Volume 81, Issue S1, May 1987
Index of content:
- PROGRAM OF THE 113TH MEETING OF THE ACOUSTICAL SOCIETY OF AMERICA
- Session QQ. Physical Acoustics VII: Propagation Phenomena
- Contributed Papers
81(1987); http://dx.doi.org/10.1121/1.2024088View Description Hide Description
The acoustic radiation produced by an infinite flat plate with sinusoidally varying properties and driven by a distributed load is presented. The plate is assumed to be thin with constant thickness, but with stiffness varying sinusoidally along the axis. The analysis is an extension of the work previously reported [M. Pierucci, J. Acoust. Soc. Am. Suppl. 1 79, S35 (1986)]. The analysis consists of solving the coupled fluid‐structure equations with the added difficulty of having nonconstant coefficients. The system reduces to a convoluted equation, which has been solved analytically. The results indicate the presence of acoustic pressure components radiating in directions that are related to the difference between the forcing function wavenumber and the wavenumber of the stiffness variation. Acoustic radiation patterns will be presented.
- Session D. Underwater Acoustics I: Propagation of Underwater Sound: Range‐Dependent, Three‐Dimensional, and Nonlinear Media
81(1987); http://dx.doi.org/10.1121/1.2024089View Description Hide Description
Leaky Rayleigh waves propagating along the bottom side of floating sea‐ice plates experience cylindrical geometrical spreading and have their low‐frequency limit determined by the ice plate thickness that corresponds to about 560 Hz. A second type of leaky Rayleigh wave can exist along ice plate edges propagating with no geometrical spreading and has its low‐frequency limit determined by the large plate horizontal dimensions. Laboratory ultrasonic modeling results are presented demonstrating the detection of low‐frequency Rayleigh waves radiated from ice plate edges. In one experiment, the plate‐edge leaky Rayleigh wave wavelength was greater than 12 times the ice plate thickness. The results suggest that a 3‐m‐thick ice plate in the Arctic may easily radiate 10‐ to 100‐Hz components from its edges. The ice plate edge acting nearly as a finite length line source produces high directionality of the radiated leaky Rayleigh wave into the water. The low‐frequency plate‐edge leaky Rayleigh waves may be detected in the water at significant distances from the ice edge. Additional physical insights into the controversial leaky Rayleigh wave are provided by comparing water/ice and Freon‐113/ice interface conditions where a high‐density low‐velocity liquid is used. [Work supported by ONR.]
81(1987); http://dx.doi.org/10.1121/1.2024090View Description Hide Description
The nonlinear progressive wave equation (NPE) model has been used to investigate the propagation of linear and nonlinear broadband signals in a range‐dependent shallow water waveguide. The NPE is the nonlinear time domain equivalent of the frequency domain parabolic wave equation (PE). Results from each of the studies are given in short computer‐generated movies illustrating a four‐cycle sine‐wave pulse propagating a total of 20 km in isovelocity water (c = 1500 m/s) over a bottom (c = 1600 m/s) whose depth profile contains a shelf of slope 0.15 connecting regions of constant depth 225 and 150 m. Effects evident in the movies are: radiation into the bottom resulting in head waves that re‐radiate back into the water column; energy dumping following reflection from the sloping bottom and surface; and mode separation at late times. Early time results suggest that bottom penetration is greater when nonlinearity is included.
81(1987); http://dx.doi.org/10.1121/1.2024091View Description Hide Description
A parabolic equation model requires a starting field; that is, the values of the field must be given as a function of depth at a fixed range. A desirable starting field must satisfy two requirements: (1) It should excite the propagating modes with the proper amplitudes; and (2) it should excite nonpropagating modes only to the extent that their effects can be filtered out with increasing range by the false bottom usually added under the ocean floor, which can more than double the depths involved. In practice, the usual approach is to use an easily computed result such as a Gaussian, filtered Gaussian, sine, or uniform ocean starting field. The false bottom is then increased to meet the second criterion. The advantage of the normal mode starting field is that it does not excite unwanted nonpropagating modes, which minimizes the need for the false bottom, which in turn reduces the running time of the model proportionally—once the starting field is computed.
