Your lectures are always helpful and easy to understand. So I always recommend this channel to my colleagues. Thank you for your work and hope your channel continues to grow :)
Thanks a lot for the video, it is truly useful to people who don't really understand how CFD codes do the calculations like me. I'm looking forward to more videos like this!
I can tell you as someone who writes cfd codes this is not the whole thing. It is bit more complicated than that. The problem is that one can not really talk about what fluent and cfx would do without working on these software as a developer.
Of course. As a user the best we can do is try and deduce enough for us to get a solution. Often the user guides aren't enough, so we have to make educated guesses 😅
@@fluidmechanics101 yes it is true. Btw fluent omits lots of important details. Two examples in presto scheme they give vague idea and one can never figure is out (it needs detective work to get what they are doing) Also notice in rhie and chow term it is given as df (p1 - p0) but what is df and how its calculated is not mentioned last 15 years in their manuals. Fluent is very secretive. The whole heat transfer calculations are very complicated even without including radiation.
A great talk, I'm always looking forward to your videos. I prefer the new approach with focus on application with the theoretical background. Especially using some examples brings the message home, which helps to support the argument on different approaches.
Really loved the video, very useful to know exactly where to look at on the software and a perfect introduction for heat transfer on CFD. Also love the heat transfer on your background light :D Thank you Aidan
I always wait for your videos... Can you please make a video stating the difference between a fem solver and a fvm solver? How those 2 different solvers solve the differential equations to find the fields such as temperature velocity etc and most important thing Happy new year :)
Sir ur videos are really amazing . Most of the doubts are solved which was comes in background of cfd . Please make some videos on meshing topics also . Please sir
Really love your content thanks a lot! Something I would be very interested in would be a video about conjugate heat transfer CHT. Just as a suggestion if you are looking for new video ideas :) All the best!
Hi, Thank you for the helpful videos, Just as a remark, the reference temperature has the same definition for both internal and external flows, which you gave at time 6:08. But for external flows, this can be simplified to $T_\infty$ due to the uniform velocity distribution at the inlet of the domain. Happy new year, RAS
Another great video. Can you make a video for the following topics? 1. Favre averaged N-S compared to Reynolds averaged N-S 2. Reynolds stress transport turbulence model for anisotropic flows
I am simulating double pipe heat exchanger. Would reference temperature be the bulk mean temperature (average of inlet and outlet temperature of either hot or cold fluid side) in my case? If yes, then how we can get the outlet temperature before doing the simulation and without knowing the outlet temperature of either side how we can get the bulk mean or reference temperature? Are you saying that we have to put the reference temperature after performing simulation for post processing the results or to find the actual heat transfer coefficient and Nusselt number by using Ansys expressions and calculator in post processing? Thanks...
You need to take several cross section slices along the length of the heat exchanger, on both the hot and cold sides. For each of the slices, the mass flow averaged temperature on the slice is the reference temperature at that location. So Tref will vary and will be different on the hot and cold sides. Naturally this is quite a complicated calculation and you will need to do it as a post processing calculation
Great video Dr. Aidan, I have a question regarding the heat transfer coefficient provided by Fluent, does it account for radiation (radiative heat transfer coefficient) when a radiation model is being used, thank you!
I don't think so. I'm pretty sure (without checking the manual) that the HTC provided automatically by fluent is for convective heat transfer only. This is another reason to define your own heat transfer coefficient in the post processor. You can always be sure which contributions it contains!
Newtons law of cooling can be applied to anything. When there is a temperature difference, there will be heat transfer, and the rate is proportional to the temperature difference. If you want a reference for Newtons Law of Cooling then any heat transfer textbook will do 🙂
I referred Heat and Mass Transfer by Yunus Cengel 4e. Unfortunately I did not find Newton's Law of Cooling applied to soild-soild interaction. Let me explore more text books.
