Thank you very much Doctor Aidan for your excelent explaination. But there might be a mistake about the definition of the "superficial velocity". I have read different explainations about the "superficial velocity" online, The superficial velocity ist defined as the volume flow rate divided by the cross area, which is, i think, the velocity "U", that you mentioned in your video. The "Us", which you called superficial velocity is maybe the physical velocity. In the the superficial velocity is equal to porosity times physical velosity, which means the superficial velocity is smaller than the physical velocity. But according to your definition the superficial velocity will be bigger than the physical velocity. maybe its just a mistake :)
Mate, I’m a research engineer with a full-on doctorate in CFD and after all these years I still come back to your videos over any text book or technical paper if I need to recap or familiarise myself with some aspect in OpenFOAM or fluid dynamics in general. There are students out there paying over 10 grand for an engineering education that is 10 thousand times more shitty in quality and you’re providing all of this for free. You’re an absolute legend. Keep up the good work.
Great video as always. Another resource with a lot of empirical C2 or K coefficients is the "Handbook of Hydraulic Resistance" by I. E. Idelchik If there are no manufacturer's data or you can't find a C2 or K coefficient in literature there is a 3rd way Run a virtual wind tunnel CFD simulation of a small portion of your grill/louver/heat exchanger etc for increasing face velocities and plot the pressure drop. You can then extract C2 or K with a polynomial fit as with the 2nd method you described.
@@danmorris9533 I think what @themanolis1984 meant was to make a small part of the material that you want to study as porous zone, as realistic as you can; meaning actually designing the geometry of the domain (as a realistic porous zone, not using porous zone option in your CFD simulation) and running the simulation to get the "pressure drop vs velocity" data, then use that data for simulation in larger scale using "porous zone" option in CFD code. One assumption for this approach I think is that the porous material is isotropic and homogeneous.
Thank you very much Doctor Aidan for your excelent explaination.I would like to ask whether it is necessary to add boundary conditions at the interface between the porous medium and the fluid region? When I simulate porous media, there is always a droplet splash phenomenon at the free surface or side wall, and then the vof analysis problems, the program can not run, I have not found the reason, want to ask the teacher what may cause this situation?
I think the definition of superficial velocity is incorrect. It is a volume averaged velocity where averaging is done over the volume containing both fluid and solid. This velocity is basically the product of porosity times the intrinsic average velocity where now the averaging is done over the fluid volume only. The velocity you mentioned in the schematic is the pore velocity or the intrinsic average velocity.
@@fluidmechanics101 Hi Aidan and Sayuj, So in method 1 where data for pressure difference vs velocity is given by suppliers, is that "velocity" : superficial velocity?
Thank you for your very detailed explanation about porous domain in CFD. Could you please handle about heat transfer in porous domain also? I'm now using porous domain to calculate conjugate heat transfer of plate heat exchanger, I have difficulty in setting a heat transfer model.
10 or 100 times the value in the fluid direction is normally fine. You just need the resistance to be higher than the fluid direction, so that the flow is encouraged to straighten and flow in the fluid direction. A small amount of flow in the solid direction is fine and makes convergence a lot easier
Nice video, thank you! I am using a porous medium to represent a spiral tube bank so the flow has a non negligible radial and circumferential velocity, how can I calculate c1 and c2 values for this?
Thank you Dr for this seance, I have questions: 1- in higher reynolds number what is the model suitable for pouros media Darcy or Darcy-Forchheimer? 2- How does the setup of pours media in ansys fluent?
Thank you very much sir. Please, how can i calculate these factors when ther is a melting of a phase change material (wax for example). So there is no flow. Just the solid wax melt inside the closed foa(porous media) when the temperature rises. Thank you before.
Solidification and melting is very hard to do in CFD. Unfortunately I don't think I can help you. You might need specialist software for those calculations
Thank you for this video. You really helped me. But I have a doubt in slide No. 16… when you are talking about the units. I can see the permeability unit is m^2 but what happened with the unit of seconds of the dynamic viscosity?
Hi! I found the same doubt, the units of pressure and C1/mu are both incorrect. Dimension of pressure shall be [M][L]^(-1)[T]^(-2), and dimensional of C1/t shall be [M][L]^(-3)[T]^(-1). However, the resulting dimension of alpha is correct as [L]^(2), LOL.
Hi, sir thank you so much for such a clear and lucid explanation about porous media. I am calculating the porous coefficients for a thin perforated flow in the turbulent regime. While calculating the porous constants from polynomial curve fittings, I am getting viscous porous resistance, C1 as negative. Can you please explain such scenarios are possible and the physics behind the same.
Dear Mr. Wimshurst, I appreciate your work, it's a big help for me. All your videos are very helpful. Can I have a question: are you planning to make a video about FAN boundary conditions (in OpenFoam etc.)? Thank you :-)
Thank you for this video it is really helpful. I just have one question, since we know the pressure drop across the porous media, during simulation can i convert my 3D geometry into a 2D geometry and while putting in coefficients divide by the actual length of porous media.?
Thanks for your video. It is very useful to understand how to porous media is work. I want to model blood clot as a porous media but some paper publish only porosity and permeability values. How can I get C2 (head loss coefficient) without using velocity-pressure drop data?
