Overset Grid Capability Research Articles

High-Fidelity Numerical Simulation of a Chevron Nozzle Jet Flow

We report results from the simulation of a moderate Reynolds number cold jet flow exhausting from a chevron nozzle with six symmetric chevrons that have an approximately 18-degree penetration angle. The flow inside the nozzle geometry and the free jet flow outside are computed simultaneously by a high-order accurate, multi-block, large eddy simulation code with overset grid capability. The simulation is performed on 400 million grid points using 2048 processor cores in parallel. The main emphasis of the simulation is to compute the jet flow in maximum detail possible and accurately capture the physical processes that lead to noise generation. It is especially critical to capture the enhanced shear layer mixing due to chevrons that takes place within the first few diameters downstream of the nozzle exit. Our calculations resolve the jet flow field at an unprecedented level of detail. Despite some issues such as an order of magnitude lower simulation Reynolds number (relative to experimental value) and unknown turbulence intensity levels within the experimental nozzle, it is shown that both the near-field jet turbulence and far-field noise predictions are in good agreement with the experimental measurements. Because of the enormous number of grid points required to resolve the near-nozzle region, the computational domain size is constrained to ten nozzle exit diameters downstream of nozzle exit by the limited computational resources. As a result, not all of the low frequency generating noise sources are resolved in the calculation, resulting in some errors in the low frequency range of the predicted noise spectrum. Nevertheless, it is shown that the high frequency noise generation in the near-nozzle region is predicted quite well.

PREDICTION OF NOISE GENERATED BY A ROUND NOZZLE JET FLOW USING COMPUTATIONAL AEROACOUSTICS

This paper demonstrates an application of computational aeroacoustics to the prediction of noise generated by a round nozzle jet flow. In this study, the nozzle internal flow and the free jet flow outside are computed simultaneously by a high-order accurate, multi-block, large-eddy simulation (LES) code with overset grid capability. To simulate the jet flow field and its radiated noise, we solve the governing equations on approximately 370 million grid points using high-fidelity numerical schemes developed for computational aeroacoustics. Projection of the near-field noise to the far-field is accomplished by coupling the LES data with the Ffowcs Williams–Hawkings method. The main emphasis of these simulations is to compute the jet flow in sufficient detail to accurately capture the physical processes that lead to noise generation. Two separate simulations are performed using turbulent and laminar inflow conditions at the jet nozzle inlet. Simulation results are compared with the corresponding experimental measurements. Results show that nozzle inflow conditions have an influence on the jet flow field and far-field noise.

Investigation of high frequency noise generation in the near-nozzle region of a jet using large eddy simulation

This paper reports on the simulation of the near-nozzle region of an isothermal Mach 0.6 jet at a Reynolds number of 100,000 exhausting from a round nozzle geometry. The flow inside the nozzle and the free jet outside the nozzle are computed simultaneously by a high-order accurate, multi-block, large eddy simulation (LES) code with overset grid capability. The total number of grid points at which the governing equations are solved is about 50 million. The main emphasis of the simulation is to capture the high frequency noise generation that takes place in the shear layers of the jet within the first few diameters downstream of the nozzle exit. Although we have attempted to generate fully turbulent boundary layers inside the nozzle by means of a special turbulent inflow generation procedure, an analysis of the simulation results supports the fact that the state of the nozzle exit boundary layer should be characterized as transitional rather than fully turbulent. This is believed to be most likely due to imperfections in the inflow generation method. Details of the computational methodology are presented together with an analysis of the simulation results. A comparison of the far field noise spectrum in the sideline direction with experimental data at similar flow conditions is also carried out. Additional noise generation due to vortex pairing in the region immediately downstream of the nozzle exit is also observed. In a second simulation, the effect of the nozzle exit boundary layer thickness on the vortex pairing Strouhal frequency (based on nozzle diameter) and its harmonics is demonstrated. The limitations and deficiencies of the present study are identified and discussed. We hope that the lessons learned in this study will help guide future research activities towards resolving the pending issues identified in this work.

Large-Eddy Simulation of a Wing Tip Vortex on Overset Grids

A multiblock large-eddy simulation (LES) code with overset grid capability is reported on that has been developed primarily to study the tip vortex noise problem arising on rotating wind-turbine blades. The LES code can also be used to simulate the tip vortices around nonrotating wing geometries typically found in aeronautical applications. It employs state-of-the-art compact finite differencing, implicit spatial filtering, and nonreflecting boundary conditions on free boundaries, as well as characteristic-type boundary conditions on solid walls. High-order-accurate interpolation is used for the transfer of information among the individual component grids that make up the overset grid topology. The multiblock nature and the overset grid capability of the code allow high-order accurate numerical solutions in complex domains. Results for the tip vortex around a nonrotating wing with a rounded tip are presented to demonstrate the capability of the code. Comparisons of the numerical results with experimental data are also carried out.

Parallel Overset-Grid Euler Solution of Generic Wing Pylon and Finned Store.

Introduction S INCE the 1980s, after Steger et al.1 solved twoand threedimensional multibody configurations, overset (chimera) grid systems became very popular because of their advantages in easing the grid generation. Later in 1994, Kao et al.2 used overset grids in order to perform grid adaptation in the flow regions where higher gradients exist. Tuncer3 preferred to use overset grids for unsteady Navier–Stokes solutions of sinusoidally pitching airfoil. Cho et al.4 solved the generic wing/pylon/finned store (WPFS) configuration as a steady three-dimensional problem by using the overset grids. Later Lee et al.5 used overset grid system for the solution of a store release problem. Furthermore, Holst6 also developed a new search algorithm for locating the donor cells associated with the overset grid system used in the solution of wing-fuselage geometry by full potential equations. In the present work, a three-dimensional finite differencing scheme, multiblock structured Euler and thin-layer Navier–Stokes7 solver is modified in order to enhance overset grid capability and to run in a parallel-computing environment. A wing-store configuration and Eglin WPFS8 configuration are selected as the test cases for the validation purposes.