Hello Aidan, for low Reynolds number turbulence models like the k-omega SST, and considering the Prandtl number of the fluid, what is the maximum y+ allowed for the mesh in thermal CFD simulations? Thank you
If your Prandtl number is ~ 1 (for air say) then you can use the standard values that are recommended for y+, as the momentum and thermal boundary layers are similar in shape. If Pr >> 1 then the thermal boundary layer is thinner than the momentum boundary layer, so you have to be careful, as you have a single mesh but the first cell height will be in a different region of the thermal boundary layer and the momentum boundary layer. It is quite case specific (depending on the specific fluid you are working with), so what fluid are you working with and what is its Pr?
Hello Aidan, thanks a lot for the answer! Taking into account that in the viscous sublayer the formula for thermal wall function is T* = Pr . y* for y* < 5 and for momentum is U+ = y+ for y+ < 5, and y+ and y* are essentially the same numbers but calculated in different ways (depending on friction velocity formula), if the mesh is resolved for momentum (obtaining y+ < 5) it will be automatically resolved for heat transfer too (only because of y* ~ y+ and the formula for viscous sublayer thermal wall function), despite the fact that different boundary layer thicknesses (momentum and thermal) arise from a Prandtl number different than one. If in the formula T* = Pr . y* one makes T* = y_mod, with y_mod = Pr . y* , so that y_mod < 5 for a resolved thermal boundary layer, it would be in agreement with the Prandtl number physical meaning (the ratio of boundary layer thicknesses), and if the fluid has Pr > 1 , the value of y cell center to get y+ < 5 in the momentum boundary layer will need to be divided by the Prandtl number in order to get y_mod < 5 in the thermal boundary layer. What are your thoughts about it? Thanks and regards
Hi Aidan, I am using StarCCM+ right now. As you may know, it offers 4 types of heat transfer coefficients to choose from, Local HTC, Specified y+ HTC, HTC, and Virtual HTC. Can you help me in deciding which one should be used for a problem of thin channel flow as it may have different Tref. Again, I know what these HTCs mean but which one would be best for comparing with tests, is confusing me. Thanks.
Dear Aidan, Quick question, are we agree that the HTC is not used by the CFD code to solve the case simulated ? I mean, the temperature field is calculated first and the HTC is calculated afterwards to be compared with the reality. Correct ? Thank you very much for your help and for this excellent video as usual.
As you have an internal flow, you want the mass flow average of the flow through the passage you are looking at. Regardless of what you choose for Tref, make sure you make it clear what value you have used (so that the reader can back calculate the heat flux if they want to)
You can normally do the integral in the post processor of your CFD code. You want a 'mass weighted integral'. Some post processors have this option. If not you can create a plane, calculate the mass flow rate on the plane, multiply this by temperature on the plane and then integrate over the plane
Do you know about any study that compares the heat transfer coefficient using the Nusselt equations (empirical equations) and the heat transfer equation (Q/DT)?
Do I need to set reference values in the case of double pipe heat exchangers? If yes, which zone should I compute from? and Which reference zone will I select?
This will be a tricky one as fluent only lets you specify one reference value. I would do the post processing yourself manually and define different reference values for each of the fluids
@@fluidmechanics101 I understand. Thanks a lot for your advice. I also have to say, your content is highly beneficial! Always great to watch your videos and the breakdown of Fluid Mechanics. Thank you.
Hey, at my company to calculate the heat transfer coefficients we run 2 simulations. 1 adiabatic and 1 wall temp. We use the adiabatic case to find the T_ref at all cells. Then export the results and do the calculations in a spreadsheet for each cell. How does this compare to the other methods to calculate T_ref? Is it more accurate?
Ah yes, I have recently found out about 'the adiabatic wall method' but I don't have much experience myself. Neither method is necessarily more accurate as both are just post processing operations to calculate the HTC. The value you actually care about is the heat flux and the wall temperature. As long as your CFD code is getting these correct then you are fine 👍
3:29 You defined the terms strangely: Heat: energy Heat Flux: energy per unit time (i.e. power) Heat Flux per unit area: power per unit area This is the standard convention: Heat: energy Heat Flow Rate: energy per unit time (i.e. power) Heat Flux: power per unit area Following the standard convention, "heat flux per unit area" is superfluous.
In mass flow averaged Htc, there is velocity and Temperature term. Is this the free stream velocity & temperature ? or velocity and temperature of cell centroid next to wall ? or of wall (no slip has vel wall = 0) ?