Thanks a lot. However, the pressure drop data is mostly available with respect to flow rate and not versus velocity in industry so curve fitting of DP versus flow rate to get the coefficients remain valid ? thanks
Great job, thanks a lot. I am trying to model a capillary membrane, can I use this model? Do u have any tips to calculate the c1,c2 if the porous media is as fine as a membrane (pore size is 7 µm)
The flow is probably laminar through the holes (you can check Re). If it is laminar then you can use the analytical solution for flow through a pipe / parallel sides channel to derive the coefficients C1 and C2. In pretty sure if it is laminar then you will find that C2 = 0, and the C1 will depend on the viscosity and pore size
@8:27 you said that velocity will be increased due to reduction in area but as you considered before a cross section of equal area as that of pore then there will be no increase in velocity ,so what about superficial velocity then
Thanks for this lecture and your clear art of explanation! One question, why you choose to set C_2 as big number and not C_1 to describe wall or very high flow resistance in particular direction?
@@fluidmechanics101 thank you for answering. I‘m struggling now with one problem connected to this lecture. I‘m modeling two-phase flow (h2o + h2) through porous media with Eulerian approach, capillary pressure and relative permeability in Fluent. Maybe you have some lecture or can suggest some good material dealing with similar issues and explains the theory behind it. Thanks & Kind Regards
I see in many cases that people say we use the porous medium to define a geometry where there is no flow actually. Sir, do you know how we define a porous medium that does not allow any flow through it?
Yep, just set the resistance coefficients to be high. This will block most of the flow. Values of around 1000 seem to work quite well. If you set them too high the flow will stop completely and the solver may diverge, so be careful!
Thank you very much Doctor Aidan the porous media can be used for tube bundle. please , How I can used porous media with tube bundle using ANSYS, or what is the equations of C1/t and C2/t thank you very much Hayder kraidi Rashid
You will need to look up empirical equations for the pressure drop (Euler number) for arrays of horizontal cylinders. Then you can equate these equations to the formulae for the source terms in Fluent and rearrange for the coefficients C1 and C2
In the video you mention that coefficients can be found in some textbooks. Could ypu specify which textbooks? I am particulary interested in heat exchangers. Thank you!
Quick question. When you were talking about C1/t being equal to dynamic viscosity/permeability it should be in kg/(m^3*s) should´t it? Because dynamic viscosity unit is kg/(m*s) and permeability is m^2. So when you divide it it should get kg/(m^2*s) or I am missing something? Because when I use polynomial formula in my pressure drop data the units check up because of delta_pressure=C1*U+C2*0,5*density*U^2 you need to get Pa = kg/(m/s^2). Correct me please if I´m wrong. :) Anyway thank you for great contect it was really helpful.
ı wathced the lesson carefully. thanks a lot. Do these coefficients change for heterogeneous and homogeneous porous regions?It does not change when it is calculated as an area. but i think it can change physically. my case study is airflow through people who are standing cramped up so heterogeneous permeability.
I suppose you could either average the people and do a homogeneous porous media or break your porous media into regions and assign different resistance coefficients for each region? (So effectively heterogeneous)
Hi Dr.Aidan. In a car company, We use one set of the viscous and inertial coefficients for the pressure drop across the porous media(heat exchange), instead of three set of coefficients. that means air will flow in all direction, which is not correct behaviors, we expect air will only flow in one direction. Do you know why? Nobody knows in our company, we bought CFD methods.
Hi there! If you are expecting flow in a single direction, then you can make the loss coefficients much larger in the other directions. This will encourage the flow to go in the direction you want and give the correct loss coefficients. Remember: the flow will always try and follow the path of least resistance 👍
One suggestion . Can u present the topic regarding the modelling volumetric heat source within the components .The reason for asking this is it follows the same approach as the porous medium (Defining the source terms and how this source term works in defining the temperature) .Would highly appreciate .Thanks in advance .
What kind of heat source are you thinking? Im pretty sure you just take the actual heat source (in Watts) then divide by the volume of the cell zone (to get Watts/m3). You might need to use a UDF or User-Fortran or fvOptions if your source is more complicated
Thanks, super helpful! Just to clarify, you should mesh your porous zone as 1 cell thick in the porous direction, so when you divide your Coeff.s by t, it would come out right, correct? or if you have n-cells in the flow direction, you need to divide t by n?
Just divide by t. The source terms are applied per unit volume so they will multiply by the volume of the cells. More cells just distributes the source over more cells
thank you so much professor, How to input the values of relative humidity (moisture content) as inlet condition and check the moisture content in the porous zone? Please tell us how to do this.
Hello Aidan, excellent presentation again, congratulations. Would you say we ought to use a porous zone method if the viscous forces are dominant in the flow? Im just trying to figure if it would be possible to simplfy even further the computational cost when one deals with a porous zone and choses to deal with the problem using Darcy law (first order velocity equation) instead of the NS equation... Best regards
I suppose it depends on your geometry. Adding an extra source term to NS doesnt really increase the computational cost (adding a porous zone). So if your overall geometry is simple, you could probably just use Darcys Law in a hand calculation. If your overall geometry is complex, just throw in the porous zone 😊
Thank you for a great explanation of porous models! Saves a ton of time having to go through lots of notes to understand it. I can see that the correction about superficial velocity explanation is already mentioned in the comments. Can you please explain the difference between superficial velocity and true velocity?
