CFD Results - How to analyse OpenFOAM data with ParaView - 25-minute Tutorial

[@BombermanOfComet]

I am currently writing my Bachelor-Thesis with OpenFOAM and this video is really helpful. It gets me an good overview for post processing. A tutorial well made. Thank you for the good work here :)

[@AirShaper]

Thank you, it's very kind of you to let us know the video was helpful - good luck with your Thesis!

[@obkf-too]

Finally someone explaining things the right way. Thanks for the tutorial.

[@AirShaper]

Thank you very much Oussama!

[@eduserenza]

Very clear presentation, I am thinking in start study CFD with OpenFoam, I did not know about that big capability of OpenFoam. Really is impressive!

[@AirShaper]

Thanks Eduardo, yeah OpenFOAM is really impressive - best of luck!

[@docd7779]

Clear, concise, informative and interesting - excellent.

[@AirShaper]

Thank you very much Doc!

[@IRC2005]

Beautifully clear presentation.

[@AirShaper]

Thanks a lot Glig

[@DhruvHaldar-b1f]

Thanks for this, I am in the process of validating the drivAer case (notchback) with our commercial solver.

[@MountainPassPerformance]

Wonderful, thank you!

[@thucpham4598]

Clear and informative! Thank you so much!!

[@AirShaper]

Thank you, you're most welcome!

[@MarcelNKemet]

NB: wallShearStress vector data is not included by default. You need to specify it as a function in system/controlDict.

[@francescapalanza2480]

Hello, sorry how do I do this?

[@Rahmahhusayn]

hello, nice and understandable tutorial, thank you sir. I would like to ask why we need to subtract with wallshearstress? is the P*A the total forces already? so that if we want to get the pressure force need to subtract with wallshearstress?

[@AirShaper]

The total force on a surge is the sum of the perpendicular force (pressure) and tangential force (friction. That is why you need to sum them. The minus sign is because of conventions on how to define force direction, has no specific meaning.

[@marvin1711]

Thank you so much! It really helped me with my project!

[@AirShaper]

You're most welcome!

[@AlvaroLleonartMoix-g8w]

Hi there! Perfect tutorial and explanation! I would like to know when do we know if our data is normalized or not? I've read that dimensional values are known as "no normalized data" and adimensional values are know as "normalized data". Could you tell me where I can find this information in my OpenFoam folders? Thanks!

[@ahmetkirhan5896]

Hello Sir, Thanks a lot for this tutorial. I have a question. What is the formula for Cp for compressible flow and how can we apply it to paraview?

[@saveriolecce5888]

Hi. I think your channel is very interesting. I have one question about this video. When you computed the total force acting on the vehicle, why are you subtracting instead of summing the wall shear stresses? Is that becasue you are interested only in the pressure forces rather than in the effect of both pressure and friction or I am just missing something?

[@AirShaper]

Dear Saverio, good question - I don't know! If I look purely at the friction forces and integrate them across the car, then I get a negative value in the driving direction. That is physically not possible. When I apply the minus sign, it matches perfectly with the forces we calculate in OpenFOAM. My guess is that perhaps this has something to do with the normals, which can be inside-out on the surface (because the surface you see is actually the boundary of the air volume, not really the surface of the car). If you find out more, thanks for letting me know!

[@turtlemicro]

20:39 Why do you multiply the static pressure with the density? Isn't the calculated pressure in OpenFOAM the static pressure? Is it possible to use the formula for the incompressible case cp = 1 - (c / c_infinity)^2 ? Thanks for the great video!

[@AirShaper]

By default, OpenFOAM (at least for the simpleFOAM solver) will write the normalized pressure, which means it is divided by the density. To obtain the real pressure (and thus the correct force) you indeed need to multiply it with the density. You can indeed calculate the pressure coefficient as well (which can be done using both the static and the total pressure) to get a "relative" distribution. When we show "pressure clouds" in our reports, we use an isosurface for the total pressure coefficient for a value of 0 (which nicely shows you the wake!). There are commands in openfoam to calculate the total pressure as a new field - those are pretty useful as well.