Hi there! Absolutely great video! The question of how to correctly get the HTC in complex applications, as you mentioned, is still quite open. I've been in touch with Ansys and they also agree a robust post-processing methodology to estimate HTC is pretty hard to get... Have you already thought about some kind of automatic boundary layer post-processing routine to estimate local bulk temperature to be used as Tref?
Yes 😄 a trick I use is to write a fluent journal file that writes out a series of line plots (all normal to the wall). Then write a MATLAB or python script to read in these data files and calculate the HTC for each line plot. Then you can plot HTC as a function of distance along the wall 👍 this works quite nicely for internal flows as well as you can calculate the friction factor and Tref (by averaging the line plot) as well
Hello! Great explanation! I think it is really insightful to refer to particular cfd codes as in this example! Could you please answer the following question: If we have a multi-stage axial compressor or turbine, do we have to specify ref. temperature for each and every stator/rotor domain ? Will we have to wright a UDF or CEL expression for that application ? Thank you in advance!
@@fluidmechanics101 Thank you for your reply! Could you please suggest me any resource to study that could help me right CEL expression in CFX for that application? Can I specify an expression rather than a value to tbulk (expert parameters)? Will I have to write expressions in CFD-Post or CFX-Pre and resolve ?
Hi Jon, if you have already run your case I would try and do it in CFX post (rerunning your case is expensive, so you should try and avoid it unless you absolutely have to)! I would probably sample / divide up the surfaces of interest in CFX post, export them and then calculate the HTC in MATLAB / python. This gives you more control and you can check it along the way as you do your calcs
@@fluidmechanics101 Really insightful! I don't know how to extract q (heat fux per unit area) because in my mind is like a vector field variable and not a scalar one. Could you please share your opinion about that! I think many CFD Engineers are struggling when it comes to HTC calculations. It's quite interesting to clarify that! Thank you for your response! Already , you have helped me a lot!
The heat flux per unit area should be available as a variable in CFX post. It is a scalar and you should be able to find it by checking the units of the variables 😄
Thank you for your work, Aidan! Here is a question: can I find out the nearest point to the target wall at which the temperature gradient along normal direction is 0, and then define the temperature at that point as the reference temperature?
I can not agree to the statement around @15:25 that the h value can not be used for a comparison. The reason why we define T+ is to eliminate the mesh dependence of h. Tref is mesh dependent but in the end, the h is calculated in such a way through the usage of T+ that, this dependence of Tref on the mesh is not transmitted to h value. Therefore, h is calculated independent of mesh and can be used. From the other way around, who would need to calculate a useless h value and why?
Hi sir all your videos are superb. realy i thank for that sir, i have doubt, For heat flux values, how ansys fluent will calculate if the geometry is 2d? ( for example 2d pipe flow, here, wall is an edge only not area)
You can use that one temperature measurement that you have. Just make sure to state the value of Tref that you are using in your report / document.That way, others can compare to your results if they want to
very nice talk on HTC and Reference Temperature. Can you guide me for calculation of HTC in case of heat transfer through fins of an electric motor, Please?
Please help. I have a pipe with steam in it. I would like to estimate the heat transfer coefficient between the wall and the fluid (using Fluent ANSYS), how I can set the boundary conditions within FLUENT. If I put convection, I need to put the heat transfer coefficient that I am looking for? What should I use for boundary conditions? Thanks million!
You need a fixed temperature boundary condition not a convective boundary condition. The convective boundary condition is applied if you have convection on the outside of your domain 👍
@@fluidmechanics101 But, I do not know the wall temperature and I know there is a convection between the surface and the steam. If I put the Tref as the fluid temperature and the surface temperature as the fluid temperature, the difference will be zero? Please, explain?
You can do this with a heat flux boundary condition. The CFD code will calculate the wall temperature for you. The reference temperature is the mass flow average temperature for the pipe cross section. The HTC is then: q / Twall - Tref (where q is the wall heat flux you applied as a boundary condition)
Just a general question: Regarding this equation Q = h (Tw-Tref). How much input we need to give to ANSYS to determine the rest. For example, if I put only the Tref, is that enough? or I need to put two pieces of information: Tref and h or Tref and Q, or Tref and Tw?