I would have a look and see if you can find any papers or textbooks with loss coefficients for heat exchangers. I cant recommend any off the top of my head but you should be able to find some with a quick google search
Yep, I am working on this one. It all really depends on the interface sharpening algorithm used for the convection term (have a look at my Eulerian multiphase video for a brief overview)! Interface sharpening is quite tricky, so I am still working on it 😄
How can we double the porosity value for the current case with respect to governing equations. Means which term in the governing equation will tell me about the porosity value?
Hello, I have been looking into modeling porous domains in regards to grain storage and was wondering if you had any thoughts on this. Looking at the literature, I have found pressure drops of (say corn) across a meter bed. Using the curve fitting method, I have found values of the linear term to be in the order of 10+ Million and the quadratic term to be in the order of 10,000. I would be happy to share the google spread sheet with you if you want to see it. Thank you.
From your values it looks like you can just neglect the viscous term and go with just the quadratic term. 10,000 is high so maybe do a quick easy CFD simulation with flow in a cuboid shaped domain with a block of porous material and check that you are getting the pressure drop you expect. If that works and you are happy then go ahead with your full CFD calculation
@@fluidmechanics101 Went back and redid the papers test in cfd, and the values are correct. I ended up using simpleFoam with different relaxation factors (0.2 for pressure and 0.3 for velocity and turbulence values) and everything is working better and, more importantly, converging. Thank you.
Ive just been checking through and it looks like ive made a mistake 😅 especially after insisting on being careful with units! C1/t should have units of [kg/m3 s]. Hence the time unit in the numerator should cancel out to give the correct units of [m2] for alpha (the face permeability). Sorry for the slip up guys, and well spotted on the mistake!
Thanks a lot for your excellent videos, really helpfull ! I have a question regarding method 1 for deriving c1 and c2 (Curve fitting on experimental data). Whic velocity is considered as x-axis ? Local velocity through the obstacle or velocity just upstream the obstacle ?
Thank you very much! One more question about the relationship between the volume porosity and the pressure loss through the porous domain. i am using the CFX to simulate the flow through a gauze. And I've ran lots of cases and found that the total pressure loss not only determined by the loss coefficient but also the volume porosity as well. can you explain this a little bit? fantastic video , thank you again.
Hi, thank you for Your great content. I wanted to ask if it would be possible to model a metal foam for heat transfer in natural and forced convection like this. :) Thank you.
You can find a good description of the discretisation of N-S in my fundamentals course (which is on my website: www.fluidmechanics101.com ) or check out my video titled ‘The Finite Volume Method’ 👍
Thanks for the very nice lecture! One question: If a porous media is made of entangled fibers and the liquid is wicked into it by capillary force, how the C values should be calculated there? (Do we need a microscopic image to measure the average size of microchannels between the fibers, or what)
Thank you so much for this wonderful videos! And I still have a question: How about the energy equation of fluid in the Porous zones? Some papers seems only considered the heat conduction between fluid & solid porous layers, while Ansys Fluent seems used different method, and I'm still quite confused about this.
How does somebody get familiar with CFD? Mainly I am talking about the programming. I am comfortable using python, MATLAB and very proficient with calculus as well as vector calculus. Any advice would be extremely appreciated. However, I have no experience with CAD.
I can recommend you this course from Boston University: th-cam.com/video/35unQgSaT88/w-d-xo.html with additional material on this website: lorenabarba.com/blog/cfd-python-12-steps-to-navier-stokes/
Or you could just jump in and try some OpenFOAM 😄 all the source code is there and there are plenty of test cases, so you wont need to do any CAD/meshing if you just want to have a go and see how the code works. (Just google OpenFOAM and you can find it)
Hello sir, if I have open channel flow with gravel bed can I simulate it as a porous zone and is thickness of gravel bed can be consider the thickness for porose media (t )
Your cell zones should be a fluid. Then you need to tick the box 'porous zone'. This will open up the options which allow you to fill in the coefficients for the zone
I have a cooling circuit including multiple components like motor, inverter, compressor etc. I want to model the circuit in Fluent while incorporating the pressure drop of each component using porus media. However i am not sure how to take care about the thickness as each component has its cold plate and i do not know their dimensions. Is this a correct approach i am adopting ? thanks
Do you know the pressure drop from each component? Or do you have any data / can approximate it? You can use an incorrect thickness as long as you ensure that the pressure drop is the same by adjusting the loss coefficient
Yep, to make sure that the pressure drop is in the opposite direction to the velocity vector, regardless of coordinate system (- U |U|) will always have a sign that is the opposite of U)
Yes, it is a type of heat exchanger. If you think of a cross-sectional slice through the tube banks, then you can model this as a porous zone. The reason that you might not want to model the tubes explicitly is that the flow separation off the cylinders is complex, unsteady and requires a high level of grid (and temporal) resolution. So it is often easier to model a set of cylinders as a porous zone instead. This also allows you to perform a steady simulation 😊
@@fluidmechanics101 Thanks for the feedback. Extra layers on all the tubes would be excessive and vortex street for each one. My other question was how to handle something that doesn't just hold up the flow like a filter but something that redirects flow like a louvre or parallel plates at an angle to the incoming flow? Another great video and explanation all the best for 2020.
