CFX Spray Breakup Setup

Spray_Ansys · May 25, 2018, 16:22

Quote:

Originally Posted by ghorrocks

If you fire the drops in a straight line, they will stay in a straight line. You need to fire them at the spray angle of your injector to get a spread.

Do you have an idea how to set up a good expression for this? I thought about something like a half of a circle and set the velocity as normal (perpendicular) on the function. But it seems to be quite difficult, doesn't it?

ghorrocks · May 25, 2018, 20:51

If you are using lagrangian particle tracking CFX has several built in models to generate various spray shapes, such as conical. As this model is used to model IC engine fuel injectors it covers typical fuel injector spray shapes quite well. So it is quite easy to do.

Spray_Ansys · May 26, 2018, 04:34

Quote:

Originally Posted by ghorrocks

If you are using lagrangian particle tracking CFX has several built in models to generate various spray shapes, such as conical. As this model is used to model IC engine fuel injectors it covers typical fuel injector spray shapes quite well. So it is quite easy to do.

Thanks, but I want to try the Eulerian model and investigate how it will look like to spray with an angle. Do you now a good expression for a planar inlet surface?

ghorrocks · May 26, 2018, 04:48

If the spray centreline is on y=0 and the spray extends from -1[mm] to 1[mm], then if you set the:
U Velocity to the spray velocity
V velocity = y/1[mm]*10[m/s]
W Velocity = 0

Then the spray will do a 2 dimensional fan from -10[m/s] to +10[m/s].

But I don't understand why you would want to do this as a Eularian model cannot model spray breakup which appears to be the point of this analysis. Your only option is a Lagrangian model and that is set up completely differently.

Spray_Ansys · May 26, 2018, 05:21

Thank you, this looks fine!
Spray breakup isn't point of the analysis anymore. I try something on a different way.

Spray_Ansys · June 8, 2018, 07:13

Further question:
I have now a suitable result of my spray flow. At the beginning (nozzle outlet) the Volume Fraction is commonly with value 1.0 and to the end of my simulation region it is going to spread and the Volume Fraction is decreasing.

I would like to know now, if there is a correlation between Volume Fraction and Drop Distribution. Does somebody know a good way to figure this out? I hardly found any papers focusing on this issue...

Thank you in advance for your answers!

ghorrocks · June 9, 2018, 07:55

What do you mean "drop distribution"? If you mean drop size distribution - then if you are doing a normal Eularian simulation like you mentioned before, then the drop size is defined in the model setup, so the drops will be the size you defined them to be. If you mean drop velocity distribution then just look at the appropriate variable in the post processor.

Spray_Ansys · June 9, 2018, 08:07

Yes, I mean drop size distribution.
What if I setup an multiphase Eulerian model but with two continuous fluids? How can I sort it out then? Did someone find some literature about that before?

ghorrocks · June 9, 2018, 08:37

Please have a careful read of the CFX theory documentation on multiphase modelling. You need to know what the models mean, what they do and what they assume before you can make sensible choices about which model to use. So you really should be answering that question yourself if you are doing this sort of work.

Eularian models are not generally used to have varying drop size. Lagrangian models are more commonly used for this. There are some Eularian models which cover variable drop size (eg MUSIG) - you might want to investigate these models. I am no expert on MUSIG, but I suspect MUSIG is not going to be appropriate. Then you are forced back to Lagrangian models again

May 25, 2018, 20:51		#22
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	If you are using lagrangian particle tracking CFX has several built in models to generate various spray shapes, such as conical. As this model is used to model IC engine fuel injectors it covers typical fuel injector spray shapes quite well. So it is quite easy to do. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

May 26, 2018, 04:48		#24
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	If the spray centreline is on y=0 and the spray extends from -1[mm] to 1[mm], then if you set the: U Velocity to the spray velocity V velocity = y/1[mm]*10[m/s] W Velocity = 0 Then the spray will do a 2 dimensional fan from -10[m/s] to +10[m/s]. But I don't understand why you would want to do this as a Eularian model cannot model spray breakup which appears to be the point of this analysis. Your only option is a Lagrangian model and that is set up completely differently. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

June 9, 2018, 07:55		#27
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	What do you mean "drop distribution"? If you mean drop size distribution - then if you are doing a normal Eularian simulation like you mentioned before, then the drop size is defined in the model setup, so the drops will be the size you defined them to be. If you mean drop velocity distribution then just look at the appropriate variable in the post processor. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

June 9, 2018, 08:37		#29
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	Please have a careful read of the CFX theory documentation on multiphase modelling. You need to know what the models mean, what they do and what they assume before you can make sensible choices about which model to use. So you really should be answering that question yourself if you are doing this sort of work. Eularian models are not generally used to have varying drop size. Lagrangian models are more commonly used for this. There are some Eularian models which cover variable drop size (eg MUSIG) - you might want to investigate these models. I am no expert on MUSIG, but I suspect MUSIG is not going to be appropriate. Then you are forced back to Lagrangian models again __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Predict % evaporation of a spray ffrom excel based on flow analysis results from CFX	kar.coep	ANSYS	0	December 16, 2015 03:11
The CFX Setup case file no longer exists. Continuing in an initial state.	ced30	CFX	1	November 13, 2015 13:00
Script CFX Setup for dynamic domain generation	kar.coep	CFX	1	May 18, 2015 03:29
CFX Distributed Parallel Setup	Behzad	CFX	6	November 25, 2010 21:54
CFX 4.4 installation problem	Pandu Sattvika	CFX	1	December 1, 2001 05:07

May 26, 2018, 05:21		#25
Spray_Ansys Member Join Date: Mar 2018 Posts: 30 Rep Power: 8	Thank you, this looks fine! Spray breakup isn't point of the analysis anymore. I try something on a different way.

June 8, 2018, 07:13		#26
Spray_Ansys Member Join Date: Mar 2018 Posts: 30 Rep Power: 8	Further question: I have now a suitable result of my spray flow. At the beginning (nozzle outlet) the Volume Fraction is commonly with value 1.0 and to the end of my simulation region it is going to spread and the Volume Fraction is decreasing. I would like to know now, if there is a correlation between Volume Fraction and Drop Distribution. Does somebody know a good way to figure this out? I hardly found any papers focusing on this issue... Thank you in advance for your answers!

June 9, 2018, 08:07		#28
Spray_Ansys Member Join Date: Mar 2018 Posts: 30 Rep Power: 8	Yes, I mean drop size distribution. What if I setup an multiphase Eulerian model but with two continuous fluids? How can I sort it out then? Did someone find some literature about that before?