Fluids

Fluids

convergence problem with k and epsilon

Tagged: 

    • kailash
      Subscriber
      Whenever my residual spikes up, the k and epsilon spikes the most. Is there anything that I can do to prevent this spike?n
    • Surya Deb
      Ansys Employee
      Hello, nWhat is the application that you are solving? Also please check the region where you have excess TKE getting produced. Do you think you need mesh refinement in those regions? Check your wall Y+ values as well.nFurther more is this Standard k-epsilon with Standard wall functions? Did you also try the option Production Limiter?nIf you have very wide range of wall Y+ variations, then you could try and use Scalable Wall Functions as well as it makes sure to scale the mesh virtually in regions where the wall Y+ is low. This ensures that the logarithmic profile is maintained.nBut again this depends on the type of application, the mesh resolution, the boundary conditions specially in terms of the turbulence parameters.nRegards,nSuryan
    • kailash
      Subscriber
      Hi. Yes I have wide variations in wall Y+. My application is studying how the mixing performance is when gas is injected into a large tank (about 100,000 gallons). I am not concerned about any drag or lift or flow separation, so I am using k-e model only with scalable wall function. nI haven't tried Production limiter, what is that? can you please tell more about it?nThanks and Regards,nKailashn
    • kailash
      Subscriber
      I just got to know about the production limiter. I am getting a default value of 10. Should i just leave it as such and run? What does it exactly does? does it limit the upper limit of the epsilon value when those spikes occurs?nRegards,nKailash R.n
    • Surya Deb
      Ansys Employee
      Hello, nYou have a couple of options like Production Limiter and Production Kato-Launder. RANS equations consider a very high degree of correlation of Reynolds Stress Tensor with Strain Rate Tensor. This is why near stagnation regions, where the velocity suddenly goes to zero, there can be very high strain rates and velocity gradients. So RANS models can predict very large K values in these regions whereas it should actually decrease the production of K near those regions due to velocity being very small.nThe Production limiter works on that and calibrates the local level of the production of K term. This is also based on the local levels of dissipation. 10 is the default value but you can change it as well.nProduction Kato-Launder on the other hand utilizes a production limiter term that employs vorticity rates as well. nDo you see these high values of K near the stagnation point where the jet impinges on the tank wall?nRegards,nSuryan
    • kailash
      Subscriber
      You are perfectly right. I am getting high values at the stagnation point only. Thanks for explaining the reason behind it.nRegards,nKailash R.nn
Viewing 5 reply threads
  • You must be logged in to reply to this topic.