Learning Forum

[JoaoAlex]

Hello folks,

I need to confirm a few topics about Bladegen and DM software in the creation of turbomachinery geometry:

(1) Are those 4 parameters sufficient to completely define a general impeller geometry on Bladegen?

- Meridional profile

- angle distribution (represented by beta or theta)

- thickness distribution

- number of blades

(2) In the standard procedure for geometry creation, the angle distribution is defined in the 0% span layer and 100% span layer, how does Bladegen interpolate or create the angle distribution for the other 3 span layers (0.25/0.5 and 0.75)

(3) Why the blade-to-blade view uses the M' vs theta coordinates system (or M vs R*theta; or M vs theta')? Usually, a see people defining the airfoil parameters in a 2D plane, where there is the axial chord, tangential chord, blade angles, leading-edge radius, and so on. I read that the beta angle is defined in the (m,s) coordinate systems, but still can't relate this to this blade-to-blade view. I get confused when converting cascade parameters from twisted rotors to 3D geometry in Bladegen.

[JoaoAlex]

[Rob]

It's enough to parameterise those values. As to whether it's enough for Bladegen, I have no idea.

[rfblumen]

1.) Specifying those four parameters (meridional profile, angle distribution, thickness distribution blade count) in general will describe the blade. There are other features that can be included in addition to these (blade leading edge/trailing edge ellipse/cutoff, overfile/underfile, tip clearance, etc). Recall that BladeGen can't be parameterized. To parameterize the geometry, the design from BladeGen needs to be transferred to BladeEditor (i.e. DesignModeler) via a "Load BGD" in BladeEditor or exporting a Neutral Data File in BladeGen and importing the NDF into BladeEditor.
2.) The exact details on how the angle interpolation is done isn't provided, but I would assume that given just the hub and shroud layers it's using a linear interpolation at a given non-dimensionalized streamwise location in the meridional space. You can observe this looking at the beta graph or theta graph in BladeGen.
3.) In the old days of blade design layouts, a 3D blade section at a constant span fraction (say at 50% span) would be collapsed to 2D using a conversion from Cartesian (x,y,z) to (m,s) coordinates where m is the meridional coordinate (m=integral(sqrt(dr^2+dz^2))) and s is integrated arc length (s=integral(r*dtheta)). The coordinates (m,s) are equivalent to (M,Theta') in BladeGen. The reason why (M,Theta') is not the default coordinates for displaying the developed view of the blade section in BladeGen is that for geometries where radius is increasing (i.e. radial impellers) the periodic boundaries "fan out" in the developed view and make displaying neighboring blades impossible without distorting them. By normalizing (m,s) by radius, we get (M',Theta) which allows displaying neighboring blades in the developed view for radial machines without distorting the neighboring blades.
Note that all span fractions in BladeGen represent a surface of revolution in 3D Cartesian space. If dealing with axial blade geometry where the geometry is described in terms of actual 2D slices through the 3D geometry, to properly transfer the geometry to BladeGen it would need to be described on constant radius surfaces.

[JoaoAlex]

thanks!