Multiphase Flow - Implementing AIAD model

saviosas · March 8, 2016, 09:45

I am trying to reproduce an article of Thomas Hohne about multiphase flow (water and air), with objective to use the same model that he used for the drag coefficient in the free surface in my work. The article is: application of new drag coefficiente model at cfd-simulations on free surface flows relevant for the nuclear reactor safety analysis.

Right now, I am having some issues with implementing the area of free surface of the Algebraic Interfacial Area Density model (AIAD). According to article this surface area is the gradient of liquid volume fraction. The expression for that I wrote below:

Afs=sqrt((Water.Volume Fraction.Gradient X)^2+(Water.Volume Fraction.Gradient Y)^2+(Water.Volume Fraction.Gradient Z)^2))

This expression always results in value of 0. Is this expression right? Or there is another way to write this gradient expression?

af123 · December 19, 2016, 09:03

Do you had any succeed with the model? I'm also interested in the model.

stardust111 · May 11, 2017, 19:45

Hi Savio，

Does it work now ?

ves · May 9, 2020, 10:50

Hi, Salvio I have my AIAD GENTOP implementation on CFX Expression Language, but mass exchange not working. Do you interesting in this model yet?

sachin tom · May 15, 2020, 01:51

Hi Ves,
I have certain doubts AIAD GENTOP implementation. Is it default available as CCL expression in CFX? Also is the technique used for blending from bubbly to annular flow regime?

ves · July 18, 2020, 15:36

ANSYS CFX have not AIAD and GENTOP model, Fluent 2020R1 have AIAD. I have my own implementation with some bugs

Stel · July 19, 2020, 10:43

The very last CFX version has an AIAD model implemented as a beta feature, which as far as I understand is a very similar model to the one proposed in the Haensch et al. (2012) paper (of which Dr. Hohne is a coauthor). It is hard to tell exactly because unlike Fluent no documentation is provided by ANSYS for CFX'S Beta features. However, notice that the default volume fraction limits in the software (0.3) are the same as proposed in the paper, as well as the minimum CD value in the free surface regime (0.01). Given the long term relationship between HZDR and ANSYS (notice that Liao's breakup and coalescence models, also from HZDR, were also included as beta features in the recent version) I would bet this AIAD model is the one that you are looking for.

Good luck with it BTW. I like the concept behind this model, but those critical limits are hard to define for general purposes and besides, convergence seems to be very difficult with it.

ves · July 24, 2020, 14:59

Dear Stel!
Thank you very much for you responce. I had tried solving plunging jet problem from the Haensch et al. (2012) in Fluent 2020R1 and received unstable solver behavior AIAD model for phase coupled SIMPLE.I had soved it with Coupled solver with adaptive timestep 10^-7-10^-8 s first 1000 steps, then with 10^-5 s. AIAD+Ingomogeneous Discrete had diverged. I had tried AIAD in CFX 2020R1 as you said and reseived good results with timestep 10^-5 s. AIAD are not compaatible with polydispersed fluid in CFX 2020R1. How i may improve stabiliy in Fluent?

Stel · July 24, 2020, 16:22

Well, as I said, convergence with this model seems to be very hard with its current implementation (hope they improve its stability in future versions). A timestep of 10^-5 s is already a very short one. And I didn't know it cannot be used together with the polydispersed fluid definition in CFX, but something you could try: once I was trying a particular setup with a polydispersed fluid involved and got a warning message telling me that one specific model I was trying to use was not allowed with a polydispersed fluid in the simulation; I tried to start the solver in spite of the warning message and it worked anyway. You could try it yourself (at your own risk; carefully analyze the results afterwards to see if it makes sense).

As for your Fluent problem: you should ask people in the Fluent forum about that, but I'll try to suggest the obvious: 1) try to use implicit solution for the volume fraction equation; 2) if your are solving it as a pseudo-transient, try reverting to transient and use several coefficient loops between steps; 3) try a solution first without the continuum surface force activated.

Good luck.

March 8, 2016, 09:45	Multiphase Flow - Implementing AIAD model	#1
saviosas New Member Savio Join Date: Mar 2016 Posts: 1 Rep Power: 0	I am trying to reproduce an article of Thomas Hohne about multiphase flow (water and air), with objective to use the same model that he used for the drag coefficient in the free surface in my work. The article is: application of new drag coefficiente model at cfd-simulations on free surface flows relevant for the nuclear reactor safety analysis. Right now, I am having some issues with implementing the area of free surface of the Algebraic Interfacial Area Density model (AIAD). According to article this surface area is the gradient of liquid volume fraction. The expression for that I wrote below: Afs=sqrt((Water.Volume Fraction.Gradient X)^2+(Water.Volume Fraction.Gradient Y)^2+(Water.Volume Fraction.Gradient Z)^2)) This expression always results in value of 0. Is this expression right? Or there is another way to write this gradient expression? stardust111 likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
multiphase flow ，mass and volume fraction imbalance	sope111	CFX	16	September 3, 2018 01:10
Non-Newtonian liquid/air multiphase flow models	Wonder	Fluent Multiphase	0	April 7, 2015 07:59
multiphase flow vof model with mass transfer	frikz	FLUENT	1	October 27, 2014 06:12
multiphase flow in porous media	zhou	FLUENT	2	August 9, 2012 08:10
Multiphase flow in atomiser	santhosh1987	FLUENT	0	May 12, 2011 05:26

