[Ivan]

Dear All,

I am simulating a round free jet spreading in coflow air. Since I have not the pipe inside the domain, I need to generate correlated fluctuations at inlet in a way they are not dumped immediately after the inlet. For the purpose I use an algorithm from Klein (J. of Comp. Physics 186 pp652-665). Nevertheless, I have the opposite problem:
Fluctuations increase too quickly along the centerline and, consequently, the mean velocity decrease too quickly as well. Without the correlated fluctuations, the data muches pretty well the experiments instead (apart for the first few diameters, since fluctuations would start from zero in that case).

I am using a dynamic Smagorinsky model. Does someone have any clue of the reason of this behavior? May it be due to the kind of correlations I'm giving in input (i.g. does the length of the turnover and the integral length scale associated with the fluctuations I give, have an influence? and how?). Or should I look to the way fluctuations are generated by the smagorinsky algorithm?


Hope someone can help!