[titanan]

Hey there. I was suspicious my drag values were lower than they supposed to be. So ı analised a 100mm radius sphere at 30m/s. By the results from internet ı was expecting around 0.1 cd. but my result is 0,005. I used the sst-k omega model and the y+ value is between 5 and 250. My mesh orthagonal quality is 0,4min. and skewness is 0,6 max. Can any of you re-do the experiment and check the values? Where can ı be wrong?

[FMDenaro]

Originally Posted by titanan

Hey there. I was suspicious my drag values were lower than they supposed to be. So ı analised a 100mm radius sphere at 30m/s. By the results from internet ı was expecting around 0.1 cd. but my result is 0,005. I used the sst-k omega model and the y+ value is between 5 and 250. My mesh orthagonal quality is 0,4min. and skewness is 0,6 max. Can any of you re-do the experiment and check the values? Where can ı be wrong?

A good drag evaluation requires a well refined grid around the surface.

[LuckyTran]

I did it real sloppily with a wall y+ up to 170 and quickly got a Cd of 0.072


I agree, your results are really sus

[CFDfan]

Originally Posted by titanan

Hey there. I was suspicious my drag values were lower than they supposed to be. So ı analised a 100mm radius sphere at 30m/s. By the results from internet ı was expecting around 0.1 cd. but my result is 0,005. I used the sst-k omega model and the y+ value is between 5 and 250. My mesh orthagonal quality is 0,4min. and skewness is 0,6 max. Can any of you re-do the experiment and check the values? Where can ı be wrong?

what is the Reynolds number you've calculated?
SST K-Omega is good for Y+<1. For your Y+ range you better use Realizable K-E or the standard K-E.

[titanan]

Originally Posted by CFDfan

what is the Reynolds number you've calculated?
SST K-Omega is good for Y+<1. For your Y+ range you better use Realizable K-E or the standard K-E.

Reynolds number is around 400000 ı did the same analysys with those methods too but cd nearly stays same. (still around 5e-03) I'm going to try refinig the mesh around and behind the sphere and try again. Thanks for the advice.

[LuckyTran]

I'm not sure if the issue is actually a meshing issue.

Pretty much the only thing that has a drag coefficient that low is a plate aligned parallel with a flow. That is, essentially 0 drag, and almost 0 disturbance on the flow field. I have to entertain the thought that the drag force is not being calculated correctly at all. If you run CFD and see any form of a wake behind the sphere you should already be seeing higher Cd's

[titanan]

Originally Posted by LuckyTran

I'm not sure if the issue is actually a meshing issue.

Pretty much the only thing that has a drag coefficient that low is a plate aligned parallel with a flow. That is, essentially 0 drag, and almost 0 disturbance on the flow field. I have to entertain the thought that the drag force is not being calculated correctly at all. If you run CFD and see any form of a wake behind the sphere you should already be seeing higher Cd's

I was thinking the same. But it seems like i tried everything. changing the mesh resulted in a increase to around 1.7e-02 but thats still way too low. I pretty much dont know what to do. Guess ımma keep tweaking the settings.

[LuckyTran]

Let me be clearer what I mean. Stop tweaking settings for a clerical error.

Please provide a plot of the velocity field. If I see any wake behind your sphere, then you are calculating the drag coefficient incorrectly because there's no way you get that low of a Cd with any kind of wake. If I see no wake then I will question whether you are doing an inviscid flow around a sphere.

[CFDfan]

Originally Posted by titanan

Reynolds number is around 400000 ı did the same analysys with those methods too but cd nearly stays same. (still around 5e-03) I'm going to try refinig the mesh around and behind the sphere and try again. Thanks for the advice.

Something is probably wrong with your postprocessing. If the flow is in X direction, then CD=2*Fx/(ro*V^2*S), where S=0.031415, V=30 and ro=1.206 (or whatever ro_air you have used).
I got CD=0.09 without any mesh tuning and Realizable k-e.

[titanan]

Originally Posted by LuckyTran

Let me be clearer what I mean. Stop tweaking settings for a clerical error.

Please provide a plot of the velocity field. If I see any wake behind your sphere, then you are calculating the drag coefficient incorrectly because there's no way you get that low of a Cd with any kind of wake. If I see no wake then I will question whether you are doing an inviscid flow around a sphere.