Hello, Great explanation. Is it possible to use this method for the simulation of flow through swirl diffuser? I am doing the simulation of airflow in a clean room which has diffuser inlets. Thanks
Yep, you can use it for the filters on the inlet and outlets to the clean room. Try and get some measures data for the pressure drop across the filters (usually the manufacturer will have this)
@@fluidmechanics101 how can the porous zone discharge with the swirl flow won't it discharge perpendicular. Same concerns for a louvre. How do you make sure it has the desired polarizing effect in terms of flow.
Hi! Thanks for the amazing video, it' very interesting. I have one question about how the solver takes in account the of porous media. Does the volume of fluid recognize there is a solid medium in the domain? So for example, if I have a full region of porosity with a percentage of void area of 50%, will the fluid be able to fill only half of the region?
Thanks a lot for this video, it helped me understand where to start learning about the porous zone. I have to model a capillary-membrane, whose wall is permeable and the water enters from the right side. Hier one has a laminar flow, which means Darcy's law is accepted. I am trying to find out c1, c2, could you please support me with some tipps? best wishes, Venesa
And In course 2, section 4 wall functions topic was explained quite fast😢..I have to go through 2 more times to understand I think🙂. Apart from that, Both courses and materials provided were Great. Thank you.
Thats really useful feedback, thanks Nikhil. It is quite a difficult subject and maybe i get ahead of myself making things too concise! I have just started fundamentals 3, so its going to be a while before i finish it. I want to make sure it is really good 😄 the course is finally going to look at unstructured meshes and how we think about putting together cfd codes and solvers. Its going to be fantastic, i know you will love it. However, it does take quite a while for me to make, as the content is original and doesnt exist anywhere on the internet, so i have to be extra careful with checking everything. Im going to aim for December time to get it out to you 😄
thank you very much for this lecture I have a very fine grid in my geom so i decided to instead it with filtering region but i dont have the chart to find "C2" the area's ratio is equal 0.00101085 !! do you think its work with this?? thank you
@@fluidmechanics101 yes it is.... in the last week i work in method that run the grid model and take the pressure drop before and after it and calculate the velocity then draw the graph .....but also i had very large error between the two pressure drops ....is the method i used work here ??
![[CFD] The k-omega Turbulence Model](http://i.ytimg.com/vi/26QaCK6wDp8/mqdefault.jpg)
![[CFD] The k-omega Turbulence Model](/img/tr.png)
![[CFD] Residuals in CFD (Part 1) - Understanding Residuals](http://i.ytimg.com/vi/v9OnNeYH4Ok/mqdefault.jpg)
![[CFD] Pressure-Inlet Boundary Conditions](/img/n.gif)
Thank you very much Doctor Aidan for your excelent explaination.
But there might be a mistake about the definition of the "superficial velocity". I have read different explainations about the "superficial velocity" online, The superficial velocity ist defined as the volume flow rate divided by the cross area, which is, i think, the velocity "U", that you mentioned in your video. The "Us", which you called superficial velocity is maybe the physical velocity. In the the superficial velocity is equal to porosity times physical velosity, which means the superficial velocity is smaller than the physical velocity. But according to your definition the superficial velocity will be bigger than the physical velocity.
maybe its just a mistake :)
Yes this is a mistake. I have pinned your comment so everyone can see 😊
Mate, I’m a research engineer with a full-on doctorate in CFD and after all these years I still come back to your videos over any text book or technical paper if I need to recap or familiarise myself with some aspect in OpenFOAM or fluid dynamics in general. There are students out there paying over 10 grand for an engineering education that is 10 thousand times more shitty in quality and you’re providing all of this for free. You’re an absolute legend. Keep up the good work.
Thank you, it means a lot 🙂
Excellent lecture! The unit of C1/t should be kg/(m^3-s) instead of kg/m^3.
Yes, well spotted 😅 thanks for the correction
Thanks a lot, this video particularly helped me to get a an understanding of simulating flows through barrier.
Explained it in a real easy way..thank you.
Great video as always.
Another resource with a lot of empirical C2 or K coefficients is the "Handbook of Hydraulic Resistance" by I. E. Idelchik
If there are no manufacturer's data or you can't find a C2 or K coefficient in literature there is a 3rd way
Run a virtual wind tunnel CFD simulation of a small portion of your grill/louver/heat exchanger etc for increasing face velocities and plot the pressure drop.
You can then extract C2 or K with a polynomial fit as with the 2nd method you described.
Yes! Completely agree with all your points. Thanks Manolis
do you run this CFD sim assuming no porous zones?
@@danmorris9533 I think what @themanolis1984 meant was to make a small part of the material that you want to study as porous zone, as realistic as you can; meaning actually designing the geometry of the domain (as a realistic porous zone, not using porous zone option in your CFD simulation) and running the simulation to get the "pressure drop vs velocity" data, then use that data for simulation in larger scale using "porous zone" option in CFD code. One assumption for this approach I think is that the porous material is isotropic and homogeneous.
Thank you ! I just search info about the porous region and i got ur video))))
Absolutely amazing channel!
It was amazing thank you so much for perfect explanation. Have a good day
Great video, helped me a lot with my understanding! Good job!
Dude , your videos are brilliant! Thanks so much!
