Hi, I just briefly checked the boundary conditions of your case file. I suggest you may try to use a Totalpressure condition for p at the inlet as well as Totaltemperature for T. I wouldn’t specify U. I’d rather use zero gradient or maybe a pressureinletoutlet condition. If the Mach Number at the throat of the nozzle is equal to one, which should be the case, then the mass flow and thus the velocity at the inlet is not independent of the Totalpressure and Totaltemperature. Therefore you shouldn’t fix it, instead let the solver figure it out. Otherwise you overspecify the problem and numerical errors, which differentiate your solution from the analytical one may lead to huge pain in the ass problems and unphysical solutions. I am currently working on a similar case. I am quite new to supersonic flow simulations and I am still learning, so there is no guarantee that this helps or works as expected, but it could be a more sensible solution for your boundary conditions. If I find the time I will try your case and check it out. ✌️
It is possible to avoid a negative temperature at the inlet by imposing in the system folder a file named fvOptions which limits the T. You can state the lower bound, let's say 200K and an upper bound 4000K for istance. In this way you should avoid any negative Temperature. For the pressure reflecting at the end of the domain, it could be dut to a too low length of the domain or the parameter lInf.
Thank you for the tutorial sir. I have a question about one topic that you have mentioned. In 0:54 you were saying "when i had no inlet velocity, openfoam would crash warning". Currently, I am trying to analyse nozzle also and i only want to give pressure inlet value and i do not want to give velocity inlet value at the nozzle inlet. I try to use Hisa as a solver. From the research I made, I guess I have to use custom bc but have no clue. Do u have any suggestions? Best regards.
Fantastic video, what scale is the nozzle compared to the real deal? I was going to use your coordinate points file to roughly CAD a geometry of the DLR RETPRO geometry (essentially a falcon 9) to look at aerodynamics of retropropulsion. Thanks!!
Hi, just to ask, so it’s unlikely to convert Fluent mesh (one-cell-thick block mesh) in open foam as axisymmetric mesh (pie shape) am I right? I’ve done research and came out w the utility makeAxialMesh but it seems not able to use in openfoam? Correct me if I’m wrong.
Hello, I am working on a nozzle model with a serpentine shape. I was wondering why your domain starts from the throat of the nozzle, as I believe it should start at the throat exit. Not sure I understand this. Can you clarify?
@@interfluo6420 This code assumes straight characteristics line coming out of sharp corner and intersecting the axis. But in the kernel zone characteristic lines are intersecting hence they will not remain straight. Look axisymmetric nozzle design code, which will have characteristic lines as curvature.
Also using this code will not give correct area ratio for a given Mach number. You should really look for axisymmetric method of characteristics nozzle design code.
The procedure should be fairly similar to the one outlined in the video, you will need a different mesh and boundary conditions. Depending on your mesh resolution, it may be an extremely computationally expensive simulation to run so you may need to run it on aws or other cloud hpc service.
My latest video is a cfmesh tutorial which you can try to adapt to a nozzle case. I will consider doing a nozzle case with cfmesh or snappyhexmesh in the future, but I have a couple of projects I am working on first. Good luck!
You can also use gmsh, which is open source. OpenFOAM has a mesh translator called gmshToFoam. There is a good example of the workflow on the OpenFOAM wiki.
Hello, Thank you for making this great work. I have some question about this project. How can I calculate epsilon (turbulence kinetic energy for this project)? Initially every where it was selected 266000. When I run I get negative T0 error, How much should I decrease Delta T in order to avoid this error? Thank you and Best Regards.
Great tutorial. Thank you very much. Before the end you have mentioned to convert meshing physics to cfd,please tell how to do it. Also while running openfoam for my conditions I am getting time continuity error, is there any solution to avoid it? Thanks
I would recommend watching part 2, it may become more clear. As far as the time continuity error I am not sure without the full screenshot; I recommend posting your concern to an online cfd forum or group, here are a few that may be useful: www.reddit.com/r/CFD/ www.cfd-online.com/Wiki/Fluid_dynamics grabcad.com/groups/unofficial-openfoam-community
Hello Interflow, great work! the reflection you encountered originates from using Linear div schemes. Using upwind will help. One question: I watched some video on test-firing of Merlin engine. the flame looks narrower than this. Any idea why it happened?
According to some preliminary reading I have done in the forums, it has been done with rhocentralfoam although I suspect an external solver would need to be used.
Likely yes, it may be down to a couple of different factors. The two biggest culprits in my opinion are: 1. As someone previously commented the code for MOC has some errors; I found this code online and assumed it was correct but it may be wrong in some areas 2. I specified the nozzle to be designed for an altitude of 3000m, the expansion ration will change based on what altitude you chose. For instance if you want to design a vacuum optimized nozzle the expansion ration will be larger than a sea level optimized nozzle. Hope this helps!
Hi, I just briefly checked the boundary conditions of your case file. I suggest you may try to use a Totalpressure condition for p at the inlet as well as Totaltemperature for T. I wouldn’t specify U. I’d rather use zero gradient or maybe a pressureinletoutlet condition. If the Mach Number at the throat of the nozzle is equal to one, which should be the case, then the mass flow and thus the velocity at the inlet is not independent of the Totalpressure and Totaltemperature. Therefore you shouldn’t fix it, instead let the solver figure it out. Otherwise you overspecify the problem and numerical errors, which differentiate your solution from the analytical one may lead to huge pain in the ass problems and unphysical solutions. I am currently working on a similar case. I am quite new to supersonic flow simulations and I am still learning, so there is no guarantee that this helps or works as expected, but it could be a more sensible solution for your boundary conditions. If I find the time I will try your case and check it out.
