Search

The current effort characterizes the details of flow interactions and wake interference effects between two cylinders in a tandem configuration. This setup is representative of several component-level flow interactions that occur when air flows over the main landing gear of aircraft. Such interactions are likely to have a significant impact on the noise radiation associated with the undercarriage. This paper focuses on two-dimensional, time-accurate flow simulations of two distinct tandem cylinder flow regimes, associated with short and intermediate separation distances between the two cylinders. Unsteady Reynolds averaged Navier-Stokes simulations using a two-equation turbulence model run at a Reynolds number of 1.66 x [10.sup.5] and a Mach number of 0.166 are presented. Emphasis is placed on understanding both the time-averaged and unsteady flow features between the two cylinders and in the wake of the rear cylinder. Predicted mean-flow quantities and vortex shedding frequencies show reasonable agreement with measured data for both cylinder spacings. Computations for the short separation distance exhibit a nonphysical decay of flow unsteadiness with time: however, the predicted sensitivity of the mean lift coefficient to small variations in the upstream flow angularity explains the asymmetric flowfield observed in the present and previous measurements.

Results demonstrate that high-amplitude, high-frequency acoustic signal, similar to those obtained in high-lift system, is caused by vortex shedding from the trailing-edge slat. The acoustic analysis is based on the Ffowcs Williams and Hawkings equation.

The steady flow calculations over a finite wing were performed using the high-order spatial operators of a computational aeroacoustics (CAA) scheme, as well as a time-marching algorithm. The equations of motion in three dimensions are provided, and there is emphasis on three-dimensional results for low-speed flow over a NACA 0012 wing. Vortex growth was larger than in experimental measurements, although other parameters were well predicted.

A research was conducted to determine the nature of broadband blade noise and determine the reliability of utilizing computational aeroacoustics procedures and parallel computers to further understand broadband blade noise. The two dimensional Navier-Stokes code was utilized in calculating numerical data. Results revealed that computational aeroacoustics techniques are reliable when applied in more realistic problems and that multigrid procedures and parallel computers can be used to obtain realistic mean flows.

High-order and high-bandwidth algorithms, which can function efficiently even in the absence of fine grids, provide accurate solutions for computational aeroacoustics problems. Design of high-bandwidth operators needs to be performed in an appropriate manner to mitigate possible errors in computations involving longer wavelength.

The slat of a high-lift wing can be a significant noise source during approach and landing. This paper describes an experimental study of the effects of two passive devices (namely, slat extensions and a cove filler) on the aerodynamics and acoustic radiation from a 30P30N model. Test conditions include three effective free-air angles of attack of 5.5, 7.5, and 13 deg, as well as three Reynolds numbers of 1.2, 1.5, and 1.71×106. Steady surface pressure measurements are used to assess the aerodynamics near the midspan section. Unsteady surface pressure and far-field acoustic array measurements are performed to evaluate the near- and far-field pressure fluctuations, respectively. Delay-and-sum beamforming is applied to localize the noise sources and to provide integrated spectra. The slat cusp extensions barely affect the mean Cp distribution, whereas the longest extension achieves up to a 5.4 dB reduction in the overall sound pressure level. The slat-cove filler yields an even higher noise reduction that is lower than the baseline by up to 14 dB. Even though the aerodynamic assessment of the cove filler is incomplete due to the covered pressure ports, the mean Cp distribution over the remaining surface of the model indicates a rather small change compared with the baseline measurements. [ABSTRACT FROM AUTHOR]