Error and Divergence

ouce_2930 · April 9, 2013, 06:28

We are doing a simulation of a 2-phase nozzle(air-glassbeads) in fluent..
The error we got is floating point exception , Temperature limited to 1 in 997 cells on one zone ,one domain ....
Reversed flow in 12 faces on pressure outlet 7...etc etc
We entered the value of outlet pressure as 101325 pascal and inlet is 395900 pascal...Flow selected is laminar
Also tell us whats the significance of solution controls and what are the values that can be given and wats the significance of relaxation factors in this problem we used default values..
As of now we are gettin divergent solution

vasava · April 10, 2013, 06:55

The floating point error is due to divergence. The warning of reversed flow is common as you are specifying pressure and not the direction of velocity on the boundary. By the way How big is your domain? You have a pressure drop of almost 4 atm. What exactly are you simulating?

ouce_2930 · April 12, 2013, 04:49

We are doing a simulation of a CD nozzle in which we use air and glassbeads are the phases..Can you elaborate what exactly does domain mean ??
And we are facing one more problem where when we look at the plots of velocity the air velocity is showing some plot where as glass beads velocity is remaining constant in that plot.. That means it is not taking glass beads into the account
What may be the problem and what changes should be made so that we get plot for glass beads velocity even??

RodriguezFatz · April 12, 2013, 05:44

"Domain" means "setup" or "experiment" here.
You should not think about any strange results you get since your solution does not converge.

It sounds quite unusual that you have a pressure drop of 3 atm but your flow is laminar. With a domain of moderate size you will get huge velocities with such a pressure drop.
Please post pictures of your geometry and you mesh and show where your outlets and inlets are!

ouce_2930 · April 12, 2013, 08:24

FFF.jpg

air-inlet.jpg

These are the attachments... We are going with turbulent flow( k-epsilon) where the solution got converged but the problem is the glass beads were not taken into account by the processosor. We gave all the boundary conditions for mixture but its only taking air in the mixture.

RodriguezFatz · April 12, 2013, 08:35

Quote:

Originally Posted by ouce_2930

These are the attachments...

What...!?

ouce_2930 · April 13, 2013, 01:08

sorry v dint check upon d pics... here r d pics of the geometry and mesh
the left edge is the inlet
and the right edge is the outlet

geometry.jpg

mesh.jpg

oj.bulmer · April 13, 2013, 04:47

It is never robust to specify just pressure values at inlet and outlet. It's preferable to use a quantifiable flow (mass flow rate/velocity etc) for a more stable setup.

Why do you need to simulate only converging-diverging portion of nozzle, why not extend it to few pipe diameter before and after the nozzle? This should be helpful in mitigating backflow. This will use fully developed flow before coming into nozzle giving a more accurate represenation.

Are you sure mixture model is setup properly? Have you had a go at the tutorials on this?

OJ

April 9, 2013, 06:28	Error and Divergence	#1
ouce_2930 New Member Roopa Join Date: Apr 2013 Posts: 4 Rep Power: 13	We are doing a simulation of a 2-phase nozzle(air-glassbeads) in fluent.. The error we got is floating point exception , Temperature limited to 1 in 997 cells on one zone ,one domain .... Reversed flow in 12 faces on pressure outlet 7...etc etc We entered the value of outlet pressure as 101325 pascal and inlet is 395900 pascal...Flow selected is laminar Also tell us whats the significance of solution controls and what are the values that can be given and wats the significance of relaxation factors in this problem we used default values.. As of now we are gettin divergent solution

April 12, 2013, 05:44		#4
RodriguezFatz Senior Member Philipp Join Date: Jun 2011 Location: Germany Posts: 1,297 Rep Power: 26	"Domain" means "setup" or "experiment" here. You should not think about any strange results you get since your solution does not converge. It sounds quite unusual that you have a pressure drop of 3 atm but your flow is laminar. With a domain of moderate size you will get huge velocities with such a pressure drop. Please post pictures of your geometry and you mesh and show where your outlets and inlets are! __________________ The skeleton ran out of shampoo in the shower.

April 10, 2013, 06:55		#2
vasava Senior Member Paritosh Vasava Join Date: Oct 2012 Location: Lappeenranta, Finland Posts: 732 Rep Power: 22	The floating point error is due to divergence. The warning of reversed flow is common as you are specifying pressure and not the direction of velocity on the boundary. By the way How big is your domain? You have a pressure drop of almost 4 atm. What exactly are you simulating?

April 12, 2013, 04:49		#3
ouce_2930 New Member Roopa Join Date: Apr 2013 Posts: 4 Rep Power: 13	We are doing a simulation of a CD nozzle in which we use air and glassbeads are the phases..Can you elaborate what exactly does domain mean ?? And we are facing one more problem where when we look at the plots of velocity the air velocity is showing some plot where as glass beads velocity is remaining constant in that plot.. That means it is not taking glass beads into the account What may be the problem and what changes should be made so that we get plot for glass beads velocity even??

April 12, 2013, 08:24		#5
ouce_2930 New Member Roopa Join Date: Apr 2013 Posts: 4 Rep Power: 13	FFF.jpg air-inlet.jpg These are the attachments... We are going with turbulent flow( k-epsilon) where the solution got converged but the problem is the glass beads were not taken into account by the processosor. We gave all the boundary conditions for mixture but its only taking air in the mixture.

April 13, 2013, 01:08		#7
ouce_2930 New Member Roopa Join Date: Apr 2013 Posts: 4 Rep Power: 13	sorry v dint check upon d pics... here r d pics of the geometry and mesh the left edge is the inlet and the right edge is the outlet geometry.jpg mesh.jpg

April 13, 2013, 04:47		#8
oj.bulmer Senior Member OJ Join Date: Apr 2012 Location: United Kindom Posts: 473 Rep Power: 20	It is never robust to specify just pressure values at inlet and outlet. It's preferable to use a quantifiable flow (mass flow rate/velocity etc) for a more stable setup. Why do you need to simulate only converging-diverging portion of nozzle, why not extend it to few pipe diameter before and after the nozzle? This should be helpful in mitigating backflow. This will use fully developed flow before coming into nozzle giving a more accurate represenation. Are you sure mixture model is setup properly? Have you had a go at the tutorials on this? OJ