[@turtlemicro]

@@AirShaper okay, makes sense. Thank you very much!

[@nikhilbhamare6761]

Yo very good bro. Love your vdeos.👍🏻

[@AirShaper]

Thanks Nikhil, I hope it's useful!

[@pepijnhuvenaars]

@9:45 I can't see which symbol you type between kHat and coords.

[@AirShaper]

it's a minus sign! --> ...kHat-coords)

[@Robert-dv3uz]

how do we came from cp=(p-p_infinity)/(0.5*rho*U_infinity^2) to (p+0.5*rho*U_mag^2)/(0.5*rho*U_infinity^2)?

[@AirShaper]

In OpenFOAM, the pressure values are relative to a reference pressure we set ourselves (p-infinity). And we set p_infinity to 0 in our simulations, so it drops from the equation.

[@architsaxen]

Can you please let me know how can I clear a surface feature error for open manifolds in my geometry?

[@서지완학생공과대학기]

Thanks for your services. I have a question. In many videos about calculating drag force, during the calculator step, they often don't multiply by density. Can you explain this part in detail? I'm not good at English. Thanks read my question.

[@AirShaper]

It depends on how they store the pressure field. In some calculations, the pressure field p is the real pressure. In other calculations, it is the normalized pressure (divided by the density). It depends on which solver you use, if you use a reference pressure, compressible flow, etc.

[@서지완학생공과대학기]

Thanks :)@@AirShaper

[@LT72884]

question, in paraview 5.12.0-RC, when you search for "generate normals" it is greyed out and i can not select it? why would that be?

[@gamergoohoo]

Thank you !

[@TheBadoumba]

Hi Wouter. Thanks for the video. Could you explain the origin of the difference between AirShaper and Paraview visual regarding the total pressure bubble (additional layer on the surface of the car). Is it something we can correct somehow in Paraview? Thanks!

[@AirShaper]

Hi Sylvain, for the AirShaper report, we first cut away the cells that are touching the surface of the object (in OpenFOAM). This avoids having the entire object surrounded in a "pressure bubble". We haven't found a way to do that in Paraview, but we stopped looking after we solved it in OpenFOAM directly. I would guess that with some creativity this should be possible in Paraview as well.

[@TheBadoumba]

Ok Wouter 👍 The result is perfect! Great work.

[@simprolougebycadkonstrukti4654]

Hello, very nice tutorial! Thank you! One question: Do the forces and tourque relate to the Center origin of the car? So to coordinates of 0 0 0 in world space? So if my cog would be middle of front axis in the car 3d model, would the force pull on that point? Hope you understand my question…

[@AirShaper]

Yes, the forces and moments relate to 0 0 0. If you want to change that, you can do a translate operation on the entire internal mesh, to reposition the origin.

[@nilmarosa24]

hello, one question where can I download the file to follow up the tutorial?

[@AirShaper]

Dear Nilma the Drivaer data is not available. But currently, you get a coupon for a free simulation when you create an AirShaper account. And that one comes with the OpenFOAM result data you can use for this tutorial!

[@HoaiThanh-wm6ze]

why is the total force calculated by (p*normal-wallshear)*rho. as I know this is need by equation (p*normal+wallshear)*rho. what is the correct?

[@andrisitepu4750]

Can you simulate heat transfer on the hatsink?

[@hessmess9504]

Excellent. Thank you :-)

[@AirShaper]

Thank you very much Hess!

[@LT72884]

Hello : ) I love this video as well as your other videos. I JUST started learning cfd and fea this week. It is a very overwhelming subject. With that in mind, do you have the video that shows how you made the "virtual wind tunnel", meshing and the use of OpenFOAM? Thank you :)

[@AirShaper]

We don't have a video on this we have some very old OpenFOAM course material: airshaper.com/courses

[@LT72884]

@@AirShaper thanks for the link, ill at least check that out haha.

[@esarworks4963]

Hello, Thanks for the tutorial, its very helpful for someone new to ParaView. I wanted to know if you could share some info on what computational resources were need to run the DrivAre model. Runtime, CPU, hours, etc. that kind of thing.

