[9mile]

Hi all.

Recently i've simulated the2-D flow past a circular cylinder.

But i got some strange picture from the vorticity contour.

With the zero of vorticity, the contour display wiggles in the whole domain.

Without zero, the picture looks good.

I don't know what happen.

I attached both pictures.

So, please take a look and give some opinions

Is there someone who experienced same situation ?

Any comment or advice is welcome.

[FMDenaro]

It's quite normal that plotting iso-vorticity = 0 you get a sort of wiggles...
the best you can do is working with double precision and using a low threshold in your iterative solvers but that would not always work better.
The important issue is to check the flow field variables

[9mile]

Hi FMDenaro,

Thank you for a quick reply.

I am going to try to post-process with double precision.

But if it would happen same result and it is quite normal,
Can i proceed with excluding the iso-vorticity=0 region ?

What does it mean to check the flow field variables?

Pressure and velocity show good results in quantitatively and qualitatively way.

I can't find the wiggles on zero-region for the pressure and velocity.

Even both flow variables satisfied the convergence criteria.

Regards.

[yanbin]

Just try to set a smaller limit of vorticity to display,such from -5 to 6.

[FMDenaro]

Originally Posted by 9mile

Hi FMDenaro,

Thank you for a quick reply.

I am going to try to post-process with double precision.

But if it would happen same result and it is quite normal,
Can i proceed with excluding the iso-vorticity=0 region ?

What does it mean to check the flow field variables?

Pressure and velocity show good results in quantitatively and qualitatively way.

I can't find the wiggles on zero-region for the pressure and velocity.

Even both flow variables satisfied the convergence criteria.

Regards.

I suppose you compute the vorticity from the velocity field, thus the discrete derivative causes possibly errors, furthermore the "zero" value is something that the isolines interpolate from small values, for example think about what would happen interpolating between 10^-6 and -10^-6 (that wolud be zero in a single precision).

[9mile]

Thanks for reply.

Dear yanbin, it still doesn't work despite i adjusted the limit.

Dear FMDenaro,

As you mentioned, i calculated the vorticity from the velocity field.

And it is true that small variations make the iso-lines of zero with the wiggles by interpolating single precision values between the nodes.

But i am not sure that it is possible to fix the wiggles only with the double precision.

If not, can i go with intentionally excluding the wiggles?

Regards.

[FMDenaro]

Originally Posted by 9mile

Thanks for reply.

Dear yanbin, it still doesn't work despite i adjusted the limit.

Dear FMDenaro,

As you mentioned, i calculated the vorticity from the velocity field.

And it is true that small variations make the iso-lines of zero with the wiggles by interpolating single precision values between the nodes.

But i am not sure that it is possible to fix the wiggles only with the double precision.

If not, can i go with intentionally excluding the wiggles?

Regards.

the key is understanding what wiggles mean...so check the magnitude, it should be very low, around zero.
However, you solve for velocity and pressure, therefore are such variables you really solve in the code to be checked

[9mile]

Originally Posted by FMDenaro

the key is understanding what wiggles mean...so check the magnitude, it should be very low, around zero.
However, you solve for velocity and pressure, therefore are such variables you really solve in the code to be checked

Ok, i got it.

Thanks.