GOAT fluid dynamics teacher
Thanku so much sir. It is so interesting and informating. Please keep it up
Excellent information. Thank you for the video.
Great video as always. Thanks!
Thank you. It is very clear and useful.
This lecture is so helpful and so easy to follow thanks so much!
Thanks a lot !!!!! this video is very very very helpful!!!!!
Thanks Aidan. Other applications are items like filters (gas or liquid) and exhaust baffles.
Yes!
your explanations is very useful and great. thank you so much
thank you for your brief explanation endeed..your presentation has always been so helpfull.and detailed ...but, isnt the unit of (c1/t=kg/m3s)?
Hmmm I think I might have made a small typo here. Let me get back to you on this one
muchas gracias por la explicación
Thank you very much Doctor Aidan for your excelent explaination.I would like to ask whether it is necessary to add boundary conditions at the interface between the porous medium and the fluid region? When I simulate porous media, there is always a droplet splash phenomenon at the free surface or side wall, and then the vof analysis problems, the program can not run, I have not found the reason, want to ask the teacher what may cause this situation?
I think the definition of superficial velocity is incorrect. It is a volume averaged velocity where averaging is done over the volume containing both fluid and solid. This velocity is basically the product of porosity times the intrinsic average velocity where now the averaging is done over the fluid volume only. The velocity you mentioned in the schematic is the pore velocity or the intrinsic average velocity.
Thanks for the correction Sayuj! Im going to pin this post so everyone can see it 😄
Thank you Aidan. I like all your videos.
Is the intendic average velocity also the true velocity?
@@fluidmechanics101 Hi Aidan and Sayuj, So in method 1 where data for pressure difference vs velocity is given by suppliers, is that "velocity" : superficial velocity?
Yep, it is the volume averaged velocity
very important lesson .... thank you
Thank you for your very detailed explanation about porous domain in CFD.
Could you please handle about heat transfer in porous domain also?
I'm now using porous domain to calculate conjugate heat transfer of plate heat exchanger, I have difficulty in setting a heat transfer model.
Good suggestion, I will see what I can do
Thanks. This is easy to understand porous zone.
Thank you very much for the explanation!
No problem, glad you found it useful 😊
Excellent explanation. I have a question about C1, what is the recommended value for it in the solid direction ?
10 or 100 times the value in the fluid direction is normally fine. You just need the resistance to be higher than the fluid direction, so that the flow is encouraged to straighten and flow in the fluid direction. A small amount of flow in the solid direction is fine and makes convergence a lot easier
Nice video, thank you! I am using a porous medium to represent a spiral tube bank so the flow has a non negligible radial and circumferential velocity, how can I calculate c1 and c2 values for this?
Thank you Dr for this seance, I have questions:
1- in higher reynolds number what is the model suitable for pouros media Darcy or Darcy-Forchheimer?
2- How does the setup of pours media in ansys fluent?
Thank you very much sir.
Please, how can i calculate these factors when ther is a melting of a phase change material (wax for example). So there is no flow. Just the solid wax melt inside the closed foa(porous media) when the temperature rises.
Thank you before.
Solidification and melting is very hard to do in CFD. Unfortunately I don't think I can help you. You might need specialist software for those calculations
Thank you for this video. You really helped me. But I have a doubt in slide No. 16… when you are talking about the units. I can see the permeability unit is m^2 but what happened with the unit of seconds of the dynamic viscosity?
Hi! I found the same doubt, the units of pressure and C1/mu are both incorrect. Dimension of pressure shall be [M][L]^(-1)[T]^(-2), and dimensional of C1/t shall be [M][L]^(-3)[T]^(-1). However, the resulting dimension of alpha is correct as [L]^(2), LOL.
Hi, sir thank you so much for such a clear and lucid explanation about porous media. I am calculating the porous coefficients for a thin perforated flow in the turbulent regime. While calculating the porous constants from polynomial curve fittings, I am getting viscous porous resistance, C1 as negative. Can you please explain such scenarios are possible and the physics behind the same.
Dear Mr. Wimshurst,
I appreciate your work, it's a big help for me. All your videos are very helpful. Can I have a question: are you planning to make a video about FAN boundary conditions (in OpenFoam etc.)?
Thank you :-)
Thank you for this video it is really helpful. I just have one question, since we know the pressure drop across the porous media, during simulation can i convert my 3D geometry into a 2D geometry and while putting in coefficients divide by the actual length of porous media.?
Thanks for your video. It is very useful to understand how to porous media is work. I want to model blood clot as a porous media but some paper publish only porosity and permeability values. How can I get C2 (head loss coefficient) without using velocity-pressure drop data?
Thanks a lot. However, the pressure drop data is mostly available with respect to flow rate and not versus velocity in industry so curve fitting of DP versus flow rate to get the coefficients remain valid ? thanks
If you think carefully about the flow area, then you can probably do the conversion to velocity
Great job, thanks a lot.
I am trying to model a capillary membrane, can I use this model? Do u have any tips to calculate the c1,c2 if the porous media is as fine as a membrane (pore size is 7 µm)
The flow is probably laminar through the holes (you can check Re). If it is laminar then you can use the analytical solution for flow through a pipe / parallel sides channel to derive the coefficients C1 and C2. In pretty sure if it is laminar then you will find that C2 = 0, and the C1 will depend on the viscosity and pore size
@8:27 you said that velocity will be increased due to reduction in area but as you considered before a cross section of equal area as that of pore then there will be no increase in velocity ,so what about superficial velocity then
Thank you very much Doctor! Do you have references for the theory?