[@AirShaper]

This sim would typically take around 3 hours on a machine with around 170 virtual CPUs and a few 100gb of memory. But you can run it on smaller machines as well.

[@praveen667]

Can you help me to get the access of the car model you used sir? pls?

[@ilserojas9152]

Thank you so much for this video, it was very clear and helpful. When I try to use Surface LIC in the internal mesh, Paraview suddenly closes and the Output Messages window pops up but with no messages inside. Could you please explain what are the possible reasons and solutions to this? Thanks in advance (:

[@AirShaper]

Ow haven't had that before yet - surfaceLIC perhaps only works on surfaces (a slice, object surface, ...). Perhaps that is the reason?

[@pedrobrito8080]

Hello, I enjoyed the presentation very much. Very helpful, as I´m new to car aerodynamics and OpenFoam. Can you clarify the physical meaning of the Cp=0 isosurfaces in regard to being a drag source? I think you called it "red clouds". The reference pressure was 0 Pa and set at the outlet BC, I´m guessing? Best regards, Pedro

[@AirShaper]

Hi Pedro, I use Cp-total in tje tutorial. That is the pressure coefficient calculated using the total pressure. If you make a 3D isosurface of that in the air (so not at the outlet boundary) you get a 3D surface that denotes the split between zones with a total pressure coefficient above and below zero. Everything contained within those clouds has a Cp-total lower than zero (so lower than the total pressure of the free stream). These low pressure 'bubbles' hold the vehicle back, as this negative pressure creates a suction effect on faces pointing towards the rear of the car in many areas.

[@pedrobrito8080]

@@AirShaper Thank you! I found the channel today and it has been very educational. The chats with real world designers are a great touch. Please keep producing content.

[@AirShaper]

Thanks Pedro, will do, good luck to you!

[@AirShaper]

@@pedrobrito8080 Thanks a lot! And good luck with your projects!

[@braedenswid8084]

Great video! How do I know if my data is normalized with density? Is this always the case? Thank you again.

[@AirShaper]

Hi Braeden, the field files of OpenFOAM don't provide you with that information I'm afraid. The best way is to check, if you have the case files, in the constant properties & controlDict files which density & solver were used. Then, checking OpenFOAM documentation, you can deduct if the density was normalized. I'm not sure if it's always normalize by the density, it could be!

[@broniaa1536]

Hi, thanks for this video, it's been very helpful. Can you clarify one thing for me please? After integration of force and torque, how do I use the single integrated vector in following calculations instead of a "map" of force vectors?

[@AirShaper]

Hi! I don't think I completely understand the question - what do you mean with the single integrated vector (force & torque are not combined, I think, in our video?)?

[@broniaa1536]

@@AirShaper No no I meant using i.e. resulting (so integrated) force vector instead of force field. But I can see that after IntegratingVariables it already uses the single vector in further calculations. Thanks again

[@broniaa1536]

@@AirShaper Do you maybe have some hints on calculating the center of pressure? I already calculated the arm of the force and not sure how to proceed

[@peterb8387]

Why does the data from OpenFOAM (pressure, velocity, etc.) sometimes now show up? Just a grey block in the display portal/window? I use Salome to make my meshes. Typically I past the new mesh in another case file and usually run into this issue.

[@AirShaper]

Sometimes it could be because you need to go to the last time step, or select the decomposed case data

[@jaydeepchawda9105]

Very helpful

[@AirShaper]

Thank you Jaydeep!

[@2JFF]

How is it possible to get the wallShearStress? or how to calculate it?

[@AirShaper]

It's not a default in an Openfoam calculation, but there is a wallshearstress command to generate it.

[@alvarolleonart6662]

Hi there! How could I obtain the wallShearSress if I already did the simulation and it does not appear in paraView? Do i need to change any file or include "wallshearstress" somewhere? It would be very helpful to know some information, I need this variable to calculate my forces on my 3D F1. Thanks!

