[winterboy]

Hi everybody,

I used fluent to model flow around NACA 0015.
I got cl & Cd for the same flow and airfoil from Xfoil and want to compare Xfoil results with fluent results.
I want to find Cl and Cd by fluent and my problem is the numbers which I have to enter for X,Y,Z of force vector.
do you know what should I put in there and how many iterations should I have at least to get acceptable answer for Cl and Cd?
Thank you in advance

[-mAx-]

enter 1 0 0 for Force Projection on x-axis
enter 0 1 0 for Force Projection on y-axis
enter 0 0 1 for Force Projection on z-axis
Now choose the right force you want to observe, and enable the plot option in monitor/force
Fluent will plot the accorded Coeff while iterating.
Your solution will be full converged if the the plotted coefficient remains constant while iterating.
You also have to set the right reference values for getting the right Coefficient

[3g_nitro]

Originally Posted by -mAx-

enter 1 0 0 for Force Projection on x-axis
enter 0 1 0 for Force Projection on y-axis
enter 0 0 1 for Force Projection on z-axis
Now choose the right force you want to observe, and enable the plot option in monitor/force
Fluent will plot the accorded Coeff while iterating.
Your solution will be full converged if the the plotted coefficient remains constant while iterating.
You also have to set the right reference values for getting the right Coefficient

when you do that it shows the values for the coefficient of pressure, so how do you work out CLift ??

[kdrbrk]

before iterations, set "monitors-lift" and define the lift vector (ex: y=1)
and select your airfoil (must be a wall) for which the lift will be monitored.
also dont forget to set correct ref. values.

than begin iteration and it will show you "cl" in each step by default.

[RGRUIZ]

Originally Posted by -mAx-

enter 1 0 0 for Force Projection on x-axis
enter 0 1 0 for Force Projection on y-axis
enter 0 0 1 for Force Projection on z-axis
Now choose the right force you want to observe, and enable the plot option in monitor/force
Fluent will plot the accorded Coeff while iterating.
Your solution will be full converged if the the plotted coefficient remains constant while iterating.
You also have to set the right reference values for getting the right Coefficient

Hi max, one question..I am modelling a 2D airfoil, the right references values should be:

Area: What value should i use?? The edge length of the airfoil multiply by the depth???

Depth: is this the length of the wing??

[-mAx-]

Check tutorial 3: Modelling external compressible Flow.
The references values you give are independant from your solution.
Compute your model, and then you can apply your refences values.
Even if you gave wrong Ref. Values, you can correct them without re-iterating.
They are just factors.
Important are pressure and viscous forces computed during iterations

[John222]

I have the same problem and your answer indeed helped me a lot. I have monitored the Cl and Cd and while iterating I found out that my coefficients are not constant in the plot what I am supposed to do in this situation. Please explain in details I am a beginner.

Check pictures attached.

R.I.P for your dear friend Herve.

Thanks

[RGRUIZ]

Hi max...i figure it out that if i change the depth on reference values i get differents cl y cd results..why is that if the equation is: Cl=F/(0.5*density*velocity*velocity*Area)..

Another question what if the angle of attack is different than zero, which velocity value on farfield should be taken to calculate the lift coefficient???

[-mAx-]

Originally Posted by John222

I found out that my coefficients are not constant in the plot what I am supposed to do in this situation. Please explain in details I am a beginner.
Thanks

You are iterating. Your computation doesn't converged yet, and your coefficient (monitor) will oscillate till it reaches its converged value.

[-mAx-]

Originally Posted by RGRUIZ

Hi max...i figure it out that if i change the depth on reference values i get differents cl y cd results..why is that if the equation is: Cl=F/(0.5*density*velocity*velocity*Area)..

Another question what if the angle of attack is different than zero, which velocity value on farfield should be taken to calculate the lift coefficient???

*Check Online Help >>> "For 2D problems, an additional quantity, Depth, can also be defined. "

*I would say you can take the cosine of your angle as factor. But since the angles are small, it won't affect your coeff (needs to be confirmed/invalidate from aero experts).
Check tutorial 3 and compare velocity value given in Ref. with Mach Number

[John222]

Originally Posted by -mAx-

You are iterating. Your computation doesn't converged yet, and your coefficient (monitor) will oscillate till it reaches its converged value.

OK I did 500 iterations and I found that my line is stable if you can see in the picture attached.

