Method to Predict Circulation Control Noise

Underwater vehicles suffer from reduced maneuverability with conventional lifting appendages due to the low velocity of operation. Circulation control offers a method to increase maneuverability independent of vehicle speed. However, with circulation control comes additional noise sources, which are not well understood. To better understand these noise sources, a modal-based prediction method is developed, potentially offering a quantitative connection between flow structures and far-field noise. This method involves estimation of the velocity field, surface pressure field, and far-field noise, using only non-time-resolved velocity fields and time-resolved probe measurements. Proper orthogonal decomposition, linear stochastic estimation and Kalman smoothing are employed to estimate time-resolved velocity fields. Poisson's equation is used to calculate time-resolved pressure fields from velocity. Curle's analogy is then used to propagate the surface pressure forces to the far field. This method is developed on a direct numerical simulation of a two-dimensional cylinder at a low Reynolds number (150). Since each of the fields to be estimated are also known from the simulation, a means of obtaining the error from using the methodology is provided. The velocity estimation and the simulated velocity match well when the simulated additive measurement noise is low. The pressure field suffers due to a small domain size; however, the surface pressures estimates fare much better. The far-field estimation contains similar frequency content with reduced magnitudes, attributed to the exclusion of the viscous forces in Curle's analogy. In the absence of added noise, the estimation procedure performs quite nicely for this model problem. The method is tested experimentally on a 650,000 chord-Reynolds-number flow over a 2-D, 20% thick, elliptic circulation control airfoil. Slot jet momentum coefficients of 0 and 0.10 are investigated. Particle image velocimetry, unsteady pressure and phased-acoustic-array data are acquired simultaneously in an aeroacoustic wind-tunnel facility. The velocity field estimation suffers due to poor correlation with the unsteady pressure data, especially in the 0.10 momentum coefficient case. The prediction without slot jet blowing matches single microphone measurements within 0-10 dB over the frequency range of interest while the prediction with the jet active is quite poor and differ from measurements by as much as 35 dB. Suggestions for improvement of the proposed method are offered. Data from the acoustic array are then investigated. Single microphone spectra are obtained, and it is shown that background noise is significant. In order to circumvent this problem, beamforming is employed. The primary sources of background noise are from the tunnel collector and jet/sidewall interaction. DAMAS is employed to remove the effects of the array point spread function. Spectra are acquired by integrating the DAMAS result over the source region. The resulting DAMAS spectral levels are significantly below single microphone levels. A scaling analysis is performed on the processed array data. With a constant free-stream velocity and a varying jet velocity the data scale as M6. If momentum coefficient is held constant and free-stream velocity is varied the data scale as M7.