Thanks for this lecture and your clear art of explanation!
One question, why you choose to set C_2 as big number and not C_1 to describe wall or very high flow resistance in particular direction?
If you just want to block the flow, you could do either C1 or C2 👍
@@fluidmechanics101 thank you for answering. I‘m struggling now with one problem connected to this lecture. I‘m modeling two-phase flow (h2o + h2) through porous media with Eulerian approach, capillary pressure and relative permeability in Fluent. Maybe you have some lecture or can suggest some good material dealing with similar issues and explains the theory behind it.
Thanks & Kind Regards
Excellent explanation. thank you very much
I see in many cases that people say we use the porous medium to define a geometry where there is no flow actually. Sir, do you know how we define a porous medium that does not allow any flow through it?
Yep, just set the resistance coefficients to be high. This will block most of the flow. Values of around 1000 seem to work quite well. If you set them too high the flow will stop completely and the solver may diverge, so be careful!
Thank you very much Doctor Aidan
the porous media can be used for tube bundle. please , How I can used porous media with tube bundle using ANSYS, or what is the equations of C1/t and C2/t
thank you very much
Hayder kraidi Rashid
You will need to look up empirical equations for the pressure drop (Euler number) for arrays of horizontal cylinders. Then you can equate these equations to the formulae for the source terms in Fluent and rearrange for the coefficients C1 and C2
Hello, this video was very useful for me thank you very much
Awesome Video, Thanks!
In the video you mention that coefficients can be found in some textbooks. Could ypu specify which textbooks? I am particulary interested in heat exchangers. Thank you!
‘Internal Flow Systems’ by Miller is a really good source for loss coefficients. Sorry i forgot to mention it!
Quick question. When you were talking about C1/t being equal to dynamic viscosity/permeability it should be in kg/(m^3*s) should´t it? Because dynamic viscosity unit is kg/(m*s) and permeability is m^2. So when you divide it it should get kg/(m^2*s) or I am missing something? Because when I use polynomial formula in my pressure drop data the units check up because of delta_pressure=C1*U+C2*0,5*density*U^2 you need to get Pa = kg/(m/s^2). Correct me please if I´m wrong. :) Anyway thank you for great contect it was really helpful.
so useful, THANK YOU SO MUCH
ı wathced the lesson carefully. thanks a lot. Do these coefficients change for heterogeneous and homogeneous porous regions?It does not change when it is calculated as an area. but i think it can change physically. my case study is airflow through people who are standing cramped up so heterogeneous permeability.
I suppose you could either average the people and do a homogeneous porous media or break your porous media into regions and assign different resistance coefficients for each region? (So effectively heterogeneous)
Hi Dr.Aidan.
In a car company, We use one set of the viscous and inertial coefficients for the pressure drop across the porous media(heat exchange), instead of three set of coefficients.
that means air will flow in all direction, which is not correct behaviors, we expect air will only flow in one direction.
Do you know why?
Nobody knows in our company, we bought CFD methods.
Hi there! If you are expecting flow in a single direction, then you can make the loss coefficients much larger in the other directions. This will encourage the flow to go in the direction you want and give the correct loss coefficients. Remember: the flow will always try and follow the path of least resistance 👍
Excellent thank you so much for your time
What about C1x & C1y? What should be the value as there will be high resistances in those directions as well
Yes, you can usually just pick a very high resistance coefficient and that is enough to stop the flow in those directions. Around 5000 should do it 👍
Hi.. Your lectures are very easy to understand.. Can you please make a video on conjugate heat transfer...
Good idea! I will add it to my list
One suggestion . Can u present the topic regarding the modelling volumetric heat source within the components .The reason for asking this is it follows the same approach as the porous medium (Defining the source terms and how this source term works in defining the temperature) .Would highly appreciate .Thanks in advance .
What kind of heat source are you thinking? Im pretty sure you just take the actual heat source (in Watts) then divide by the volume of the cell zone (to get Watts/m3). You might need to use a UDF or User-Fortran or fvOptions if your source is more complicated
Thanks, super helpful! Just to clarify, you should mesh your porous zone as 1 cell thick in the porous direction, so when you divide your Coeff.s by t, it would come out right, correct? or if you have n-cells in the flow direction, you need to divide t by n?
Just divide by t. The source terms are applied per unit volume so they will multiply by the volume of the cells. More cells just distributes the source over more cells
thank you so much professor,
How to input the values of relative humidity (moisture content) as inlet condition and check the moisture content in the porous zone? Please tell us how to do this.
Hello Aidan, excellent presentation again, congratulations. Would you say we ought to use a porous zone method if the viscous forces are dominant in the flow? Im just trying to figure if it would be possible to simplfy even further the computational cost when one deals with a porous zone and choses to deal with the problem using Darcy law (first order velocity equation) instead of the NS equation... Best regards
I suppose it depends on your geometry. Adding an extra source term to NS doesnt really increase the computational cost (adding a porous zone). So if your overall geometry is simple, you could probably just use Darcys Law in a hand calculation. If your overall geometry is complex, just throw in the porous zone 😊
I have a question. the velocity we use in pressure drop equation to calculate C1 and C2 is physical velocity or superficial velocity in porous medium?