The turbulent kinetic energy on the face of the cell might be zero or very small, but the value at the cell centroid will be non zero 😄 the trick is to notice the difference between the face of the cell and the cell centroid 👍
@@fluidmechanics101 thx for the quick reply. As I understand for a y+ value of e.g. 1 one may have more than one mesh layer in the viscous sublayer. The entire viscous sublayer is laminar and in a laminar flow the TKE is zero since the velocity fluctuations are zero. If the TKE values on the cell face on the wall and the cell face towards the 2nd layer are both zero than the centroid value should be zero too, right?
Hello sir In open foam I wanted a developed flow at my jet inlet so i simulated channel flow and i want it's outlet as my jet inlet so sir i did use mapfield dict but it's didn't work it's putting half of inlet only
Hmmm I can't debug your problem for you but I can suggest maybe double checking the coordinates and patch names? Maybe try with a simpler problem and check you can get it to work. These things can be tricky with OpenFOAM
@@ashutoshsingh-et7vm The easiest way is to just consider the inlet of your domain coordinate, then in a separate simulations with the exact patch of your inlet extend it into 2 cells only with the same mesh density. Apply cyclic flow conditions, then map the field! If u couldn't do it then u need to share your case
![[CFD] Convection (Heat Transfer Coefficient) Boundary Conditions](http://i.ytimg.com/vi/ygR31_vhvJI/mqdefault.jpg)
![[CFD] Convection (Heat Transfer Coefficient) Boundary Conditions](/img/tr.png)
![[CFD] How Fine should my CFD mesh be?](http://i.ytimg.com/vi/60fDz2cVdy8/mqdefault.jpg)
![[CFD] The Energy Equation for Solids and Fluids in CFD](/img/n.gif)
Very nice talk! Concise, precise, and to the point! Thank you!!!
Your lectures are always helpful and easy to understand. So I always recommend this channel to my colleagues. Thank you for your work and hope your channel continues to grow :)
we want more talk in this style. Your talks are always interesting. Thank you so much
literally doing this for a project!
I hope that you will continue to educate in some form since you have a real talent.
Very clearly explained. Your videos r crisp n very focussed. Thanks n keep making 👍
That was very useful for me personally. I needed this.
Keep up the great work!
Thanks a lot for the video, it is truly useful to people who don't really understand how CFD codes do the calculations like me. I'm looking forward to more videos like this!
I can tell you as someone who writes cfd codes this is not the whole thing.
It is bit more complicated than that.
The problem is that one can not really talk about what fluent and cfx would do without working on these software as a developer.
Of course. As a user the best we can do is try and deduce enough for us to get a solution. Often the user guides aren't enough, so we have to make educated guesses 😅
@@fluidmechanics101 yes it is true.
Btw fluent omits lots of important details.
Two examples in presto scheme they give vague idea and one can never figure is out (it needs detective work to get what they are doing)
Also notice in rhie and chow term it is given as df (p1 - p0) but what is df and how its calculated is not mentioned last 15 years in their manuals.
Fluent is very secretive.
The whole heat transfer calculations are very complicated even without including radiation.
Great content with very organized organized explanation. Thanks a lot.
Great video. Excellent presentation and explanation. Thanks!
You saved my life another time. Thanks brother ❤
I love this kind of video! I totally understood how heat transfer coefficient is calculated in ANSYS Fluent :)
Thanks. Very helpful. Managed to get to the point just looking at the slides. Great Job!
A great talk, I'm always looking forward to your videos. I prefer the new approach with focus on application with the theoretical background. Especially using some examples brings the message home, which helps to support the argument on different approaches.
Great talk. Really appreciate the content, quality and the effort that you put into producing it.
Excellent video !!! Thanks for sharing.
Great talk. Really appreciate the content,
Real helpful explanation you have there, thanks!
Thank you so much for your brilliant explanations ! Really great video
Very organized video. Thank you very much.
Really good explanation. Thank you for the content and keep going!
