Differences in 1D and 2D flow results using Gauss-Seidal/Jacobi

u05harrisb · November 19, 2016, 18:21

Dear All,

Thank you in advance for reading and contributing! I've had a good search around and haven't found a previous thread, so thought I'd start a new one.

I am using Gauss-Seidal and Jacobi iterative methods to calculate a simple square duct flow in one and two dimensions. In theory, in my mind, the 1D case should result in precisely the same flow characteristics as the 2D does down its symmetrical center. In practice I have found that the max velocities found differ quite significantly from one another, and I do not understand why.

I am coding these problems using a Fortran code given by my supervisor. As such, I believe the code to be accurate having checked it very closely. Which leads me to the conclusion that there is a known difference between calculating 1D and 2D flows using these methods?

If any of you can shine some light on this matter it would be terrifically appreciated. Is there indeed a reason as to why this is happening, and if so can you elaborate or point me to material that can explain further?

Many thanks,
Ben

FMDenaro · November 20, 2016, 05:01

To tell the true, I do not understand what you want to compare... 1D and 2D model produce different physics... The 1D duct is a model where the tangential stress can only be "modelled" in the sense of Fanno theory. The 2D model produce a flow rate (and a velocity at centerline) that depends on the effect of the stress.

What kind of physical models are you using? For sure, the problem is not the iterative solver.

November 19, 2016, 18:21	Differences in 1D and 2D flow results using Gauss-Seidal/Jacobi	#1
u05harrisb New Member Benjamin Harris Join Date: Nov 2015 Posts: 2 Rep Power: 0	Dear All, Thank you in advance for reading and contributing! I've had a good search around and haven't found a previous thread, so thought I'd start a new one. I am using Gauss-Seidal and Jacobi iterative methods to calculate a simple square duct flow in one and two dimensions. In theory, in my mind, the 1D case should result in precisely the same flow characteristics as the 2D does down its symmetrical center. In practice I have found that the max velocities found differ quite significantly from one another, and I do not understand why. I am coding these problems using a Fortran code given by my supervisor. As such, I believe the code to be accurate having checked it very closely. Which leads me to the conclusion that there is a known difference between calculating 1D and 2D flows using these methods? If any of you can shine some light on this matter it would be terrifically appreciated. Is there indeed a reason as to why this is happening, and if so can you elaborate or point me to material that can explain further? Many thanks, Ben

November 20, 2016, 05:01		#2
FMDenaro Senior Member Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,896 Rep Power: 73	To tell the true, I do not understand what you want to compare... 1D and 2D model produce different physics... The 1D duct is a model where the tangential stress can only be "modelled" in the sense of Fanno theory. The 2D model produce a flow rate (and a velocity at centerline) that depends on the effect of the stress. What kind of physical models are you using? For sure, the problem is not the iterative solver.