Is my solution converged, and how to compare 1st order with 2nd I am running now in 2nd order (pressure and momentum).

I see no difference really is the difference from the drag and lift forces or the plot of the iterations to drag.

How to do a GIS, I probably done it already?



R.I.P to your friend again

[-mAx-]

you need also to check your residuals (plot/residuals...)
If all the residuals fall down, let iterate and check if your monitoring still stay constant.
Regarding your comparison between first and second order: write your force monitoring on a file (one file for 1st order, and one for second).
Write also case and data before switching to second order.

What is GIS?

[John222]

Originally Posted by -mAx-

you need also to check your residuals (plot/residuals...)
If all the residuals fall down, let iterate and check if your monitoring still stay constant.
Regarding your comparison between first and second order: write your force monitoring on a file (one file for 1st order, and one for second).
Write also case and data before switching to second order.

What is GIS?

- What do you mean by all residuals fall down, I know how to get the residuals plt but how to know if they are down?

- grid independence study?

[-mAx-]

*Residuals charts decreasing (>> converging)
*Grid independant study: then you need to recompute your stuff on a refined grid.
You reached the solution if your solution isn't influenced by the grid refinement.

[John222]

Originally Posted by -mAx-

*Residuals charts decreasing (>> converging)
*Grid independent study: then you need to recompute your stuff on a refined grid.
You reached the solution if your solution isn't influenced by the grid refinement.

- does my graphics & animations plots of static pressure changes when changing from 1st order to 2nd order discretization, and what actually changes in 2nd order discretization.

- and I still didn't get the full idea of grid independent study, I have tried to refine my grid as best as I could. Do I have to do another very well done mesh and compare both answers or what?

- Please check your email.


Thanks dude.

[John222]

Originally Posted by -mAx-

*Residuals charts decreasing (>> converging)
*Grid independant study: then you need to recompute your stuff on a refined grid.
You reached the solution if your solution isn't influenced by the grid refinement.

Here is my mesh. Is it possible to do a 2nd order with it and is it converged enough.

[RGRUIZ]

-John..

The grid independent study is like this:

You have to have different mesh faces: example 10.000-20.000-30.000 and fix a parameter you want to study like Cf, Cl, Cd if any this values change less than 5% from mesh 10.000 and 30.000 means that it won´t be any different if you do a mesh with 40.000 faces because you are going to get very similar results, so you considerer that the 30.000 faces is okay for what you are doing.

hope it help you

[-mAx-]

Originally Posted by John222

- does my graphics & animations plots of static pressure changes when changing from 1st order to 2nd order discretization, and what actually changes in 2nd order discretization.

- and I still didn't get the full idea of grid independent study, I have tried to refine my grid as best as I could. Do I have to do another very well done mesh and compare both answers or what?

-read case for with 1st order>> initialize>> compute and write your coeff in a file. When converged, write case & data.

read case & data from previous 1st order solution>> switch to 2nd order >> Initialize or don't(if you start from converged 1st order solution you will get faster convergence) >> compute and write your coeff in a different file.
Load your 2 graphs in excel, or whatever... and compare


- You don't need to generate a very fine grid: follow what RGRUIZ said

[John222]

Originally Posted by RGRUIZ

-John..

The grid independent study is like this:

You have to have different mesh faces: example 10.000-20.000-30.000 and fix a parameter you want to study like Cf, Cl, Cd if any this values change less than 5% from mesh 10.000 and 30.000 means that it won´t be any different if you do a mesh with 40.000 faces because you are going to get very similar results, so you considerer that the 30.000 faces is okay for what you are doing.

hope it help you

Sorry but when you say different mesh faces do you mean the Sizing (maximum face size) I am using (car.x_t) geomtry file from Ansys.
and in the max face size by default its 1.78m.
can someone tell me the full steps of how to change mesh faces 10.000-20.000-30.000.

[John222]

Originally Posted by -mAx-

-read case for with 1st order>> initialize>> compute and write your coeff in a file. When converged, write case & data.

read case & data from previous 1st order solution>> switch to 2nd order >> Initialize or don't(if you start from converged 1st order solution you will get faster convergence) >> compute and write your coeff in a different file.
Load your 2 graphs in excel, or whatever... and compare


- You don't need to generate a very fine grid: follow what RGRUIZ said

Very helpful understood you, and testing now...