Really loved the video, very useful to know exactly where to look at on the software and a perfect introduction for heat transfer on CFD. Also love the heat transfer on your background light :D
Thank you Aidan
Спасибо за видео ! С Новым годом :)
I always wait for your videos... Can you please make a video stating the difference between a fem solver and a fvm solver? How those 2 different solvers solve the differential equations to find the fields such as temperature velocity etc and most important thing Happy new year :)
Thanks man. I spent last 2 hours to find the meaning
Thank you for this great video
I love the fact that you mentioned the different codes and how they operate,
Looking forward to your future works!
That’s exactly what I need, thanks!
Thank you! Very helpful overview
Thanks a lot. It is very informative and useful.
Your new trail is so useful
Very helpful! Thank you!
Thanks for the video! Well done!
Very useful, thank you.
Thank you sir for posting this excellent vedio, this style is good, cuz we are acutally able to correlate with different CFD software.
Good and very useful info.
thanks for the video =) Happy New Year!
It's really helpful. Thanks a lot!
Very impressive, thanks a lot!
Sir ur videos are really amazing . Most of the doubts are solved which was comes in background of cfd . Please make some videos on meshing topics also . Please sir
Very Informative video
Thank you so much for this talk, it will be good to have blend of theory and UI. If at all possible.
Easily explained!Thanx
Really love your content thanks a lot! Something I would be very interested in would be a video about conjugate heat transfer CHT. Just as a suggestion if you are looking for new video ideas :) All the best!
It is a good vedeo Dr, good luck.
You are awesome.....very helpful
You are THE BEST 🙌🌋.
Thank you so much you are really awesome
Hi,
Thank you for the helpful videos,
Just as a remark, the reference temperature has the same definition for both internal and external flows, which you gave at time 6:08.
But for external flows, this can be simplified to $T_\infty$ due to the uniform velocity distribution at the inlet of the domain.
Happy new year,
RAS
Thanks RAS! Happy new year to you too
Another great video.
Can you make a video for the following topics?
1. Favre averaged N-S compared to Reynolds averaged N-S
2. Reynolds stress transport turbulence model for anisotropic flows
They're on my 'to do list' 😄
I am simulating double pipe heat exchanger. Would reference temperature be the bulk mean temperature (average of inlet and outlet temperature of either hot or cold fluid side) in my case? If yes, then how we can get the outlet temperature before doing the simulation and without knowing the outlet temperature of either side how we can get the bulk mean or reference temperature? Are you saying that we have to put the reference temperature after performing simulation for post processing the results or to find the actual heat transfer coefficient and Nusselt number by using Ansys expressions and calculator in post processing? Thanks...
You need to take several cross section slices along the length of the heat exchanger, on both the hot and cold sides. For each of the slices, the mass flow averaged temperature on the slice is the reference temperature at that location. So Tref will vary and will be different on the hot and cold sides. Naturally this is quite a complicated calculation and you will need to do it as a post processing calculation
Really nice one sir but eagerly waiting for LES
Great job
Great video!
at 6:17 you have a formulla for Tref . which book is that from ? and thanks a lot for your videos
Incropera et al. 'Fundamentals of Heat and Mass Transfer'
Great video Dr. Aidan, I have a question regarding the heat transfer coefficient provided by Fluent, does it account for radiation (radiative heat transfer coefficient) when a radiation model is being used, thank you!
I don't think so. I'm pretty sure (without checking the manual) that the HTC provided automatically by fluent is for convective heat transfer only. This is another reason to define your own heat transfer coefficient in the post processor. You can always be sure which contributions it contains!
thank you!!!
At 1.26 can you please give reference to literature which says Newton's law of cooling can be applicable to solid to solid heat transfer? Thank you
Newtons law of cooling can be applied to anything. When there is a temperature difference, there will be heat transfer, and the rate is proportional to the temperature difference. If you want a reference for Newtons Law of Cooling then any heat transfer textbook will do 🙂
@@fluidmechanics101 thank you. I was under impression that HTC is calculated for fluid-solid heat transfer
Yep, it normally is. But you can calculate it for anything with a temperature difference 👍
I referred Heat and Mass Transfer by Yunus Cengel 4e. Unfortunately I did not find Newton's Law of Cooling applied to soild-soild interaction.