81(1987); http://dx.doi.org/10.1121/1.2024092View Description Hide Description
A particular shallow‐water range‐dependent environment was chosen and the stepwise coupled mode model COUPLE was applied. The problem consisted of a refracting water layer overlying an isospeed liquid sediment. The water depth varied between 120 and 210 m over a range of 50 km. The calculations were done at 50 and 300 Hz. The main range‐dependent effect considered was the redistribution of energy due to forward and backscatter caused by both large and small scale bathymetic changes. The large scale changes, that evolve slowly with range, produced a smooth transition of energy between different sized sets of locally propagating modes. The more rapid, small scale, changes caused a pronounced energy loss due to forward scattering into highly attenuated high‐angle modes. The effects of backscatter were negligible. The results are compared in this paper to IFD/PE and, in a subsequent paper, with other parabolic equationmodels.
A comparison of parabolic equation models with stepwise coupled modes in a shallow‐water range‐dependent environment81(1987); http://dx.doi.org/10.1121/1.2024093View Description Hide Description
Parabolic equation calculations are compared with stepwise coupled mode COUPLE calculations for a set of runs for which the two models have overlapping capability. The PE calculations were then extended to consider the additional case of refracting sediment. Normal mode start‐up fields were used in the calculations. PE models have capabilities differing in the areas of wide angle capability, treatment of the bottom, and numerical methods. The intent of this paper is to discuss which of these differences make a difference. The environment chosen was that presented in the previous paper. Calculations were performed at 50 and 300 Hz. Water depths varied somewhat, but averaged about 150 m, corresponding to 3 wavelengths and 15 wavelengths at 50 and 300 Hz, respectively. The dependence of depth with range took two forms. The first consisted of smoothly varying linear segments, the second was the same as the first with a number of spikelike irregularities added in. Differences are discussed in terms of numerical issues and the physics incorporated in the models.
81(1987); http://dx.doi.org/10.1121/1.2024094View Description Hide Description
Surface guiding of high‐frequency sound fields may take place near a concave boundary in a homogeneous medium or near a plane boundary in a refracting medium. Multiple reflected ray fields excited by a source near the boundary form caustics that pile up at long ranges around the source depth. Conventional caustic corrections are inadequate to deal with this catastrophe. It is shown here that the narrow angular spectrum interval occupied by the troublesome caustic‐forming rays can be filled alternatively by a narrow‐angle parabolically approximated (PE) wavefield. The hybrid combination of legitimate rays and narrow‐angle PE can be regarded either as correcting via PE the failure of ray acoustics at long ranges near the source depth, or as removing the narrow‐angle restriction from PE by filling the wide‐angle spectra with rays. The theory is illustrated with calculations for a range‐independent ducting environment, and also for a range‐dependent ducting‐to‐antiducting transition. [Work supported by ONR.]
HYPER: A hybrid parabolic equation‐ray model for underwater sound propagation in the vicinity of smooth caustics81(1987); http://dx.doi.org/10.1121/1.2024095View Description Hide Description
The hybrid parabolic equation‐ray model (HYPER) uses modified ray tracing of geometrical acoustics to locate important propagation paths. Then HYPER solves a modified parabolic equation in a small region near the rays to compute the pressure field. This paper applies the HYPER model to investigate acoustic propagation in the vicinity of smooth caustics. Results are compared with ordinary ray theory, modified ray theory, and the wave solution for a range‐independent case. Some range‐dependent examples are also presented. [Funding for this work was provided by the AEAS Program Office, ONR Detachment, Code 132.]
