Inlet flow direction at axial turbomachinery with twisted Blades
 

Hi,
i have a cascade of seven twisted blades of a axial turbomachinery. At the hub the angle between the leading edge and the inlet surface is ~-65° and at the tip it ist ~35°. So i tried to create an inlet flow profile that depends on the radius so the flow starts already at the inlet with the right angle on the blade.

I tried the 'mathematical' way with a formular like y=mr+b in the expressions and call it in the 'flow direction' -> 'cylindrical components' option, but it won't work.
Another attempt was to create a user function with linear interpolation between a start value at r_hub and an end value at r_tip, but it also didn't work the way i wanted it to.

Is there anyone who has experience with this kind of problem? Or are there any examples or literatur? The CFX-help was not very helpful.

Greetings,
Stuntmanbob


P.S.: Sorry for bad english :o

Blade twist is usually done to account for the varying circumferential velocity of the blade versus radius for an approximately constant velocity inlet flow. So why are you changing the velocity as a function of radius? Isn't it meant to be constant?

Regardless, you can easily make velocity a function of anything with CEL expressions. Please post your CCL of what you are doing and any error messages you are getting and we will try to correct it.

Hi,
of course you're right, the velocity at the inlet should be constant. I'm trying to change the direction of flow as a function of the radius. Maybe the picture will help to show.

As can be seen on the picture the flow hits the leading edge on the hub in a good angle of attack, but on the shroud the angle of attack points at the pressure side of the blade.

The aim is to adapt the angle of attack of the flow to the blade angle.
Therefore i wanted to change the flow direction under 'Boudary:Inlet', 'Boundary Conditions: Flow direction'.
In cylindrical components there are three directions: Axial-, radial- and theta-component.
With the values 1, 0, 0 (axial, radial, theta) it is the same like 'Flow direction: Normal to boundary condition'. And when changing the radiaul component the flow points at the hub or shroud surface. Now my consideration was that i have to change the theta component to a value that depends on the radius to vary the direction with the hight of the blade.
The next consideration was to create a function for the theta component that follows the form theta(r) = m*r + b. I calculated m and b and saved them as constant values in the expressions. The next problem was that ansys only excepts values that are dimensionless for theta, so i made it to theta(r) = m*(r/R) + b. I tried to change the parameters and the formular a few times, but nothing leads to the desired results.

Maybe it is a completely incorrect approach?

Greetings,
Stuntmanbob

I might be wrong at all this and would anxiously wait for Glenn's nod/ guidance...

To me it seems that the problem which makes you think you have to orient the flow direction is that usually blades are twisted (let me say they should be twisted) the opposite of what your images show.. imagine a situation where your hub and shroud aerofoils (assuming you have corrected labelled them) are reversed and obviously just like the current case you would direct the incident velocity on the hub aerofoil at a zero angle or a positive AoA with the LE (i.e. almost aligned to the LE shape).. now gradually moving upwards to shroud, the angle of attack with the chord would always increase hence increase the lift/ circumferential velocity.. i believe that is the usual approach where the incoming flow is independently introduced normal to the face/ axial and the resultant angle is formed by virtue of geometry of your blades (virtually and theoretically speaking, for twisted blades only the hub aerofoil has to be aligned at a desired AoA as moving upwards the AoA would usually always increase hence improving efficiency of the turbomachinery)...

Again, I might be over-assuming or oversimplifying and you may have a perfect reason for what you're doing, just giving my humble opinion here.. cheers.

Hi,
Sorry, they are not correct labeled. Of course the first picture shows the shroud and the second the hub :o.

Thanks for your advice :-)

Greetings,
Stuntmanbob

Alright. In that case, i believe the velocity vectors would also be a mirror of what they are right now on the horizontal axis.. in fact in your model the hub aerofoil should be aligned/ rotated to form the AoA to the normal flow velocity vector instead of the other way around..

Hi fresty,
thanks for your reply. But first let me know if i get you right.
When using the option 'flow direction: normal to boundary condition', ansys will automatically increase the flow velocity from the hub to the tip. And due to the higher radius and higher inlet velocity, the flow direction will also change and automatically fit the blade angle?
- Independent if it is a rotating or a non-rotating case?

And second that ansys automatically forms the flow angle that it best fits the geometry? :confused:

So i only have to see that the flow direction fits the blade angle at the hub an it will fit on the rest of the height of the blade?

But when i want to have a flow that hits the blade with an incidence angle? Do i only have to change the flow direction on the hub?

Greetings,
Stuntmanbob

Hi,
i made a few screenshots that shows the streamlines directing at the blade at span of 0.01 and 0.99. The flow direction option is 'normal to boundary condition' and you can see that this option does not work very well for twisted blades.

Greetings,
Stuntmanbob

You have to align the hub blade with the normal flow direction.. think of the real-time situation where the flow in axial turbomachinery initiates in the normal/ axial direction.. again, the hub blade should be oriented to make sure you have positive AoA with the flow.. from hub to shroud the AoA with the sectional aerofoils would simply increase... this comment is based on the assumption that you aim at having a positive lift, relative to the camber of your aerofoils..
Hope this helps.

Hey,
after looking into the tutorials again i found another way to specify the flow direction. In tutorial 36 they created an inlet profile for temperature and pressure and load it as initial profile data into ansys. So i calculated my on initial profile file for the circumferential component. And it worked almost fine.

The problem now is, that the flow direction fits one blade of 14. In the fifth flow channel the direction starts to turn in the wrong direction. I think the pictures shows the problem quite well.

Has anyone an idea what went wrong?

Greetings,
Stuntmanbob