[pdhyani96]

I have been simulating steady flow over a 3-D turbine stator blade with periodic boundary conditions. I simulated my flow with rotational and translational periodic boundary conditions to understand the difference.

I apply a constant total pressure at the inlet and a constant static pressure at the outlet. Upon applying rotational periodic boundary conditions, I am getting a spanwise static pressure distribution at inlet, which is high near the hub side (inner wall) and drop continuously up to the shroud side (outer wall). While for the translational periodic boundary conditions, I get a constant static pressure over from hub to shroud at the inlet. I don't understand why this happens.

That's why, I am trying to understand how rotationally periodic boundary conditions are implemented numerically in CFX. I have found some wordy explanation in the CFX solver guide which does not explain the difference in my results. I need more of a mathematical justification. Where can I find it?

[ghorrocks]

Can you post some images of the effect you are seeing?

[pdhyani96]

These are my two cases for rotationally and translationally periodic boundary conditions. Static pressure is plotted versus span. As translational periodic boundary conditions are not physically true, they make the pressure distribution unstable. But clearly, the static pressure remains constant for this case.

Why does the rotationally periodic boundary conditions give out such pressure distribution over the inlet which is higher at the inner wall and lower at the outer wall?

Note: The inlet is defined at a constant total pressure, and the outlet is defined at a constant static pressure. In real, one would have a spanwise pressure distribution over the inlet and the outlet.



Transationally periodic.jpg

Rotationally Periodic.jpg

[ghorrocks]

Please post an image of your domain which shows what you mean by rotationally periodic and translationally periodic. Also post an output file. I still do not understand what you are asking.

[pdhyani96]

My apologies. In a turbine blade row, the flow is rotationally symmetric. I can simulate the flow over one blade and replicate it over all the other blades using periodic boundary conditions.

Here is the full blade row
all blades.png

Here is the single blade
single blade.png

Here is the domain, where the arrows define the inlet and outlet.
domain details.jpg
The pitchwise planes (top and bottom planes in view) with purple rotating pairs is where I can either specify rotational or periodic boundary conditions. Basically, whatever comes out of bottom plane goes inside the top plane and vice versa.

[ghorrocks]

Why would you apply translational periodicity to those faces? They are not translationally periodic. They are rotationally periodic, so that is the only applicable condition for those faces.

Back to your original question from post #1- rotational periodicity is the same as translational periodicity, just with the angle of the face pair taken into account. So stuff which goes in one side comes out the other, but with the velocity vector changed by the angle made by the pair of faces.

I cannot answer your question on post #3 as the axes of the chart are not labelled so I have no idea what you are showing. But I question why you would apply translational periodicity to these faces as they are clearly not translationally periodic.

[pdhyani96]

I didn't realize the plots didn't show the labels. I am plotting static pressure on y-axis versus the normalized span over the inlet surface.

For rotational periodic conditions, I am getting a high static pressure near the inner wall (near 0 on x-axis) as compared to the outer wall (near 1 on x-axis). I was not able to understand why that happens. That is the reason why I applied translational periodic conditions to make sure that this pressure distribution is because of applying rotational periodic conditions.

My guess was right, I am getting a flat line pressure for translational periodic conditions. But, I am still not able to understand that why a rotational periodic boundary condition give me such pressure distribution over the inlet?

[ghorrocks]

Please be very careful to define anything you refer to. None of your images has shown what the x-axis is defined as. I think I know what you mean, but you really should be careful to define what you are talking about.

If I understand your comment correctly, then the total pressure inlet you are using could combine with the rotation (higher velocity at the outside) to result in low static pressure at the outer edge and higher static pressure at the inner edge. This would only occur if the rotation velocity was far greater than the fluid velocity. Could this explain what you are seeing?

[pdhyani96]

I will be more careful next time. But you did figure out my question correctly.
Although I am simulating a steady state flow over a stator, so it is not rotating. So I don't understand why is there a higher velocity at the outer wall.

[ghorrocks]

Have you done a validation and verification? As described here: https://www.cfd-online.com/Wiki/Ansy..._inaccurate.3F

There is no point in analysing results which are not accurate. The big ones people keep forgetting is convergence tolerance and mesh size checks.

If you have done a V&V and you are confident the results are accurate then the only thing to do is to look at the results in detail and try to work it out. There is no inherent reason I can think of why you would see what you are seeing. So there will be something special about the way you set it up.