[@AirShaper]

You need to run a separate post processing command in Open foam to get that field

[@Willy_LGM]

Is there any way we can download the sample files from this tutorial please?

[@AirShaper]

Hi Willy, not for this model directly (although the Drivaer car can be requested at the TUM - Technical University of Munchen). I recommend using the simulation data of this public project instead: app.airshaper.com/projects/sample/sports-car Just click on the button "Simulation data" to download it - the structure is entirely the same as for this tutorial!

[@piereligio_ds]

I haven't wallshearstress in my openfoam simulation. I generated it using Xsim, maybe it's not the best way. How do you recommend to run a simulation for this use case?

[@AirShaper]

I haven't heard of Xsim before. In OpenFOAM, the wallShearStress is generated during post processing with a separate command.

[@piereligio_ds]

@@AirShaper that's great to know, thanks. Xsim is a nice web based (runs locally) GUI tool that prepares the scene for openfoam, which then you'll run on your machine. But for some reason they didn't include the wallshearstress in the post processing, but the calculation of drag and lift coefficients is there. But I guess it's wrong

[@AirShaper]

@@piereligio_ds Interesting, I hadn't heard of it before, thanks a lot! Hm, then best to double check all the calculation settings if the values are wrong :)

[@piereligio_ds]

@@AirShaper i assume it’s wrong because there is no wallshearstress output, so that isn’t used for sure

[@AirShaper]

@@piereligio_ds It could still be correct, the OpenFOAM "forces" and "forceCoeffs" might take wallShearStress into account (without writing that field)

[@saidtacuri9010]

Could you share the 3D file?, please 🙏

[@diegoandrade3912]

Bravo!!!

[@AirShaper]

Thanks Diego!

[@idreeskhan1629]

Nice explanation,could you plz upload a vedio how to extract the data and from paraview and plot in gnuplot,how to compare the extracted data with some other results usin gnuplot.

[@AirShaper]

Hi Idrees, such a video would be too specific / detailed for our channel. But I've exported data from curves for example to csv, to then plot in excel!

[@idreeskhan1629]

@@AirShaper where is that vedio?

[@AirShaper]

@@idreeskhan1629 There is no video on that topic - it's just how I did it in the past. For example, you can first plot the surface data on an object and then make a 2D slice through it. That will give you a curve. Then, you can click that slice and go to "File - Save data" and save it as a csv file. You can then process that data in excel. You may need to first add coordinates to the results (using a calculator) so X, Y and Z are exported as well.

[@idreeskhan1629]

@@AirShaperthank you, kindly if you send me your email i will send you a paper whose graphs i need if you guide me how draw those graphs please,

[@AirShaper]

@@idreeskhan1629 Hi Idrees, I can't go as far as to provide custom assistance for free for everyone. If you need assistance, you can contact me at [e-mail] for regular (paid) consultancy.

[@ashok305]

hello, can we animate results..?? could you plz make a video thank you

[@AirShaper]

Dear Ahsok, yes, I believe it is possible to make animations in paraview (changing camera positions and even having particles move along a trajectory?). We don't have enough experience on that ourselves to make a video, however. There is quite a large paraview community (with forum), perhaps you can find answers there!

[@michielvanstaden2698]

Big Thumbs Up

[@AirShaper]

Thanks Michiel!

[@Tom-hg9lw]

Can you share the model please?

[@AirShaper]

This is a public model available for download here: www.mw.tum.de/en/aer/research-groups/automotive/drivaer/ Please note there are many different combinations possible (estate, fasteback, detailed or simple underfloor, ...). Good luck!

[@VinnieDreher]

Te amo !

[@AirShaper]

Hope this all makes you go faster! I saw you race motorbikes --> be sure to check our interviews with Mv Agusta: airshaper.com/videos/race-talk-with-mv-agusta-top-aerodynamic-priorities-for-moto-gp-bikes/video/BUkaejkHRn8 airshaper.com/videos/race-talk-with-mv-agusta-how-to-improve-moto-gp-aerodynamics/video/fGNdMaVft_w