Superficial 👍
Kindly do detail video for MSMD models
Thank you for a great explanation of porous models! Saves a ton of time having to go through lots of notes to understand it. I can see that the correction about superficial velocity explanation is already mentioned in the comments. Can you please explain the difference between superficial velocity and true velocity?
Really very very helpful.....
Super useful stuff, thank you for all of this! :D
I would have a look and see if you can find any papers or textbooks with loss coefficients for heat exchangers. I cant recommend any off the top of my head but you should be able to find some with a quick google search
Sir, Your videos are very good and easy to understand. Will you please upload a video on multiphase flow and free surface flow.
Yep, I am working on this one. It all really depends on the interface sharpening algorithm used for the convection term (have a look at my Eulerian multiphase video for a brief overview)! Interface sharpening is quite tricky, so I am still working on it 😄
How can we double the porosity value for the current case with respect to governing equations. Means which term in the governing equation will tell me about the porosity value?
Hello,
I have been looking into modeling porous domains in regards to grain storage and was wondering if you had any thoughts on this. Looking at the literature, I have found pressure drops of (say corn) across a meter bed. Using the curve fitting method, I have found values of the linear term to be in the order of 10+ Million and the quadratic term to be in the order of 10,000. I would be happy to share the google spread sheet with you if you want to see it.
Thank you.
From your values it looks like you can just neglect the viscous term and go with just the quadratic term. 10,000 is high so maybe do a quick easy CFD simulation with flow in a cuboid shaped domain with a block of porous material and check that you are getting the pressure drop you expect. If that works and you are happy then go ahead with your full CFD calculation
@@fluidmechanics101 Went back and redid the papers test in cfd, and the values are correct. I ended up using simpleFoam with different relaxation factors (0.2 for pressure and 0.3 for velocity and turbulence values) and everything is working better and, more importantly, converging. Thank you.
Amazing! Nice work!
U are a legend mate!!
Very nice lecture
Shouldn’t the units be m^2/s in slide 16?
Ive just been checking through and it looks like ive made a mistake 😅 especially after insisting on being careful with units! C1/t should have units of [kg/m3 s]. Hence the time unit in the numerator should cancel out to give the correct units of [m2] for alpha (the face permeability). Sorry for the slip up guys, and well spotted on the mistake!
Thanks a lot for your excellent videos, really helpfull ! I have a question regarding method 1 for deriving c1 and c2 (Curve fitting on experimental data). Whic velocity is considered as x-axis ? Local velocity through the obstacle or velocity just upstream the obstacle ?
Velocity just upstream (so based on the superficial velocity)
@@fluidmechanics101 thanks a lot 👍
can porous zones be used to model a membrane like the one used in water treatment like reverse osmosis or membrane distillation?
I love u man, great video
Thank you very much! One more question about the relationship between the volume porosity and the pressure loss through the porous domain. i am using the CFX to simulate the flow through a gauze. And I've ran lots of cases and found that the total pressure loss not only determined by the loss coefficient but also the volume porosity as well. can you explain this a little bit?
fantastic video , thank you again.
Hi, thank you for Your great content. I wanted to ask if it would be possible to model a metal foam for heat transfer in natural and forced convection like this. :) Thank you.
sir can you make videos on discreption of turbulent flow navier stokes equation
You can find a good description of the discretisation of N-S in my fundamentals course (which is on my website: www.fluidmechanics101.com ) or check out my video titled ‘The Finite Volume Method’ 👍
Thank u sir.
Thanks for the very nice lecture!
One question:
If a porous media is made of entangled fibers and the liquid is wicked into it by capillary force, how the C values should be calculated there?
(Do we need a microscopic image to measure the average size of microchannels between the fibers, or what)
Do you have any measurements of the pressure drop?
Fluid Mechanics 101 We have a graph for wicking length vs. time. Can we use the wicking rate to determine C value?
Sorry Hossein, i dont really know what a wicking length is? Maybe ask your colleagues
Thank you so much for this wonderful videos! And I still have a question: How about the energy equation of fluid in the Porous zones? Some papers seems only considered the heat conduction between fluid & solid porous layers, while Ansys Fluent seems used different method, and I'm still quite confused about this.
How does somebody get familiar with CFD? Mainly I am talking about the programming. I am comfortable using python, MATLAB and very proficient with calculus as well as vector calculus. Any advice would be extremely appreciated. However, I have no experience with CAD.
I can recommend you this course from Boston University: th-cam.com/video/35unQgSaT88/w-d-xo.html with additional material on this website: lorenabarba.com/blog/cfd-python-12-steps-to-navier-stokes/
Or you could just jump in and try some OpenFOAM 😄 all the source code is there and there are plenty of test cases, so you wont need to do any CAD/meshing if you just want to have a go and see how the code works. (Just google OpenFOAM and you can find it)
Hello sir, if I have open channel flow with gravel bed can I simulate it as a porous zone and is thickness of gravel bed can be consider the thickness for porose media (t )
Yep! You sure can
Hello professeur
In the porous zone is the value of the source terms in the solid or fluid (in fluent )?