Let me explore more text books.
Hey, I referred few more books unfortunately could not able find answer to above question.
Hello Aidan, for low Reynolds number turbulence models like the k-omega SST, and considering the Prandtl number of the fluid, what is the maximum y+ allowed for the mesh in thermal CFD simulations? Thank you
If your Prandtl number is ~ 1 (for air say) then you can use the standard values that are recommended for y+, as the momentum and thermal boundary layers are similar in shape. If Pr >> 1 then the thermal boundary layer is thinner than the momentum boundary layer, so you have to be careful, as you have a single mesh but the first cell height will be in a different region of the thermal boundary layer and the momentum boundary layer. It is quite case specific (depending on the specific fluid you are working with), so what fluid are you working with and what is its Pr?
Hello Aidan, thanks a lot for the answer! Taking into account that in the viscous sublayer the formula for thermal wall function is T* = Pr . y* for y* < 5 and for momentum is U+ = y+ for y+ < 5, and y+ and y* are essentially the same numbers but calculated in different ways (depending on friction velocity formula), if the mesh is resolved for momentum (obtaining y+ < 5) it will be automatically resolved for heat transfer too (only because of y* ~ y+ and the formula for viscous sublayer thermal wall function), despite the fact that different boundary layer thicknesses (momentum and thermal) arise from a Prandtl number different than one. If in the formula T* = Pr . y* one makes T* = y_mod, with y_mod = Pr . y* , so that y_mod < 5 for a resolved thermal boundary layer, it would be in agreement with the Prandtl number physical meaning (the ratio of boundary layer thicknesses), and if the fluid has Pr > 1 , the value of y cell center to get y+ < 5 in the momentum boundary layer will need to be divided by the Prandtl number in order to get y_mod < 5 in the thermal boundary layer. What are your thoughts about it? Thanks and regards
Yes I like that explanation! Great logic
Hello Aidan, thanks for the compliment! But does the explanation make sense for you? Is it correct? Best regards
Hi Aidan, I am using StarCCM+ right now. As you may know, it offers 4 types of heat transfer coefficients to choose from, Local HTC, Specified y+ HTC, HTC, and Virtual HTC. Can you help me in deciding which one should be used for a problem of thin channel flow as it may have different Tref. Again, I know what these HTCs mean but which one would be best for comparing with tests, is confusing me. Thanks.
Have a look at the definition they used in the test and then use the same definition in your CFD 👍
please,
How to apply this method for a heat exchanger with a cold fluid and hot fluid separated by a thin wall
Dear Aidan, Quick question, are we agree that the HTC is not used by the CFD code to solve the case simulated ? I mean, the temperature field is calculated first and the HTC is calculated afterwards to be compared with the reality. Correct ?
Thank you very much for your help and for this excellent video as usual.
Yep correct
Great video. I have a question in connection with brake cooling htc calculation. In that case, what would be the Tref? Thanks
As you have an internal flow, you want the mass flow average of the flow through the passage you are looking at. Regardless of what you choose for Tref, make sure you make it clear what value you have used (so that the reader can back calculate the heat flux if they want to)
Hi, thank you for this video! Can I ask how I can calculate the integral for an internal flow problem in 6:13? also, what do the variables stand for?
You can normally do the integral in the post processor of your CFD code. You want a 'mass weighted integral'. Some post processors have this option. If not you can create a plane, calculate the mass flow rate on the plane, multiply this by temperature on the plane and then integrate over the plane
Do you know about any study that compares the heat transfer coefficient using the Nusselt equations (empirical equations) and the heat transfer equation (Q/DT)?
Sadly no! Maybe someone else in the comment section does?
Do I need to set reference values in the case of double pipe heat exchangers? If yes, which zone should I compute from? and Which reference zone will I select?
This will be a tricky one as fluent only lets you specify one reference value. I would do the post processing yourself manually and define different reference values for each of the fluids
@@fluidmechanics101 I understand. Thanks a lot for your advice. I also have to say, your content is highly beneficial! Always great to watch your videos and the breakdown of Fluid Mechanics. Thank you.
