That's an amazing video mate! Especially that mention that, the dispersed-continuous and continuous-continuous interaction differs in how we define the interfacial area per unit volume. It was super helpful.
Hi Aidan. This video was very informative. A video on Lagrangian Multiphase model along with a comparison to Eulerian will be good sequel to this one. Cheers :)
Today, I watched this lecture and subscribed your channel. Your lecture was compact with quality. I am hoping some lecture related to interface tracking in multi-phase. I mean VOF model with explicit function. Thanks for this video.
Great explanation. I would request you to kindly give an insight to the modelling of mass transfer mechanisms between the different phases in your next video.
Hi Hassan, the mass transfer mechanisms are very tricky and im going to have to put some careful thought in before making a video. The essence of mass transfer is equal and opposite source terms in the volume fraction equations so that you get an effective transfer of mass between phases where you want it. For now i hope you have enough to help out with your CFD, while i research the video 😄
@@fluidmechanics101 "The essence of ... source terms in the volume fraction equations ...": this eqn reads partial r_q/partial t+div(r_q U_q)=0 in your video, do you mean in the case of mass transfer between phases, e.g., phase change, we should add source terms on the rhs of the original volume fraction eqn?
So in continuous-continuous flows, we don't normally model the ''surface drag'. For two incompressible fluids, the interface velocity is continuous. ie there is no phase velocity difference. However, I am talking about the single-fluid approach - which is the standard approach for fuel/tank slosh - maybe it is different in the multi-fluid approaches. I have never header of using the gradient as a stand-in for the surface area. I mean there is no mathematical reason for the gradient of the volume fraction to be related to the surface area, they are spatially orthogonal quantities. The gradient is used frequently as the interface normal (when normalised). For an algebraic VOF scheme, knowing the surface area of the interface is next to impossible, but for geometric methods, it is implicit in the scheme. Sorry perhaps because I work building these models I am being a bit critical. I love this channel in any case. Please don't take it personally, just a bit of confusion on my side. :D
Good comments! I always appreciate feedback from people who actually develop the models, as it really helps clarify the understanding. The best I can ever go for is reading the manuals, equations and trying to make deductions from these 😄
@@fluidmechanics101 Thank you so much for taking the comment in good spirit. I was really worried it came across wrong and I really find these videos of great quality and I appreciate the work you do. I have learned a lot, and it's nice to be able to absorb advanced content in related but different areas of CFD. If you need to reach out regarding multi-phase stuff feel free. I am a Post-Doc in InCFD at the University of Cape Town. :D Thanks again for this channel.
Nice video, thanks. However, in the drag model, please note that what counts is the frontal area of a sphere, not its surface area. The Cd of a body is generally defined using its frontal area. Therefore, equation 12 turns out to be 3/2*rp/dp, not 6*rp/dp
@@water5825 The surface area of a sphere is 4*pi*r^2, the frontal area is the area of a circle, which is pi*r^2. So as you can see, the frontal area is 1/4 of the surface area!!
Good work. Your explanations are simple enough to understand. Keep continuing with the good work. I am a masters student in CFD so it sounds familiar. I would request you to make videos on hybrid RANS LES models and probably on parameters such as q criterion and vorticity.
Wow, that is an unexpected leap - right into multiphase! Great lection again :> PS I know it's too much to ask for - but .. any plans for lections on evaportaion/condensation/boiling , break-up/coalescence, solid dispersed (granular temeprature etc)?
😄 im glad you liked it! I havent looked into any phase change processes yet (they are quite hard) but it is something im going to look into in the future! General multiphase is the best place to start for sure with mass and heat transfer terms
Thank you for this amazing video, i just have a question regarding the VOF and mixture models, in Fluent when choosing VOF model we have also the option to choose dispersed interface , and when choosing mixture model , we have the option to choose sharp/ dispersed interface, could you please give more details about that ? Thanks
Hello Aidan At 20:50 in the video, is that continuous-continuous interaction Cd is calculated same as Dispersed-continuous interaction Cd? All using the Schiller model? Thanks
Thank you, may I ask a question ? Assume we are spraying water in hot air (~1000c). some of the water will evaporate. Now, At the beginning of time t=0 we know r_vapor = 0 every where (we don't have vapor yet). If we consider the equation (4), r_vapor will not change as time progress (r_vapor = 0 for every time step). Is this correct ? Should not we add a volume fraction term (representing the transfer from water to vapor) as we did to the continuity equation ?