December 19, 2016, 09:03		#2
af123 New Member Andre Join Date: May 2015 Posts: 3 Rep Power: 11	Do you had any succeed with the model? I'm also interested in the model.

May 11, 2017, 19:45		#3
stardust111 New Member Guang Join Date: Feb 2015 Location: Stuttgart, Germany Posts: 15 Rep Power: 11	Hi Savio， Does it work now ?

May 9, 2020, 10:50		#4
ves Member Veskov Eugene Join Date: Feb 2011 Posts: 31 Rep Power: 15	Hi, Salvio I have my AIAD GENTOP implementation on CFX Expression Language, but mass exchange not working. Do you interesting in this model yet?

May 15, 2020, 01:51		#5
sachin tom New Member sachin Join Date: Dec 2016 Posts: 7 Rep Power: 10	Hi Ves, I have certain doubts AIAD GENTOP implementation. Is it default available as CCL expression in CFX? Also is the technique used for blending from bubbly to annular flow regime?

July 18, 2020, 15:36		#6
ves Member Veskov Eugene Join Date: Feb 2011 Posts: 31 Rep Power: 15	ANSYS CFX have not AIAD and GENTOP model, Fluent 2020R1 have AIAD. I have my own implementation with some bugs

July 19, 2020, 10:43		#7
Stel Member Henrique Stel Join Date: Apr 2009 Location: Curitiba, Brazil Posts: 93 Rep Power: 17	The very last CFX version has an AIAD model implemented as a beta feature, which as far as I understand is a very similar model to the one proposed in the Haensch et al. (2012) paper (of which Dr. Hohne is a coauthor). It is hard to tell exactly because unlike Fluent no documentation is provided by ANSYS for CFX'S Beta features. However, notice that the default volume fraction limits in the software (0.3) are the same as proposed in the paper, as well as the minimum CD value in the free surface regime (0.01). Given the long term relationship between HZDR and ANSYS (notice that Liao's breakup and coalescence models, also from HZDR, were also included as beta features in the recent version) I would bet this AIAD model is the one that you are looking for. Good luck with it BTW. I like the concept behind this model, but those critical limits are hard to define for general purposes and besides, convergence seems to be very difficult with it.

July 24, 2020, 14:59		#8
ves Member Veskov Eugene Join Date: Feb 2011 Posts: 31 Rep Power: 15	Dear Stel! Thank you very much for you responce. I had tried solving plunging jet problem from the Haensch et al. (2012) in Fluent 2020R1 and received unstable solver behavior AIAD model for phase coupled SIMPLE.I had soved it with Coupled solver with adaptive timestep 10^-7-10^-8 s first 1000 steps, then with 10^-5 s. AIAD+Ingomogeneous Discrete had diverged. I had tried AIAD in CFX 2020R1 as you said and reseived good results with timestep 10^-5 s. AIAD are not compaatible with polydispersed fluid in CFX 2020R1. How i may improve stabiliy in Fluent?

July 24, 2020, 16:22		#9
Stel Member Henrique Stel Join Date: Apr 2009 Location: Curitiba, Brazil Posts: 93 Rep Power: 17	Well, as I said, convergence with this model seems to be very hard with its current implementation (hope they improve its stability in future versions). A timestep of 10^-5 s is already a very short one. And I didn't know it cannot be used together with the polydispersed fluid definition in CFX, but something you could try: once I was trying a particular setup with a polydispersed fluid involved and got a warning message telling me that one specific model I was trying to use was not allowed with a polydispersed fluid in the simulation; I tried to start the solver in spite of the warning message and it worked anyway. You could try it yourself (at your own risk; carefully analyze the results afterwards to see if it makes sense). As for your Fluent problem: you should ask people in the Fluent forum about that, but I'll try to suggest the obvious: 1) try to use implicit solution for the volume fraction equation; 2) if your are solving it as a pseudo-transient, try reverting to transient and use several coefficient loops between steps; 3) try a solution first without the continuum surface force activated. Good luck.