Hey, at my company to calculate the heat transfer coefficients we run 2 simulations. 1 adiabatic and 1 wall temp. We use the adiabatic case to find the T_ref at all cells. Then export the results and do the calculations in a spreadsheet for each cell. How does this compare to the other methods to calculate T_ref? Is it more accurate?
Ah yes, I have recently found out about 'the adiabatic wall method' but I don't have much experience myself. Neither method is necessarily more accurate as both are just post processing operations to calculate the HTC. The value you actually care about is the heat flux and the wall temperature. As long as your CFD code is getting these correct then you are fine 👍
Do you have any idea how to write HTC in e.g. cross-flow around tube and how to set T_oo for Autodesk CFD ?
3:29 You defined the terms strangely:
Heat: energy
Heat Flux: energy per unit time (i.e. power)
Heat Flux per unit area: power per unit area
This is the standard convention:
Heat: energy
Heat Flow Rate: energy per unit time (i.e. power)
Heat Flux: power per unit area
Following the standard convention, "heat flux per unit area" is superfluous.
Agreed. Thank you for the clarification
@@fluidmechanics101 Great video, by the way.
Thank you
In mass flow averaged Htc, there is velocity and Temperature term. Is this the free stream velocity & temperature ? or velocity and temperature of cell centroid next to wall ? or of wall (no slip has vel wall = 0) ?
Hello Sir! I have a doubt. How does ANSYS FLUENT calculate heat flux?
Good question. I can never really know, as I don't have access to the code. We can guess? I would ask ANSYS and see if they will tell you
Tref in fluent doesnt work with me. i dont know where may i missed somthing
Hmmm it's probably best to just calculate your HTC manually in the post processor then
Hi there! Absolutely great video! The question of how to correctly get the HTC in complex applications, as you mentioned, is still quite open. I've been in touch with Ansys and they also agree a robust post-processing methodology to estimate HTC is pretty hard to get... Have you already thought about some kind of automatic boundary layer post-processing routine to estimate local bulk temperature to be used as Tref?
Yes 😄 a trick I use is to write a fluent journal file that writes out a series of line plots (all normal to the wall). Then write a MATLAB or python script to read in these data files and calculate the HTC for each line plot. Then you can plot HTC as a function of distance along the wall 👍 this works quite nicely for internal flows as well as you can calculate the friction factor and Tref (by averaging the line plot) as well
@@fluidmechanics101 Oh that method sounds nice! Can I assume that you use the converged HTC value obtained from each line plot?
Yep, take the converged value for all the lines
The undergrounds of CFD
Thank you.
Hello! Great explanation! I think it is really insightful to refer to particular cfd codes as in this example! Could you please answer the following question: If we have a multi-stage axial compressor or turbine, do we have to specify ref. temperature for each and every stator/rotor domain ? Will we have to wright a UDF or CEL expression for that application ? Thank you in advance!
Yep the reference temperature will change as you move through each stage of the rotor / compressor, so you will need CEL or a UDF to do this for you
@@fluidmechanics101 Thank you for your reply!
Could you please suggest me any resource to study that could help me right CEL expression in CFX for that application?
Can I specify an expression rather than a value to tbulk (expert parameters)?
Will I have to write expressions in CFD-Post or CFX-Pre and resolve ?
Hi Jon, if you have already run your case I would try and do it in CFX post (rerunning your case is expensive, so you should try and avoid it unless you absolutely have to)! I would probably sample / divide up the surfaces of interest in CFX post, export them and then calculate the HTC in MATLAB / python. This gives you more control and you can check it along the way as you do your calcs
@@fluidmechanics101 Really insightful! I don't know how to extract q (heat fux per unit area) because in my mind is like a vector field variable and not a scalar one. Could you please share your opinion about that! I think many CFD Engineers are struggling when it comes to HTC calculations. It's quite interesting to clarify that! Thank you for your response! Already , you have helped me a lot!
The heat flux per unit area should be available as a variable in CFX post. It is a scalar and you should be able to find it by checking the units of the variables 😄
Thank you for your work, Aidan! Here is a question: can I find out the nearest point to the target wall at which the temperature gradient along normal direction is 0, and then define the temperature at that point as the reference temperature?
