Large eddy simulation of a coaxial jet with a synthetic turbulent inlet

Abstract

This study analyzed the effects of two methods of synthetic inlet turbulence on the accuracy of coaxial, circular jet simulations with regards to experimental data. The two methods that were utilized in the study were a digital filter method and a synthetic eddy method. This study examines the implementation of these methods into an academic LES code, with extensive evaluation of the simulation data compared to experimental results. The results in the paper are presented for two different mesh resolutions, 22 and 8 million cells. The simulations were performed for three different velocity ratios of the annular to the inner jet: 0.0, 1.0 and 1.5. The inlet methods were utilized in two ways, one by setting the velocity fluctuation profile based on experimental data of the inlet flow, the second by setting an estimated isotropic-fluctuation profile. In this way, using the estimated method can be compared to using the prescribed fluctuations, which may not be usable if experimental data of the inlet is not available. When comparing the centerline profiles to the velocity data for the turbulent inlet, the data for velocity ratios 1.0 and 1.5 were improved compared to the experimental profiles. However, for the velocity ratio 0.0 case velocity data when using the turbulent inlet the data showed significant decay and did not match experimental profiles in the far field of the jet. The radial data for a velocity ratio of 0.0 showed better agreement with the experimental data in the jet spreading to develop the proper velocity profile. The velocity fluctuation data was the most accurate in the near field region when the turbulent inlet was used. A comparison can also be made between using a prescribed fluctuation inlet and just using the approximated

Budgets of Turbulent Kinetic Energy and Reynolds Normal Stresses in Coaxial Jets with and Without Swirl: A Numerical Study

“…2(b)). Correspondingly, the streamwise velocity decreases, which is also observed in other studies [Buresti et al, 1998;Rehab et al, 1997;Abboud and Smith, 2014]. Then a negligible slope for the streamwise velocity is observed until x/D ≤ 3.0, which also marks the end of inner jet potential core.…”

“…These methods affect the distribution of mean velocity components and hence can alter the vortex behavior [Lucca-Negro and O'doherty, 2001]. Apart from the mean velocity, the correlations between the instantaneous velocity components are also vital [Abboud and Smith, 2014]. Therefore, an attempt is made to improve the inflow boundary condition by using precursor simulations of jet nozzles in advance of the DNS simulations.…”