High turbulence intensity
I am simulating a valve problem using rke turbulence model. Inlet velocity is 17m/s and diameter is 0.054m while outlet diameter is 0.043m. The Re number is calculated based on inlet condition which is about 90,0000. The velocity BC is imposed on inlet with 5% turbulence intensity and hydralic diameter. Pressure BC is imposed on outlet with 5% turbulence intensity.
The solution is converged to 1.e6 (residual) and I also monitor a pressure at a selected location which converges too. The inlet turbulence kinetic energy (areaweighted average) reported by Fluent is 1.12m2/s2 from which I compute turbulence intensity of around 5%. However, the areaweighted average of inlet turbulence intensity reported by Fluent is 83% at inlet and even higher at outlet which I assume is unphysical. Some posts in this forum suggested to check the reference value for turbulence intensity computation, what does that mean and where to check the reference value? Thanks! 
I had the exact problem before. I simulated a heat rejection problem using the rke model, which gave me an overestimated turbulence intensity, as high as 50%, and also consequently an overestimated effective fluid thermal conductivity. By looking at the high turbulence intensity level and the unrealistic temperature profile, it can be tell the model is not doing well.
I switched to the komiga model as FLUENT people suggested, and all results look fine to me. Even though, I don't know if the high turbulence intensity is a result of 'the reference value for turbulence intensity computation' as you mentioned. Or, it's because the rke model is not doing well. I'd appreciate if any of you guys can give me an answer. Quote:

All times are GMT 4. The time now is 20:42. 