Here is a photo of the velocity field. https://ibb.co/G030Kjp Thanks for your time btw.

[FMDenaro]

Originally Posted by titanan

Here is a photo of the velocity field. https://ibb.co/G030Kjp Thanks for your time btw.

But did’nt get a convergence to a steady state?

[titanan]

Originally Posted by FMDenaro

But did’nt get a convergence to a steady state?

No it didn't converged. Should ı increase the number of iterations and wait for it?

[FMDenaro]

Originally Posted by titanan

No it didn't converged. Should ı increase the number of iterations and wait for it?

Without any convergence what are we discussing?

[titanan]

Originally Posted by FMDenaro

Without any convergence what are we discussing?

I managed to make it converge at 136th iteration. Nothing changed. Thanks for your answer

[FMDenaro]

Originally Posted by titanan

I managed to make it converge at 136th iteration. Nothing changed. Thanks for your answer

Are you running a steady solver? Could you show the history of convergence?


Do you reach convergence for a steady laminar flow (without turbulence model) and have a correct Cd?

[arjun]

Could you ask solver to report force on the sphere.

Cd etc could be affected by reference values. Your velocity field says that your cd shall be around 0.4 to 0.5.

Also check for stagnation pressure in front of sphere which shall be 0.5rho x velocity x Velocity.

[titanan]

Originally Posted by arjun

Could you ask solver to report force on the sphere.

Cd etc could be affected by reference values. Your velocity field says that your cd shall be around 0.4 to 0.5.

Also check for stagnation pressure in front of sphere which shall be 0.5rho x velocity x Velocity.

Hey
Thanks for your response. Sorry for writing back late.
Force report is true when calculated by the given cd by ansys. I unfortunatly don't know what you mean by stagnation pressure but the total and static pressure are both pretty close to the calculated value of 0.5rho x velocity x Velocity (by an error margin of around 5%)
Looking forvard for your responses,
A desperate guy just trying to get a close drag value.

[titanan]

Originally Posted by FMDenaro

Are you running a steady solver? Could you show the history of convergence?


Do you reach convergence for a steady laminar flow (without turbulence model) and have a correct Cd?

Hey
Thanks for your response. Sorry for writing back late.
If i understood you correctly, here is the history of convergence.


https://ibb.co/0BfdmMs
https://ibb.co/QQnN543
I tried it with laminar flow but it seems like it doesn't want to converge. Drag coefficient is higher at around 1.5e-02 but its still far away than what is is supposed to be. ı improved my mesh quality even more. I folowed that tutorial for the whole process and nothing changed.
https://www.youtube.com/watch?v=nh94...el=SNOWYOUTUBE
I would appreciate your further support. Looking forward for your answers. Thanks

[FMDenaro]

Originally Posted by titanan

Hey
Thanks for your response. Sorry for writing back late.
If i understood you correctly, here is the history of convergence.


https://ibb.co/0BfdmMs
https://ibb.co/QQnN543
I tried it with laminar flow but it seems like it doesn't want to converge. Drag coefficient is higher at around 1.5e-02 but its still far away than what is is supposed to be. ı improved my mesh quality even more. I folowed that tutorial for the whole process and nothing changed.
https://www.youtube.com/watch?v=nh94...el=SNOWYOUTUBE
I would appreciate your further support. Looking forward for your answers. Thanks

Try first to get the correct setting for the laminar flow...

[arjun]

Originally Posted by titanan

Hey
Thanks for your response. Sorry for writing back late.
Force report is true when calculated by the given cd by ansys. I unfortunatly don't know what you mean by stagnation pressure but the total and static pressure are both pretty close to the calculated value of 0.5rho x velocity x Velocity (by an error margin of around 5%)
Looking forvard for your responses,
A desperate guy just trying to get a close drag value.

You say that force values are correctly calculated but then the cd value doesn't match the velocity profile you are showing. On top of that you are using fluent and i find it hard to believe fluent got the pressure so wrong that its integral on the surface is so off.

I can't comment much on it without knowing

1. Dia of sphere.
2. Free stream velocity and density of fluid

And
3. Force value reported by the solver.

You say all these are okay but in the end cd doesn't match up.

Anyway is it possible for you to share the fluent mesh (.msh) and let us know the free stream velocity. I can try myself in my solver (which is very much like fluent).