Dear Nices,

Is the velocity magnitude in FLUENT defined as following (for non-swirl cases):

Velocity Magnittude=ABS(Ux)+ABS(Uy)+ABS(Uz),

where ABS() means the absolute value and Ux, Uy and Uz are the velocitiey components at xyz Cartesian coordinate.

Best regards Koorosh

Dear Koorosh,

If it is ok to patch velocity magnitude in initialization for calculating domain, what do you think about to use "User Field Function" in 'Define' menu? If velocity magnitude needs to be assigned as boundary condition, I think UDF is needed.

Thanks,

J.W.Ryu

The correct definition of the velocity magnitude is SQRT(Ux^2 + Uy^2 + Uz^2).

Dear Janos,

Basically the definition you wrote is the magnitude of velocity vector U=Uxi+Uyj+Uzk in 3D cartesian space. So it seems that the velocity magnitude has to be equal to the [U]=(Ux^2+Uy^2+Uz^2)^0.5. But as I checked by FLUENT software for several simulation (those I'm doing or the solved problems) the "velocity magnitude" is different from that you wrote. To check it, just please open a solved problem in FLUENT and write the x,y,z velocities and also velocity magnitude. Next to it you will find that SQR(Ux^2+Uy^2+Uz^2) is so different from "velocity magnitude" reported by FLUENT.

Best regards

Koorosh

Yep, I just checked it directly in a Fluent simulation and my definition is correct.

Remember to un-check the "node-values" option (if you report node-values interpolation errors can cause differences).

Dear Janson,

You are exactlly right. I did switch on the node values and the problem was so. Now as you told me I did switch off those node values in order to prevent the interpolations, and now the velocity magnitude is OK. For an example, with nodes value the velocity magnitude reported by FLUENT and that one obtained by SQR(Ux^2+Uy^2+Uz^2) has an average error nera 0.81%. even though it seems OK, but this error for the case of NO NODE VALUES is 0.00008%. So thanks a lot. But I have another question. I will writte it in another massage.

Best regards

Koorosh