I'm trying to model liquid atomization and primary breakup in a converging- diverging nozzle using FLUENT. The geometry is attached below. It is a cross flow of water on a high speed jet combustion vapor. To start with am trying to model a cold flow simulation of water- air, which would create a spray at the exit.
From my initial research of the problem in hand, i have decided to use Euler-Lagrange model. Gas phase described in Eulerian frame of reference and liquid phase using DPM model of FLUENT. Weber number of the liquid phase is high. At the throat region, speed is sonic (Mach no = 1). I am not sure what sort of atomizer model ( plain orifice or effervescent or any other) should i use to describe the liquid flow. In most of the built in atomizer model, the flow of two phases are co- axial. However, in my problem it is cross- flow.
So the primary question is how should i inject my liquid phase in DPM model of FLUENT? I would like to use blob injection model, so the initial diameter is that of the nozzle diameter and then use KHRT model to study the primary break up. In my problem, am interested in the primary break up, that is the region where it form liquid sheets, film for the core liquid. My aim to study the regime before the droplet formation.
Any sort of input, ideas are appreciated and it would help immensely in getting a step closer to solve the problem.
Dear, Shailesh Konda
1) I agree that the phases are in cross flow type at the inlets. It can affect to the flow structure at the downstream. However, at the region of your interest, the flow is coaxial flow in my view. It is not cross-flow type.
2) Then, according to DPM, it is applicable to study collision of discrete particles and further breakup of discrete particles (to smaller sizes).
3) I think your interest is liquid sheet formation soon after liquid nozzle outlet. In my view, your interest is a kind of collision of liquid particles. And you stated that you will set initial discrete phase diameter the same as liquid nozzle diameter according to the model you used. So, it has no other particles to collide at the same time. It will just breakup to future smaller particles. I am not quite sure that it will fulfill your interest. In my view, you may choose other model that can set smaller particle size that come out form the liquid nozzle at the same time. For that condition, you may see temporal collision before further breakup.
4) Please note that I am not criticizing your idea. I am just giving discussion on your interest. I am also interested in it. I will wait and follow up answers of other researchers.
Hi Nay Zar,
1) I am interested in studying the region close to the the nozzle exit. So, am predominantly focused on investigating the primary break up process of liquid core. For now, am not bothered about the spray (secondary breakup regime).
2) In the DPM model, how and at which location should i define the injection of liquid parcels? Should i use plain orifice atomizer model or should i go with something else. I am not still not sure, what would be the best way to define injections.
3)Since the liquid side inlets are horizontal, i believe the liquid injection will also be along the same axis. I may be wrong in my understanding.
4)I appreciate your thoughts, it has helped me understand the problem better.
If yours is a fundamental study of liquid atomization and primary breakup, DPM model may be inappropriate; you should go for VOF method. If you have certain assumptions about breakup, atomization process, Weber number you may choose appropriate model from the available list using DPM model. However, it appears from your situation that you should go for VOF method with proper grid resolution.
Hi Sumer,
Yes, that was my first approach to go for VOF method and am still trying. I hit a road block in grid resolution as you suggested. The grid resolution should be sub-micron to actually see the droplet formation. VOF looks suitable for slug or stratified type of flow. Do you know any advance interface tracking mechanism for jet two phase flow that can be implemented in FLUENT.
A suggestion: As to define the proper injection point or region. Run a VOF as you did or a Euler-Euler simulation with a more refined grid and time step and let the solution of this run indicates to you based on the volume fraction field where is the better location. Set a good time step to capture the drops formations and the superficial tension force model parameters.
Marek, my impression is that it matters how the twin-fluid mix approaches the throat of the nozzle, therefore, one needs to include some effects of fluid-solid ahesion forces. It is not as if the doplets only accellerate, membranes get to be deformed and break up in this region, hence, the starting point - the representative droplet - is not known, but can be manipulated.
Sorry for being mute for a long time, business issues became overwhelming, will update you on the Oksana project shortly.
Jurek
hallo friends,
im using DPM for injection if particle in my domain.do you have any idea how to change the absorption coeficient of the injected particle.I think i have to make a UDF. anyone some experience in it of a usefull example
when I am creating DPM injection why I am not getting water-liquid in dropdown material?
- Badges
- Science topic