Laminar flow between parallel plates with SIMPLE algorithm

cfluid · November 18, 2016, 15:21

Hello Folks,
I'm trying to code in FORTRAN incompressible laminar flow between two parallel plates. I put uniform inlet velocity. I use finite volume method. I'm using 1-D convection-diffusion equation and representing it by the power law. I'm linking velocity and pressure fields using the SIMPLE algorithm.
Dimensions of the channel are Length x Height is 2x0.02. Reynolds number is 150.
So my question is, I need to divide this domain to 160x40 control volumes but when I do it, solution blows up. When I divide domain to 40x160 it runs without any problems. So I wonder if anyone has any comments why Length (2m) cannot be divided into 160 control volumes.

cfluid · July 19, 2017, 02:02

It's been a long time I forgot about this thread.
Just in case if anyone will have the similar problem I solved the problem. I realized that in the first few iterations solution was not propagating to the end of the channel when I was dividing length to the 160 meshes. Solution is to add global mass conservation to the code which basically forces mass to be conserved for the first few iterations until solution reaches to the end of the channel.

Global mass conservation added to the code by checking if the ratio of total outlet mass to the total inlet mass is equal to 1 after each iteration. If not, it is forced to be.

Sent from my iPhone using CFD Online Forum mobile app

November 18, 2016, 15:21	Laminar flow between parallel plates with SIMPLE algorithm	#1
cfluid New Member Alper Join Date: Jul 2013 Posts: 8 Rep Power: 13	Hello Folks, I'm trying to code in FORTRAN incompressible laminar flow between two parallel plates. I put uniform inlet velocity. I use finite volume method. I'm using 1-D convection-diffusion equation and representing it by the power law. I'm linking velocity and pressure fields using the SIMPLE algorithm. Dimensions of the channel are Length x Height is 2x0.02. Reynolds number is 150. So my question is, I need to divide this domain to 160x40 control volumes but when I do it, solution blows up. When I divide domain to 40x160 it runs without any problems. So I wonder if anyone has any comments why Length (2m) cannot be divided into 160 control volumes.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Issues on the simulation of high-speed compressible flow within turbomachinery	dowlee	OpenFOAM Running, Solving & CFD	11	August 6, 2021 07:40
Setting boundary conditions for simple pipe flow with flow direction changing in time	Sipher	FLUENT	1	May 4, 2015 21:05
Is it a turbulent flow or laminar flow?	ringtail	Main CFD Forum	9	January 22, 2015 20:52
SIMPLE algorithm for variable density flow	mauz	Main CFD Forum	1	May 15, 2009 16:13
Solve Simple foam for laminar flow	nandiganavishal	OpenFOAM Running, Solving & CFD	4	January 20, 2009 01:56

July 19, 2017, 02:02		#2
cfluid New Member Alper Join Date: Jul 2013 Posts: 8 Rep Power: 13	It's been a long time I forgot about this thread. Just in case if anyone will have the similar problem I solved the problem. I realized that in the first few iterations solution was not propagating to the end of the channel when I was dividing length to the 160 meshes. Solution is to add global mass conservation to the code which basically forces mass to be conserved for the first few iterations until solution reaches to the end of the channel. Global mass conservation added to the code by checking if the ratio of total outlet mass to the total inlet mass is equal to 1 after each iteration. If not, it is forced to be. Sent from my iPhone using CFD Online Forum mobile app