Discussion Groups

Simulation CFD

Reply
Member
willnich35
Posts: 3
Registered: ‎02-28-2013

Multiphase Flow - Water Droplets

342 Views, 4 Replies
02-28-2013 08:34 AM

Hello,

 

I am currently evaluating various CFD software packages for my company. We design and manufacture louvres, and the idea is to improve the efficiency of our designs using CFD. A highly efficient louvre blade design requires both a low pressure drop across the blades and an ability to keep water droplets from passing through. I have successfully analysed the pressure drop across the blade, with a high degree of correlation between the CFD and experimental testing. 

 

I would like to know if it is possible to analyse water droplets in any of the AutoDesk Simulation packages. Something like the particle transport model in Ansys CFX. I have the trail version of AutoDesk Sim CFD 360 2013 and am struggling to find any such capabilities.

 

Thanks!

Please use plain text.
*Expert Elite*
OmkarJ
Posts: 427
Registered: ‎10-02-2012

Re: Multiphase Flow - Water Droplets

02-28-2013 11:42 AM in reply to: willnich35

From what you describe, I believe you are hinting at Langrangian particle tracking like DPM in Fluent or PTM in CFX. I think the only multiphase capability available in Autodesk CFD is scalar mixing. But you can use "massed particle traces" to see how water drops travel, after simulating with normal air. From what I read, the equations for velocity and position of the individual particles are solved, taking into account the drag and inertial forces on the water droplets.

 

Having said that, I see no mention of turbulent dispersion or virtual mass force etc but they don't seem much relevant to your kind of physics. Also, I presume it is a one way coupling and not two way coupling so point masses of the water drops do not influence the physics of the air.

 

The equation of the drag law suggests that it is possibly a Schiller-Naumann correlation, ideal for spherical water droplets in air. For serious results, you may want to do a bit of literature search to get the best constants in the correlation for your kind of physics, instead of relying on the default ones. 

 

More experienced folks may have more to add...

 

Cheers

OJ

Please use plain text.
Member
willnich35
Posts: 3
Registered: ‎02-28-2013

Re: Multiphase Flow - Water Droplets

03-01-2013 01:43 AM in reply to: OmkarJ

Thank you very much for that detailed answer. I have just found the massed particle documentation so will have a play around with that. Ideally we would like to be able to model wall film and particle 'splashing' when they hit a surface, although I may be getting ahead of myself. Il use the coef of restitution to start with and see what results I get.

 

Thanks!

 

Please use plain text.
Member
willnich35
Posts: 3
Registered: ‎02-28-2013

Re: Multiphase Flow - Water Droplets

03-01-2013 04:39 AM in reply to: OmkarJ

 

Just a quick additional question to any takers... Is there a way to create a bulk result file or summery plane which will measure the mass flow rate of massed particles through the plane?

 

Cheers

Please use plain text.
*Expert Elite*
OmkarJ
Posts: 427
Registered: ‎10-02-2012

Re: Multiphase Flow - Water Droplets

03-05-2013 12:56 AM in reply to: willnich35

Not sure if there is a direct way of doing this. But this comes to mind:

 

1) Create  a plane in the fluid domain. Right click on it and set it as summary plane.

2) Navigate to Results >> Decision center and in the left pannel visible now, right click on critical values and update

3) In summary panes at the bottom, select volume flow and using this value calculate average velocity over plane

4) With this, you can calculate the mass flow of the massed particles as : rho * a * v_avg * n

   where a is cross section area of water drop, n is no. of particles selected, rho is density of water..

 

Now there are two assumptions made here which may be unrealistic for larger water particles:

 

1) The velocity of water drops is same as that of air

2) The number of water drops depends on what you specified. You may need a bit experimental data to make sure you specify correctly.

 

That's all I can think of as of now...

 

Cheers 

OJ

 

Please use plain text.