Hi,

I am running a 3D simulation, k-e RNG, steady state, tetraheral grids, air medium with no rotating boundary.

In solving the model, I am taking a step-by-step

laminar -> std k-e -> RNG k-e

Managed to get convergence by monitoring the outlet velocity and temperature. However, upon convergence, if we let the model to continue iterating, once in a while the error message below will appear

turbulent viscosity limited to viscosity ratio of 1.000000e+005 in 1 cells

This happens once approximately every 200-250 iterations. A few questions on this:

1. Even with this occurrence, is my solution valid? 2. The error is only in 1 cell. What can be done to alleviate this? The solution mentioned is already using very fine grids at the region of high velocity gradient

Hope to further discuss this. Thank you.

Regards, Woo

hi,

u should not get that kind of message anyway..

don t know why u have used laminar option for a turb. problem. no need to do like that. u can change over the options of k-e . if u r not getting that message for 100-200 iterations u can trat it as converged. the error msg comes b coz of improper b.c s for turbulence. o.k

Thanks RAvi,

Actually I do get that error message in the beginning. Like what you mentioned, it tapers off after a few hundred iterations. However, once the solution converges, the error message will appear once in a while, approximately every 250 - 300 hundred iterations.

As it is already converged and does not happend all the time, don't think it is because of inappropriate boundary conditions. Furthermore, it only occur in one cell. I am usinga total of 600,000 cells.

Any idea?

See it is frequent problem with turbulence modelling with tetrahedral mesh and second order schemes for turb. dissipation rate ....

Ideally u should not get any messages ....

check where this cell is....may be u can improve mesh quality in this perticular area(like skewness, reduce mesh size, reduce mesh size jump etc......)

or u can use first order scheme for turb. dissipation rate equation and second order for rest equations.....

Most important thing...if ur solution looks physically right then ur solution is always right......and check/measure ur physically important quantity in domain....if they are right then ur sinulations are always right