Thank you for the clear explanation! For a project I need to simulate the dynamics and concentration profiles of polydispersed nanoparticles in a microfluidic and I was considering to approach the model using the Euler-Euler method in COMSOL (treating them as a dispersed phase), also introducing LES equations. Can this model be considered for nanoparticles? Do you think this approach could work?
Hi Samuel, yes your approach seems reasonable and will probably work. Im not sure if LES is necessary though. It depends on what quantities you are interested in studying.
Mostly, i use the CFD user manuals from ANSYS, the OpenFOAM source code and textbooks. Often these sources arent very good, so i try and work things out for myself as well 😊 I then try and compile everything together in a really clear form for you guys, as there really isnt a great source of CFD information out there. Thats what im trying to make 😄
@@fluidmechanics101 Thank you for your reply, I asked that becausa I felt the same think as you said and I wanted to confirme. I am doing my course conclusion work and I would like to have some lectures to put on the reference and explain all the things that I have done in my research (that is to develop a tool using OpenFOAM to predict the resistance curve for a ship of the Serie 60 merchant ships) . If you could help in some way with that I would be very thankful, and thank you for your help and your good videos!
Thank you for the video. I had a question though, doesn't the two-phase full eulerian model also have an inequality bound restraint on the phase fractions (i.e. 0
Part of that restraint is imposed in the way you obtain your pressure equation, additionally you have to properly integrate the phase's transport equations.
Hi Aidan, if I want to simulate in Fluent for Moisture diffusion into something (like Electronic Packaging with voids inside them that allow moisture to diffuse in), do I need Multiphase model: Eulerian in this case?
For moisture diffusion you can just use a mixture model. In fluent you need to define a custom fluid as a mixture of air and water vapour. Then enable species transport and fluent should solve for the mass fraction of water vapour which you can convert into a humidity. With moisture calculations, there are no distinct free surfaces or interfaces between the water droplets and the air so you don't need an Eulerian / Volume fraction based approach. I hope this helps!
Hello Aidan. I Have a question. In air-water multiphase flow CFD simulation's it is common to use a VOF = 0.5 to track or identify the position of interface or free surface. I would like to know why this criteria is used and if it can always be assumed the same. Otherwise, I would like to know if volume fraction can be taken different to 0.5 and, if so, ask you about a methodology to find this volume fraction. In Summary: Which value of volume fraction I can to choose for to identify free surface in an air-water simulation of open channel? What you do is great! Thanks for sharing your knowledge!
Hi Leonardo, yes you are right, it is common to use a volume fraction of 0.5 to track the free surface. This choice is arbitrary of course and you could really use any value between 0 and 1. The important thing to remember is that the CFD code can only resolve the free surface to within a cell. So perhaps a better way of defining it would be: the free surface exists in a cell where the volume fraction changes between 0 and 1. We cant say where in the cell it is, only that it is within a cell. So 0.5 is just useful as it is between 0 and 1. Interface compression algorithms usually try and steepen the gradient so that the volume fraction changes within 1 cell (instead of several) so it makes it easier to locate the free surface. I hope this helps 👍
In algebraic VoF approaches (CICSAM) this does get used. However, in the actual physics governing equations, there is no distinction, just a region of smoothly varying material properties (its why we do VoF). The 0.5 can help to localise certain terms for example surface tension and evaporation. Although even here typically you don't find 0.5 being used rather volume fraction \in (tole, 1- tole). The geometric methods have a discrete representation of the interface so for them, they 'know' where the interface is. In all, I would say the use of 0.5 is most commonly a post-processing thing.
@@thunder852za Very interesting. Do you recommend some book or paper where I can read about algebraic VoF approaches, particularly COMPRESSIVE AND HRIC MODIFIED? Thank you!
Hi Aidan, thanks for uploading such an informative video on Multiphase modelling. I have something to ask, i.e is DPM also pronounced as mixture model?
Thank you, found the talk very useful! I'm currently implementing VOF for free surface LBM (see my YT channel) and the hardest part is surface curvature calculation. A talk about this would be very appreciated!
good video, but I have some problems with the metaphors. I suspect that you mean "suspension" not "solution" when you discuss Dispersed - Continuous. Specifically this is clear in (c) of sediment transport.