✌️
please do CFD analysis for surfing!!
It is possible to avoid a negative temperature at the inlet by imposing in the system folder a file named fvOptions which limits the T. You can state the lower bound, let's say 200K and an upper bound 4000K for istance. In this way you should avoid any negative Temperature. For the pressure reflecting at the end of the domain, it could be dut to a too low length of the domain or the parameter lInf.
Thanks for the tip, it is greatly appreciated
Awesome! Please do more videos on vehicle aerodynamics and reentry!
thank you for Computational Fluid Dynamics tutorial, please more if you have time
More to come!
Which effects are kept hidden if you had to simulate just the half of the domain?
Thank you for the tutorial sir.
I have a question about one topic that you have mentioned.
In 0:54 you were saying "when i had no inlet velocity, openfoam would crash warning". Currently, I am trying to analyse nozzle also and i only want to give pressure inlet value and i do not want to give velocity inlet value at the nozzle inlet.
I try to use Hisa as a solver. From the research I made, I guess I have to use custom bc but have no clue. Do u have any suggestions?
Best regards.
Fantastic video, what scale is the nozzle compared to the real deal? I was going to use your coordinate points file to roughly CAD a geometry of the DLR RETPRO geometry (essentially a falcon 9) to look at aerodynamics of retropropulsion. Thanks!!
Hi, just to ask, so it’s unlikely to convert Fluent mesh (one-cell-thick block mesh) in open foam as axisymmetric mesh (pie shape) am I right? I’ve done research and came out w the utility makeAxialMesh but it seems not able to use in openfoam? Correct me if I’m wrong.
How about trying an MHD case in OpenFOAM (plasma or electric arc in a circuit breaker).
Hello, I am working on a nozzle model with a serpentine shape. I was wondering why your domain starts from the throat of the nozzle, as I believe it should start at the throat exit. Not sure I understand this. Can you clarify?
yes, it is because I am using the pressure from the combustion chamber as my inlet condition.
Hellow friend i am working nozzle too
Great! Not having ANSYS, is a demo grid available? (no email found, so asking here)
Who wrote the Matlabcode? Seems like there is something wrong if TT is the Throat Temperature. Throat Temperature would be T_1*2/(g+1)
I got it from a video by VDEngineering. I really appreciate it and I will look into it, Thanks!
@@interfluo6420 This code assumes straight characteristics line coming out of sharp corner and intersecting the axis. But in the kernel zone characteristic lines are intersecting hence they will not remain straight. Look axisymmetric nozzle design code, which will have characteristic lines as curvature.
Also using this code will not give correct area ratio for a given Mach number. You should really look for axisymmetric method of characteristics nozzle design code.
at the end you need Ansys to generate the mesh, is there a open source mesh generator to CFD (trustful) ?
No, other options are available like cfmesh, snappyhexmesh, or gmsh. I have a video on cfmesh and use snappyhexmesh in some of my other videos.
Hello sir
I have designed super sonic nozzle of 2.5 mach at exit but I am not able to do 3d analysis can u please help me in that..
The procedure should be fairly similar to the one outlined in the video, you will need a different mesh and boundary conditions. Depending on your mesh resolution, it may be an extremely computationally expensive simulation to run so you may need to run it on aws or other cloud hpc service.
@Interfluo Can you do meshing tutorial of nozzle on openfoam? The procedure you mentioned needs ansys and I can't afford.
My latest video is a cfmesh tutorial which you can try to adapt to a nozzle case. I will consider doing a nozzle case with cfmesh or snappyhexmesh in the future, but I have a couple of projects I am working on first. Good luck!
You can also use gmsh, which is open source. OpenFOAM has a mesh translator called gmshToFoam. There is a good example of the workflow on the OpenFOAM wiki.
Hello,
Thank you for making this great work.
I have some question about this project. How can I calculate epsilon (turbulence kinetic energy for this project)? Initially every where it was selected 266000.
When I run I get negative T0 error, How much should I decrease Delta T in order to avoid this error?
Thank you and Best Regards.
Great tutorial. Thank you very much. Before the end you have mentioned to convert meshing physics to cfd,please tell how to do it. Also while running openfoam for my conditions I am getting time continuity error, is there any solution to avoid it? Thanks
I would recommend watching part 2, it may become more clear. As far as the time continuity error I am not sure without the full screenshot; I recommend posting your concern to an online cfd forum or group, here are a few that may be useful:
www.reddit.com/r/CFD/
www.cfd-online.com/Wiki/Fluid_dynamics
grabcad.com/groups/unofficial-openfoam-community
Hello Interflow, great work! the reflection you encountered originates from using Linear div schemes. Using upwind will help.
One question: I watched some video on test-firing of Merlin engine. the flame looks narrower than this. Any idea why it happened?
thankx for sharing.!!!
Nice :)
Does OpenFOAM support hypersonic flows?
According to some preliminary reading I have done in the forums, it has been done with rhocentralfoam although I suspect an external solver would need to be used.
area expansion ratio is 21.4 for merlin 1-d in the table. But your design's ratio is about 5. Isn't there something wrong?
Likely yes, it may be down to a couple of different factors. The two biggest culprits in my opinion are:
1. As someone previously commented the code for MOC has some errors; I found this code online and assumed it was correct but it may be wrong in some areas
2. I specified the nozzle to be designed for an altitude of 3000m, the expansion ration will change based on what altitude you chose. For instance if you want to design a vacuum optimized nozzle the expansion ration will be larger than a sea level optimized nozzle.
Hope this helps!