always disconvergent

tmac1kobe8 · April 23, 2012, 22:28

hello,everyone. I was simulating the flowfield of supersonic crossflow and normal sonic jet. The configuration was a 2d plate,
By using 1st scheme of flow in FLUENT, I can get the convergent solution. however, when the 2nd SOU scheme was applied, the calculation always unstable.
Please give some advise, thank you very much.

Zigainer · April 24, 2012, 02:07

Have you tried to lower the under-relaxation factors?

tmac1kobe8 · April 24, 2012, 20:54

yes, I have lowered under-relaxation factors from the default value(0.8) to 0.5, it failed still.
the boundary condition:
crossflow: pressure inlet(supersonic) p0=2.4e5 Pa, T0=309K, M=3.5
jet: pressure inlet(sonic): p0=3.8e5 Pa, T0=293K,
upper boundary: farfield pressure: M=3.5, T=86.5K
out: pressure outlet: p=3145 Pa.

Zigainer · April 25, 2012, 02:01

Just decrease the under-relaxation factors more ..... I had problems with under-relaxation of 1/3 or less of default values

tmac1kobe8 · April 25, 2012, 03:09

I tried the value of under-relaxatioin factors of 0.1, it didn't work. I observed the field, when I turned 1st scheme to 2nd scheme, after several iteration, there was high temperatrue region right behind the mach disc of jet plume.
I think it was the source of disconvergence.
But I don't konw how to deal with this problem

Zigainer · April 25, 2012, 03:14

okay .... have you checked the mesh in this region?

tmac1kobe8 · April 25, 2012, 08:43

yes, mesh is very simple, it is H-type, I think the problem should not lie in mesh.
Now I am trying calculating with pressure-based solver instead of density-based solver.

t.krishnamohan · April 29, 2012, 21:53

You should use density based solver for compressible flows. Have you tried decreasing the courants number?

April 23, 2012, 22:28	always disconvergent	#1
tmac1kobe8 New Member Wang Join Date: Dec 2010 Posts: 8 Rep Power: 15	hello,everyone. I was simulating the flowfield of supersonic crossflow and normal sonic jet. The configuration was a 2d plate, By using 1st scheme of flow in FLUENT, I can get the convergent solution. however, when the 2nd SOU scheme was applied, the calculation always unstable. Please give some advise, thank you very much.

April 24, 2012, 02:07		#2
Zigainer Senior Member Join Date: May 2011 Location: Germany Posts: 130 Rep Power: 14	Have you tried to lower the under-relaxation factors?

April 24, 2012, 20:54		#3
tmac1kobe8 New Member Wang Join Date: Dec 2010 Posts: 8 Rep Power: 15	yes, I have lowered under-relaxation factors from the default value(0.8) to 0.5, it failed still. the boundary condition: crossflow: pressure inlet(supersonic) p0=2.4e5 Pa, T0=309K, M=3.5 jet: pressure inlet(sonic): p0=3.8e5 Pa, T0=293K, upper boundary: farfield pressure: M=3.5, T=86.5K out: pressure outlet: p=3145 Pa.

April 25, 2012, 02:01		#4
Zigainer Senior Member Join Date: May 2011 Location: Germany Posts: 130 Rep Power: 14	Just decrease the under-relaxation factors more ..... I had problems with under-relaxation of 1/3 or less of default values

April 25, 2012, 03:09		#5
tmac1kobe8 New Member Wang Join Date: Dec 2010 Posts: 8 Rep Power: 15	I tried the value of under-relaxatioin factors of 0.1, it didn't work. I observed the field, when I turned 1st scheme to 2nd scheme, after several iteration, there was high temperatrue region right behind the mach disc of jet plume. I think it was the source of disconvergence. But I don't konw how to deal with this problem

April 25, 2012, 03:14		#6
Zigainer Senior Member Join Date: May 2011 Location: Germany Posts: 130 Rep Power: 14	okay .... have you checked the mesh in this region?

April 25, 2012, 08:43		#7
tmac1kobe8 New Member Wang Join Date: Dec 2010 Posts: 8 Rep Power: 15	yes, mesh is very simple, it is H-type, I think the problem should not lie in mesh. Now I am trying calculating with pressure-based solver instead of density-based solver.

April 29, 2012, 21:53		#8
t.krishnamohan Member Krishna Join Date: Oct 2009 Posts: 34 Rep Power: 16	You should use density based solver for compressible flows. Have you tried decreasing the courants number?