[Norflow]

Hey community!

Firstly, I would like to say what a great resource this is and how thankful I am to be a part of it. I am a junior mechanical engineer, and aspiring hydraulic engineer. I am relatively new to CFX software but have a good foundation of theory behind fluid dynamics. I will continue to post my findings and questions and hopefully they can be of help to others that have relative issues.

I have a multi-domain, steady state simulation with 3D geometry of a water passage (one domain - stationary) imported from SW and inside this geometry is a 5 blade kaplan turbine (second domain - rotating). It has meshed well and I have specified inlet (mass flow-rate) and outlet (static pressure) conditions. I am interested in any areas of cavitation and a preliminary efficiency that I can compare to a physical model test that was performed.

I am wondering, initially, is this the right approach? I am trying to model a complete water passage from intake to draft tube with a rotating turbine inside it. However, the walls are also part of the rotating domain and I'm not sure if that's correct to simulate. Should I have a counter-rotating wall in this domain?

Any help is more than appreciated!

Thanks

[ghorrocks]

There is a FAQ in the CFX documentation on rotating machinery which will help you. Have you read it? That would be the best place to start.

[Norflow]

@Ghorrocks,

I have read most of the User Guides for Mesh, Pre, Solver, and Post. While all of them are great resources, tutorials and examples don't deal with large area simulations or only deal with water-to-blade passages in a symmetric nature. Also, I am not using turbo mode, as I have modeled only the water passage as a single part by taking existing geometry and creating a internal "mold" of the cavity.

I have performed a simulation and had convergence but I don't believe my results are accurate. I have attached a .jpeg of the velocity streamline through the water passage. My concern is that my stationary domains aren't properly effected by my rotating domain. Any ideas?

[ghorrocks]

What is not properly affected by it?

[Norflow]

Firstly, the blades are spinning in the wrong direction, which is quite evident from the uploaded jpeg image, and my next simulation will be corrected. What I am concerned with is the fact that there is no residual swirl in streamlines as the flow leaves the rotating domain and a sudden change in velocity. Why is that?

[ghorrocks]

the blades are spinning in the wrong direction

Looks OK to me. I think you are confused by the frames of reference. Plot velocity vectors using "velocity in stationary frame". Then you will see what you expect.

Should write an FAQ on this, this question has been asked many times.

there is no residual swirl in streamlines

For low losses good turbomachinery design tries to minimise swirl in the outlet. This suggests your design is functioning well! Operate it off the design point and I bet you will get lots of swirl.

[Norflow]

Okay, I'll give it a try. Thanks so much for your valuable feedback, I really appreciate it.