Increase in the use of composites for aerospace applications has the benefit of decreased structural weight, but at the cost of decreased acoustic performance. Stiff, lightweight structures (such as composites) are traditionally not ideal for acoustic insulation applications because of high transmission loss at low frequencies. A need has thus arisen for effective sound insulation materials for aerospace and automotive applications with low weight addition. Current approaches, such as the addition of mass law dominated materials (foams) also perform poorly when scaled to small thickness and low density. In this dissertation, methods which reduce sound transmission without adding significant weight are investigated. The methods presented are intended to be integrated into currently used lightweight structures such as honeycomb sandwich panels and to cover a wide range of frequencies.

Layering gasses of differing acoustic impedances on a panel substantially reduced the amount of sound energy transmitted through the panel with respect to the panel alone or an equivalent-thickness single species gas layer. The additional transmission loss derives from successive impedance mismatches at the interfaces between gas layers and the resulting inefficient energy transfer. Attachment of additional gas layers increased the transmission loss (TL) by as much as 17 dB at high (>1 kHz) frequencies. The location and ordering of the gasses with respect to the panel were important factors in determining the magnitude of the total TL. Theoretical analysis using a transfer matrix method was used to calculate the frequency dependence of sound transmission for the different configurations tested. The method accurately predicted the relative increases in TL observed with the addition of different gas layer configurations.

To address low-frequency sound insulation, membrane-type locally resonant acoustic materials (LRAM) were fabricated, characterized, and analyzed to understand their acoustic response. Acoustic metamaterials with negative dynamic mass density have been shown to demonstrate a significant (5×) increase in TL over mass law predictions for a narrow band (100Hz) at low frequencies (100–1000Hz). The peak TL frequency can be tuned to specific values by varying the membrane and mass properties. TL magnitude as a function of frequency was measured for variations of the mass magnitude and membrane tension using an impedance tube setup. The dynamic properties of membranes constructed from different materials and thicknesses were measured and compared to the results of coupled field acoustic-structural finite element analysis (FEA) modeling to understand the role of tension and element quality factor. To better comprehend the mechanism(s) responsible for the TL peak, a laser vibrometer was used to map the out-of-plane dynamic response of the structure under acoustic loading at discrete frequencies. Negative dynamic mass was experimentally demonstrated at the peak TL frequency.

The scale-up of the acoustic metamaterial structure was explored by examining the behavior of multiple elements arranged in arrays. Single membranes were stretched over rigid frame supports and masses were attached to the center of each divided cell. TL behavior was measured for multiple configurations with different magnitudes of mass distributed across each of the cell membranes in the array resulting in a multi-peak TL profile. To better understand scale-up issues, the effect of the frame structure compliance was evaluated, and more compliant frames resulted in a reduction in TL peak frequency bandwidth. In addition, displacement measurements of frames and membranes were performed using a laser vibrometer. The measured TL of the multi-celled structure was compared with TL behavior predicted by FEA to understand the role of non-uniform mass distribution and frame compliance.

TL of membrane-type LRAM with added ring masses was analyzed using both finite element analysis and experimental techniques. The addition of a ring mass to the structure either increased the bandwidth of the TL peak, or introduced multiple peaks, depending on the number of rings, the distribution of mass between the center and ring masses, and radii of the rings. FEA was used to predict TL behavior of several ring configurations, and TL for these configurations was measured to validate the model predictions. Finally, FEA was used to predict the mode shapes of the structure under single-frequency excitation to understand the mechanisms responsible for the TL peaks.