlet boundary conditions for the velocity at inlet and outlet of the channel, but I have problems when I want to set dirichlet B.C. for both pressure and velocity at the channel inlet.
I'm solving the u-momentum equation at the main control volume interfaces, and the pressure correction equation at the central nodes. how should I set the pressure value on the first node, in order to fulfill the inlet pressure condition? I think the problem is there.

thanks in advance

Mauri

majesty February 13, 2014 09:00

Hello Mauri,

I am having the same problem. My code works almost perfect for velocity inlet but I have problems when I try pressure BCs. Were you able to come up with a solution to set pressure inlet & outlet to your boundary?

leflix March 2, 2014 17:06

Quote:

Originally Posted by mauz (Post 223342)

when I set dirichlet boundary conditions for the velocity at inlet and outlet of the channel

How do you know the velocity magnitude at the outlet?
Imposing a neumann BC for velocity at the outlet is a better option.


Quote:

but I have problems when I want to set dirichlet B.C. for both pressure and velocity at the channel inlet.
what you do at inlet is not consistent.
If you want to impose a dirichlet BC for pressure at inlet then use neumann BC for velocity.

samurai_01 March 3, 2014 10:04

Hi!

1. The inflow conditions will be u=u_inlet, v=0, dp/dx = 0. and for outlet, we can use fully developed outflow BC, (by making few assumptions).
Fully outflow BC is du/dx = 0, dv/dx = 0 and drichlette BC on pressure, ie P=0.

2.The boundaries are always given at physical domain. I believe you are familiar with concept of fictitious cell method for imposing BCs(ref. veerstag & malasekra, or milovan Peric's CFD book). if U or V lies on the "PHYSICAL BOUNDARY" of domain, leave it as it is. If the value lies at fictitious cell, then impose P_boundary = (P_fictitious+P_internal)/2 and make P_b = 0.

Hope this helps. In case of doubts, please reply back


All times are GMT -4. The time now is 13:59.