# Problem with SIMPLE algorithm

 Register Blogs Members List Search Today's Posts Mark Forums Read

December 27, 2009, 08:24
Problem with SIMPLE algorithm
#1
Member

Join Date: Dec 2009
Posts: 49
Rep Power: 16
I have written finite difference SIMPLE algorithm to solve the Navier Stokes equation in Matlab. My problem is that I don't know how to specify boundary condition for Poisseuille flow. If i specify inlet velocity, the solution tend to blowup. I suspect that somehow I don't specify the pressure boundary correctly.(In the matlab code if I specify the pressure boundary for inlet and outlet but no velocity boundary at the inlet, the solution does not blow up but the problem is that I cannot control the inlet velocity) I managed to solve Couette flow and Driven cavity problem with no problem though..Care to advice me on what I have to do. I attached also the code for the SIMPLE algorithm that I write..
Attached Images
 Fluid Domain.jpg (70.4 KB, 63 views)
Attached Files
 FD_Method_Flow_Poisseuille.zip (2.8 KB, 63 views)

Last edited by katakgoreng; December 27, 2009 at 08:48. Reason: need to give mo info

 December 27, 2009, 10:13 boundary condition #2 Member   jk Join Date: Jun 2009 Posts: 64 Rep Power: 16 You are solving flow in a duct basically. Use velocity inlet and pressure outlet bc. Which grid r u using ? colocated or staggered

December 27, 2009, 10:21
#3
Member

Join Date: Dec 2009
Posts: 49
Rep Power: 16
Quote:
 Originally Posted by jyothishkumar You are solving flow in a duct basically. Use velocity inlet and pressure outlet bc. Which grid r u using ? colocated or staggered
I'm using staggered grid..if i set the inlet velocity to v=1 and pressure outlet p=0, how about inlet pressure, do i need to extrapolate it with inner pressure?

 December 27, 2009, 23:02 #4 Member   jk Join Date: Jun 2009 Posts: 64 Rep Power: 16 yes exactly. p(1).value = p(2).value where p is the pressure and numbers indicate the node location

December 28, 2009, 04:09
#5
Member

Join Date: Dec 2009
Posts: 49
Rep Power: 16
Quote:
 Originally Posted by jyothishkumar yes exactly. p(1).value = p(2).value where p is the pressure and numbers indicate the node location
hmmm..it's that simple? i'll try it..
for anyone who has run my code, do comment on what I have done..thanks in advance..