Problems with own CFD

Peszkow · June 20, 2016, 06:07

Hello everyone!
I am creating my own solver of unsteady Stokes flow in 2D(pressure at the inlet rises linearly and stop after few seconds). At first I solve $\nabla^{2}p=0$ and then computed pressure is put into Stokes equations. It bases on finite difference method and uses Conjugate Gradient Method to solve the pressure equation. Next the stokes equations are solved using ADI algorithm and CGM. First step of time is okey, because the solution is zero(so it is put during initiazation). Unfortunately in next step when pressure at the inlet rises pressure don't want to converge. It is linear problem and theoritically it should converge(matrix from laplace equation is symmetricall and PDM) in some cases slowly but still it should converge. Unfortunately in my case calculations are diverging. I have checked everything but I wasn't able to find an error. Have anyone got some ideas how to find where is the problem ?

FMDenaro · June 20, 2016, 07:07

I don't understand the Laplace equation for the pressure...

Peszkow · June 20, 2016, 07:26

If take a divergence of Stokes Equation(incompressible), you will get Laplace equation for pressure.

FMDenaro · June 20, 2016, 08:05

be carefull, numerically the diffusive terms in the momentum equation produce a source term into the pressure equation, that is Div (1/Re Div Grad v). The bc for the normal derivative of the pressure must ensure the compatibility

Peszkow · June 20, 2016, 09:50

Problem found, my matrix turn out to be not symmetricall. Thank you for advice.

AHMEDLAOUARI · June 21, 2016, 06:25

hi everyone
I have a question concerning the difference between conservative and aggresive length scale option for simulation

hossein3975 · January 18, 2020, 12:31

Hi Dears

could you please help me about solid particle injection in liquid phase? particles have injected to inlet but why they don't escape from outlet?

June 20, 2016, 06:07	Problems with own CFD	#1
Peszkow New Member Piotr Pierwocha Join Date: Jun 2016 Posts: 6 Rep Power: 9	Hello everyone! I am creating my own solver of unsteady Stokes flow in 2D(pressure at the inlet rises linearly and stop after few seconds). At first I solve $\nabla^{2}p=0$ and then computed pressure is put into Stokes equations. It bases on finite difference method and uses Conjugate Gradient Method to solve the pressure equation. Next the stokes equations are solved using ADI algorithm and CGM. First step of time is okey, because the solution is zero(so it is put during initiazation). Unfortunately in next step when pressure at the inlet rises pressure don't want to converge. It is linear problem and theoritically it should converge(matrix from laplace equation is symmetricall and PDM) in some cases slowly but still it should converge. Unfortunately in my case calculations are diverging. I have checked everything but I wasn't able to find an error. Have anyone got some ideas how to find where is the problem ?

June 21, 2016, 06:25	solver controle	#6
AHMEDLAOUARI New Member AHMED LAOUARI Join Date: May 2016 Posts: 2 Rep Power: 0	hi everyone I have a question concerning the difference between conservative and aggresive length scale option for simulation

January 18, 2020, 12:31	Dpm	#7
hossein3975 New Member hossein Join Date: Mar 2014 Posts: 2 Rep Power: 0	Hi Dears could you please help me about solid particle injection in liquid phase? particles have injected to inlet but why they don't escape from outlet?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
ICEM CFD Meshing Problems: Hexa Dominant	Evan	CFX	6	July 3, 2006 12:37
Technical Problems at CFD Online	Jonas Larsson	Main CFD Forum	0	May 6, 2005 13:35
ASME CFD Symposium, Atlanta, July 2001	Chris R. Kleijn	Main CFD Forum	0	September 13, 2000 04:48
CFD for fans & blower housings	David Carroll	Main CFD Forum	8	August 24, 2000 17:25
ASME CFD Symposium, Atlanta, July 2001	Chris R. Kleijn	Main CFD Forum	0	August 21, 2000 04:49

June 20, 2016, 07:07		#2
FMDenaro Senior Member Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,777 Rep Power: 71	I don't understand the Laplace equation for the pressure...

June 20, 2016, 07:26		#3
Peszkow New Member Piotr Pierwocha Join Date: Jun 2016 Posts: 6 Rep Power: 9	If take a divergence of Stokes Equation(incompressible), you will get Laplace equation for pressure.

June 20, 2016, 08:05		#4
FMDenaro Senior Member Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,777 Rep Power: 71	be carefull, numerically the diffusive terms in the momentum equation produce a source term into the pressure equation, that is Div (1/Re Div Grad v). The bc for the normal derivative of the pressure must ensure the compatibility

June 20, 2016, 09:50		#5
Peszkow New Member Piotr Pierwocha Join Date: Jun 2016 Posts: 6 Rep Power: 9	Problem found, my matrix turn out to be not symmetricall. Thank you for advice.