[wjarrah]

Hello everyone,

I have to study the k-e model, i would appreciate it if you could help me with these questions :
- When do we use it ? ( It is adapted to which situations?)
- Pros/Cons
- What do these constants 'C_mu, 6_epsilon, C1e, C2e' match correspond to ?
- When do we have to change them ?
- What are the conditions to have good results? (y+ = 1 in the first cell ?!)
- And if possible, a comparison with the K-Oméga SST, V²_f et Reynolds Stress models ?

Thanks a lot !

[Charles Rundle]

Hello,

I'm afraid I'm not sure how familiar you are with turbulence models so if I repeat what you know please forgive me.
The k-epsilon turbulence model is a two equation eddy viscosity model similar to the k-omega and SST models. Actually the SST model is a blending of the k-epsilon and k-omega models designed to minimize the weakness of both models. That's not the only thing that makes SST different but it is the most relevant to your question.

My understanding is that the k-epsilon model is best in free flows away from the walls. It also has a tendency to model separation
flows poorly and can over predict turbulence in stagnant areas. The other limitations of eddy viscosity models also apply. That being said it is possible to adapt to other situations.

With regard to the constants they can be adjusted to correct for different types of flow. If you’re interested in a specific type of flow I recommend searching academic journals. You should be able to find articles that will help you with the correct values for constants.

It would be helpful to know which k-epsilon model you are considering. i.e. is a Low Re model or are you using a wall function.

I should mention that while I believe everything I have presented is correct I do not guarantee it as such or accept any liability that may arise from its use. It is entirely my opinion and you use it at your own risk.

Hope I was helpful. Have a good day.

CAR

[harishg]

The constants in the model depend on the test cases that have been used to obtain them. Depending on the choice of the authors different models consider different test cases. Hence you would find slight differences in the model constants for the same k-e model. Usually you use y+<=1 to ensure the grid independent solution. You need to ensure that you use a low-Re model if you integrate through the viscous sub layer. Regarding the models that you are considering as k-omega SST and RSM have been shown to perform much better than k-epsilon models for many test cases. However, recently many new k-e models have been proposed which may be as good as these two models.

[wjarrah]

Thanks Charles and Harishg !

[Charles Rundle]

Hello,

You are most welcome.

Have a good day.

CAR