Yep that would also work!
@@fluidmechanics101 Thanks, Aidan!
I can not agree to the statement around @15:25 that the h value can not be used for a comparison. The reason why we define T+ is to eliminate the mesh dependence of h. Tref is mesh dependent but in the end, the h is calculated in such a way through the usage of T+ that, this dependence of Tref on the mesh is not transmitted to h value. Therefore, h is calculated independent of mesh and can be used.
From the other way around, who would need to calculate a useless h value and why?
Hi sir all your videos are superb. realy i thank for that sir, i have doubt, For heat flux values, how ansys fluent will calculate if the geometry is 2d? ( for example 2d pipe flow, here, wall is an edge only not area)
Good question. I would probably check the manual if I were you 🙂 I suspect they will use the edge length rather than the area
@@fluidmechanics101 thank you so much for immediate reply sir..
If you have only one temperature measured at the level of the pipe (inside the pipe), with a mass flow. How you can select the Tref?
You can use that one temperature measurement that you have. Just make sure to state the value of Tref that you are using in your report / document.That way, others can compare to your results if they want to
Can you demonstrate how to calculate tref for complex geometry using udf in fluent ?
Very nice lecture it's numerical method using matlb please
very nice talk on HTC and Reference Temperature. Can you guide me for calculation of HTC in case of heat transfer through fins of an electric motor, Please?
How to calculate the average axial Nu number using cfd
Please help. I have a pipe with steam in it. I would like to estimate the heat transfer coefficient between the wall and the fluid (using Fluent ANSYS), how I can set the boundary conditions within FLUENT. If I put convection, I need to put the heat transfer coefficient that I am looking for? What should I use for boundary conditions? Thanks million!
You need a fixed temperature boundary condition not a convective boundary condition. The convective boundary condition is applied if you have convection on the outside of your domain 👍
@@fluidmechanics101 But, I do not know the wall temperature and I know there is a convection between the surface and the steam. If I put the Tref as the fluid temperature and the surface temperature as the fluid temperature, the difference will be zero? Please, explain?
@@fluidmechanics101 If I use "coupled" isn't a better way to represent what happens between the surface of the pipe and the steam? Thanks
You can do this with a heat flux boundary condition. The CFD code will calculate the wall temperature for you. The reference temperature is the mass flow average temperature for the pipe cross section. The HTC is then: q / Twall - Tref (where q is the wall heat flux you applied as a boundary condition)
Just a general question: Regarding this equation Q = h (Tw-Tref). How much input we need to give to ANSYS to determine the rest. For example, if I put only the Tref, is that enough? or I need to put two pieces of information: Tref and h or Tref and Q, or Tref and Tw?
Thanks sir
Great Video, one question though: when using y+
The turbulent kinetic energy on the face of the cell might be zero or very small, but the value at the cell centroid will be non zero 😄 the trick is to notice the difference between the face of the cell and the cell centroid 👍
@@fluidmechanics101 thx for the quick reply. As I understand for a y+ value of e.g. 1 one may have more than one mesh layer in the viscous sublayer. The entire viscous sublayer is laminar and in a laminar flow the TKE is zero since the velocity fluctuations are zero. If the TKE values on the cell face on the wall and the cell face towards the 2nd layer are both zero than the centroid value should be zero too, right?
Yep, spot on 👍
Hello sir
In open foam I wanted a developed flow at my jet inlet so i simulated channel flow and i want it's outlet as my jet inlet so sir i did use mapfield dict but it's didn't work it's putting half of inlet only
Hmmm I can't debug your problem for you but I can suggest maybe double checking the coordinates and patch names? Maybe try with a simpler problem and check you can get it to work. These things can be tricky with OpenFOAM
@@fluidmechanics101 sir is there any other way
@@ashutoshsingh-et7vm The easiest way is to just consider the inlet of your domain coordinate, then in a separate simulations with the exact patch of your inlet extend it into 2 cells only with the same mesh density. Apply cyclic flow conditions, then map the field! If u couldn't do it then u need to share your case