Thanks a zillion for your awesome videos. Honestly, my CFD knowledge has improved significantly since I started watching your great informative videos. Anyway, there was something that I couldn't understand very well at the penultimate slide which was about Ap/V of the continuos-continuous phases. I was wondering if you could explain it to me how this ratio is equal to the gradient of the volume fraction.
I can't remember exactly why this is the case. It is stated in many of the CFD manuals without much explanation. I know that's not very helpful but it's the best I've got at the moment 😅
@@fluidmechanics101 Thank you Aiden. By the way, may I ask you to introduce a source from which I can learn better the discretisation methods for the convection term of the volume fraction equation of the continuous-continuous method? Or if it's possible, could you please make video on them?
As it was mentioned that the dispersed particles are dissolved in the continuous fluid (1:53), So my doubt is if it is a solid particle then will it also get dissolved in CFD simulation? If we don't want the solid particle to be dissolved then what type of modeling approach (i.e., Models in CFD) needs to be followed? Thanks in advance
If your particles are large, then you probably need to look at Discrete Particle Method / Discrete Element Method. It is available in Fluent or MFIX. Not sure about CFX or OpenFOAM though 👍
Very good video, teacher.👏👏👏. Could I know the software that you use to draw the figures in the presentation? They are excellent! Lot of thanks in advance💞
@@fluidmechanics101 Thank you very much teacher. I will learn Inkscape from the video link. I have just joined the udemy course. You are the BEST💞👏💞👏💞👏💞
Hello Aidan, It’s a Very nice video. So multiphase modeling is applied for phase transition between liquid and vapor refrigerant in a small diameter channel with larger length. But why most of the time solution diverge or temperature doesn’t not change? Thank you very much Please leave a suggestion
Hi! Thanks for another wonderful video. I was wondering, is there a way to know how much governing equations I'm solving in total? If you e.g. have three phases, but they are coupled via drag / the sum of the volume fractions is one. And only considering mass and momentum (no energy and turbulence). I don't quite understand how fluent solves/couples the equations, and thus how many equations you're eventually solving. Cheers! Michelle
You could always just check the residuals that fluent writes out? Fluent will write the residuals for every governing equation you are solving, so this is a quick way to check. For example: pressure x-mom y-mom z-mom volf-1 volf-2 would give 6 governing equations? (3 momentum, 1 pressure, 2 volume fractions) which is what I would expect for 3 species
@@fluidmechanics101 Thanks Aiden. I have another query. In a dam break simulation using VOF, what will be the bc to be applied on the inner surface of the fluid while solving the volume fraction equation?
Hello sir, may i ask a few question? Lets say i want to insert diesel fuel in chamber that contain air as their fluid. should i use multi phase model in order to separate the two fluid which is diesel and air to identify the flow of the diesel and if yes which model should i use?.. Thanks sir for answer my question before..
Yep, i would use an inhomogeneous multiphase method (seperate momentum equations for diesel and oil). I would only go for a homogenous method if you are sure that the drag forces are only significant near the interface
I would estimate the volume fraction of each phase and then calculate the volume averaged material properties (density, viscosity). You can use these to get an estimate of the Reynolds number, friction coefficient and y+
Dear sir... I referred to your video... can you add energy equation to eulerian model...??? I have checked the ansys fluent theory guide... there energy equation is not clearly interpreting in nature..in fact the energy equation contains the shear stress tensor and velocity vector....though energy equation should contain temperature or enthalpy term.. it will be helpful for me if you can reply...
I use latex with the beamer package. If you look through the patreon posts i have put up the template that i use, so you can use it yourself if you like 👍
Hi Aidan, Thank you it was really helpful, I have a question. in the solution of volume fraction equations for three-phase (sharp interface), how it avoids overlapping the interfaces? in references the explanations about the interface tracking are all about the two phases condition.
It was a great video, I got this project called modeling of liquid-gas flows using VOF-DBM method and this surely was helpful, so I will be needing more guidance, could you help me out with that?
The liquid water phase loses mass, while the gaseous vapour phase gains mass. So the source term is negative for the liquid phase and positive for the vapour phase. Of course if you have condensation, it is the other way around 👍
Sir, can you make another video on the algorithm/procedure to solve these complex equations.like I didn’t get how do we get pressure equation and solve it.
Yes, I am working on this at the moment. It is quite confusing and the pressure correction equation is definitely one of the hardest parts to understand in CFD
![[CFD] Eulerian Multi-Phase Modelling](/img/logo.svg){kind=logo}
