High resolution simulations of high Reynolds number jets with microjet injection M.E. Rife Gary Page 2134/22504 https://repository.lboro.ac.uk/articles/conference_contribution/High_resolution_simulations_of_high_Reynolds_number_jets_with_microjet_injection/9223133 Large eddy simulations have been performed of a Mach 0.9 jet at a Reynolds number of 1.3 million for both a clean and microjet injected configurations. Two numerical grids were used for the simulations differing in the number of azimuthal cells. The first with 720 cells and an azimuthal clustering near the microjet injection locations, and the other with 1440 cells and a uniform azimuthal cell spacing that matches the finest cell for the clustered case. The grids contained 100 million and 200 million cells respectively. A standard Smagorinsky sub-grid scale model was used together with a synthetic trip in the nozzle shear layer. The non-uniform grid with 720 cells azimuthally showed a variation in laminar to turbulent transition location that was a function of the clustering, with later transition in the coarser regions. However, this had little detrimental impact on mean velocity distributions further downstream. The results of the simulations were compared with PIV experimental data and good agreement of mean radial velocity and turbulent kinetic energy profiles were obtained. The microjets caused a deformation of the shear layer, reducing the radial location of peak turbulence kinetic energy in-line with the microjets. Additionally, the shear layer is translated away from the jet centreline between the microjets and becomes flat in the regions between the microjets. A Ffowcs Williams-Hawking technique was used to propagate the unsteady pressure fluctuations to the far-field. Spectral data at downstream and sideline observer locations indicated the presence of a high-frequency peak in the microjet case which is consistent with the size of the microjets. The microjets provide a blockage effect to the main jet and the peak is probably a shedding like behaviour. Overprediction of overall sound pressure levels by 6-8 dB was found when compared to the experimental data, however, the correct behaviour with observer angle was captured and more importantly the microjets showed a reduction in OASPL of around 2 dB across a range of angles, similar to the experiment results. 2016-09-20 11:07:50 untagged Engineering not elsewhere classified