[seasoul]

I am modeling a fluid impact on a cylinder. Therefore I set it in the cylindrical coordinate. The veloctiy at impact will be resolved into two components, radial and azimuthal velocity. Let's say velocity is 1 m/s, the fluid impacts on the cylinder, not through the circle center. Assume the angle is 30 degrees, the diameter of the cylinder is 1mm. Then the radial velocity is 0.86m/s, the azimuthal velocity is sin30/0.0005=10000 rad/s. But I found the fluid flew away from the computational domain within a quite short time. But physically it should impact on the cylinder. Is the velocity in y direction (theta) in a unit of radian/s?

[MuxaB]

I can't quite picture this in cylindrical coordinates. Where are you setting the inlet velocity?

Why not do it in a Cartesian grid?

Originally Posted by seasoul

I am modeling a fluid impact on a cylinder. Therefore I set it in the cylindrical coordinate. The veloctiy at impact will be resolved into two components, radial and azimuthal velocity. Let's say velocity is 1 m/s, the fluid impacts on the cylinder, not through the circle center. Assume the angle is 30 degrees, the diameter of the cylinder is 1mm. Then the radial velocity is 0.86m/s, the azimuthal velocity is sin30/0.0005=10000 rad/s. But I found the fluid flew away from the computational domain within a quite short time. But physically it should impact on the cylinder. Is the velocity in y direction (theta) in a unit of radian/s?

[seasoul]

I am actually modeling a droplet, so there is no inlet velocity issue. And I want to postprocess the data close to the cylinder surface, if in cartesian coordinate, is it possible to create a slice parallel to the curvature surface?

[MuxaB]

Originally Posted by seasoul

I am actually modeling a droplet, so there is no inlet velocity issue. And I want to postprocess the data close to the cylinder surface, if in cartesian coordinate, is it possible to create a slice parallel to the curvature surface?

Yes, that what an 2D plot in an y-z plane in the Analyze tab would do.

If you are trying to define an initial uniform droplet velocity moving towards the cylindrical axis, I don't think you can do it using the available input in FLOW-3D (if understand your problem). It is much easier in a Cartesian grid.