An alternative pressure-sensor based method for estimating the acoustic intensity, the phase and amplitude gradient estimation (PAGE) method, is presented. This method is similar to the finite-difference p-p (FD) method, in which the intensity is estimated from pressure measurements made using an array of closely spaced microphones. The PAGE method uses the same hardware as the FD method, but does not suffer from the frequency-dependent bias inherent to the FD method. Detailed derivations of the new method and the traditional FD method are presented. Both methods are then compared using two acoustic fields: a plane wave and a three monopole system. The ability to unwrap the phase component of the PAGE method is discussed, which leads to accurate intensity estimates above previous frequency limits. The uncertainties associated with both methods of estimation are presented. It is shown that the PAGE method provides more accurate intensity estimates over a larger frequency bandwidth. The possibility of using a higher-order least-squares estimation with both methods is briefly demonstrated. A laboratory experiment designed to validate the PAGE method was conducted. The preliminary results from this experiment are presented and compared to analytical predictions. Finally, the application of the PAGE method to a static rocket test firing is presented. The PAGE method is shown to provide accurate intensity estimates at frequencies that are higher than possible with just the FD method.

Microphone arrays are useful for measuring acoustic energy quantities (e.g. acoustic intensity) in the near-field of a full-scale solid rocket motor. Proper characterization of a rocket plume as a noise source will allow for more accurate predictions in engineering models that design for protection of structures, payloads and personnel near the rockets. Acoustic intensity and energy density quantities were measured in three rocket noise fields and have shown that the apparent source region of the rocket becomes smaller and moves upstream as frequency increases. Theoretical results accounting for some scattering and finite-difference errors arising in these types of energy-based measurements have been previously discussed by other authors. This thesis includes results from laboratory experiments which confirm some of this previous theoretical work as well as gives insight into the physical limitation of specific microphone array designs. Also, calibrations for both magnitude and directional response of the microphones are demonstrated. Of particular interest is the efficacy of phase calibration of array microphones for the low-frequency regime below 200 Hz.

Near-field acoustical holography (NAH) has been used extensively for acoustical imaging of infinitesimal-amplitude (or small-amplitude) sources. However, recent interests are in the application of NAH to image finite-amplitude (or high-amplitude) sources such as jets and rockets. Since NAH is based on linear equations and finite-amplitude sources imply nonlinear effects, which cause shock formation and consequently an altered spectral shape, a feasibility study is carried out to determine the effect of nonlinear propagation on NAH. Jet and rocket sources typically have a distinct spectral shape resembling a ‘haystack’ and center frequencies varying from 30 to 300 Hz. To test the effect of nonlinear propagation on jet or rocket noise, several waveforms with varying spectral shapes and center frequencies were created and numerically propagated in one dimension using a nonlinear propagation algorithm. Bispectral methods were used to determine the amount and effect of nonlinearity, showing that higher center frequencies lead to more nonlinearities for a given amplitude. Also, higher-order statistical analysis of the time derivative of the waveforms was used to determine information about the relative amount of waveform steepening and shock coalescence occurring. NAH was then used to reconstruct the original waveform magnitude and the errors were determined. It was found that the ‘haystack’ spectral shape can be preserved by the nonlinear effects leading to low amplitude-reconstruction errors, whereas a narrow-band spectral shape will become altered and reconstruct very poorly. However, if nonlinear effects become strong due to higher center frequencies, longer propagation distances or higher amplitudes, even the ‘haystack’ shape will become altered enough to cause poor reconstruction. Two-dimensional propagation studies were also performed from two point sources, showing differences between linear and nonlinear propagation.