Your cell zones should be a fluid. Then you need to tick the box 'porous zone'. This will open up the options which allow you to fill in the coefficients for the zone
@@fluidmechanics101 think you professeur
Thanks a lot !!!!! this video is very helpful
I have a cooling circuit including multiple components like motor, inverter, compressor etc. I want to model the circuit in Fluent while incorporating the pressure drop of each component using porus media. However i am not sure how to take care about the thickness as each component has its cold plate and i do not know their dimensions. Is this a correct approach i am adopting ? thanks
Do you know the pressure drop from each component? Or do you have any data / can approximate it? You can use an incorrect thickness as long as you ensure that the pressure drop is the same by adjusting the loss coefficient
Hello, can you please recommend some text book for coeff. calculation?
Yes, have a look at ‘internal flow systems’ by D.S.Miller. I think there are pdfs online, so you probably dont even need to buy the book 👍
Any specific reason to write modulus of velocity (IUI) in the inertial term of pressure drop equation ?
Yep, to make sure that the pressure drop is in the opposite direction to the velocity vector, regardless of coordinate system (- U |U|) will always have a sign that is the opposite of U)
Thanks very much. Very informative.
How would the tube banks work? Is this a heat exchanger?
Yes, it is a type of heat exchanger. If you think of a cross-sectional slice through the tube banks, then you can model this as a porous zone. The reason that you might not want to model the tubes explicitly is that the flow separation off the cylinders is complex, unsteady and requires a high level of grid (and temporal) resolution. So it is often easier to model a set of cylinders as a porous zone instead. This also allows you to perform a steady simulation 😊
@@fluidmechanics101 Thanks for the feedback. Extra layers on all the tubes would be excessive and vortex street for each one. My other question was how to handle something that doesn't just hold up the flow like a filter but something that redirects flow like a louvre or parallel plates at an angle to the incoming flow? Another great video and explanation all the best for 2020.
Sir...really helpful...thanks...I have to model an earthen dam by using ANSYS...can it possible...have you any related vedios?
What are you trying to simulate? A dam
break or the hydrostatic load?
Hi, Can you please recommend me a book where I can get empirical data for C2?
‘Internal Flow Systems’ by D.S.Miller has lots of empirical data. I normally use that 👍
@@fluidmechanics101 Thank you! Your video was very helpful!
Hello,
Great explanation.
Is it possible to use this method for the simulation of flow through swirl diffuser? I am doing the simulation of airflow in a clean room which has diffuser inlets.
Thanks
Yep, you can use it for the filters on the inlet and outlets to the clean room. Try and get some measures data for the pressure drop across the filters (usually the manufacturer will have this)
@@fluidmechanics101 how can the porous zone discharge with the swirl flow won't it discharge perpendicular. Same concerns for a louvre. How do you make sure it has the desired polarizing effect in terms of flow.
Hi! Thanks for the amazing video, it' very interesting. I have one question about how the solver takes in account the of porous media. Does the volume of fluid recognize there is a solid medium in the domain? So for example, if I have a full region of porosity with a percentage of void area of 50%, will the fluid be able to fill only half of the region?
ANSYS fluent doesn't recognize any solid domain. It only applies the momentum sink term once you activate the porous media model.
what about for the situation that Re is large and the flow is compressible?
Good question. It's not something I have looked into. I am guessing that there may be additional terms needed to account for the increased losses
Absolute legend
Thanks a lot for this video, it helped me understand where to start learning about the porous zone.
I have to model a capillary-membrane, whose wall is permeable and the water enters from the right side.
Hier one has a laminar flow, which means Darcy's law is accepted. I am trying to find out c1, c2, could you please support me with some tipps?
best wishes,
Venesa
thank you...
Dear Dr.Aidan, When can we expect Fundamentals 3, on Udemy.
And In course 2, section 4 wall functions topic was explained quite fast😢..I have to go through 2 more times to understand I think🙂. Apart from that, Both courses and materials provided were Great. Thank you.
Thats really useful feedback, thanks Nikhil. It is quite a difficult subject and maybe i get ahead of myself making things too concise! I have just started fundamentals 3, so its going to be a while before i finish it. I want to make sure it is really good 😄 the course is finally going to look at unstructured meshes and how we think about putting together cfd codes and solvers. Its going to be fantastic, i know you will love it. However, it does take quite a while for me to make, as the content is original and doesnt exist anywhere on the internet, so i have to be extra careful with checking everything. Im going to aim for December time to get it out to you 😄
@@fluidmechanics101 So, then it's for the Christmas 🎅. I will be waiting for it...Thanks again, for the quick reply...
Can porous zones be applied to filter media in CFD?
Pretty sure they can if you know the pressure drop across the filter media
Thank you ! Very interesting :-)
thanks for your eplanation
thank you very much for this lecture
I have a very fine grid in my geom so i decided to instead it with filtering region
but i dont have the chart to find "C2"
the area's ratio is equal 0.00101085 !!
do you think its work with this??
thank you
It will still work. The pressure drop will be quite high though ...
@@fluidmechanics101 yes it is.... in the last week i work in method that run the grid model and take the pressure drop before and after it and calculate the velocity then draw the graph .....but also i had very large error between the two pressure drops ....is the method i used work here ??