81(1987); http://dx.doi.org/10.1121/1.2024096View Description Hide Description
Fast field computations without range dependence are based on computing the Hankel transform of the Green's function of the separated depth‐dependent ordinary differential equation. As usual, the required Green's function is expressed in terms of two solutions of the homogeneous differential equation, the first satisfying the boundary condition at the surface of the ocean, the second satisfying the boundary condition at the bottom of the ocean, and the Wronskian of these two functions. In our approach, one solution is given in terms of one transmutation that preserves the surface boundary condition, and the other is given in terms of a different transmutation that preserves the bottomboundary condition. Once the two different transmutation kernels are found, they are applied to trigometric functions that satisfy the respective boundary conditions, and the Green's function is numerically computed. A fast Hankel transform (the usual fast Fourier transform approximation was used) completes the calculation of the field versus range.
Numerical solution for the diffraction of internal sound sources by a free‐flooded cylindrical shroud81(1987); http://dx.doi.org/10.1121/1.2024097View Description Hide Description
The problem of determining virtual source distributions for an insonified free‐flooded pipe of finite length is posed as a pair of coupled integral equations. The geometry is assumed to have a rigid outer surface and a locally reacting inner surface. Predictions are made of directivity patterns of diffracted sound, with particular interest in the effect of lining characteristics on levels of on‐axis noise. At each chosen frequency the solution to the boundary‐value problem is checked by confirming the power balance between that injected at the incident source, and the sum of corresponding radiated and lining‐dissipated values. The effect of finite length on input energy is established upon comparison to that for the similarly lined infinite waveguide.
81(1987); http://dx.doi.org/10.1121/1.2024098View Description Hide Description
The radiation of noise from machinery located inside a submerged elastic structure involves both a direct structure‐borne and flanking airborne path. Calculations were performed to assess the importance of the latter for a cylindrical shell excited by an interior monopole source. An analytic solution for the interior pressure field is derived for a prescribed wall impedance using the Helmholtz integral equation approach, Farfield pressures are evaluated for a baffled section of the shell. Transfer functions of farfield to interior pressures are presented over a frequency range encompassing the ring frequency. The excess insertion loss arising from the interior application of various absorptive blankets is evaluated.
81(1987); http://dx.doi.org/10.1121/1.2024099View Description Hide Description
During Project KIWI ONE [D. G. Browning et al., Nature 282, 820–822 (1979)] anomalously high values of low‐frequency attenuation were observed in the central South Pacific Ocean. Similar results were obtained along a track from New Zealand into the Southern Ocean during Project TASMAN TWO [R. W. Bannister et al., J. Acoust. Soc. Am. 62, 847–859 (1977)]. A recent analysis of oceanographic data by Mellen shows that both regions are included in a relatively high pH contour at the sound channel axis. The corresponding predicted values of attenuation are in reasonable agreement with the measured data.
81(1987); http://dx.doi.org/10.1121/1.2024100View Description Hide Description
Laboratory measurements of the propagation velocity of ultrasonic compressional wave pulses were made at four orientations (vertical and at three horizontal directions) on five cores from first year sea ice. Accurately measured relative velocities (± 0.9%) showed that velocities were slightly anisotropic (generally < 5%). As in freshwater ice, it was found that the velocities parallel to the c axis (which was in the horizontal plane) are faster than horizontal velocities perpendicular to the c axis. However, the measured vertical velocity is faster than velocity parallel to the c axis in our sea ice cores. This is surprising since the velocity along the c axis in freshwater ice is faster than the perpendicular velocity.Velocities at 45° to the c axis were found to be either faster or slower than velocities parallel to the c axis, depending primarily on temperature. [This work was supported by the ONR.]
- Session E. Architectural Acoustics I: Research in Building and Room Acoustics
81(1987); http://dx.doi.org/10.1121/1.2024101View Description Hide Description
Newer auditorium acousticsmeasures were used to evaluate and compare a fan‐shaped and a rectangular auditorium. Measurements included reverberation times, early decay times, clarity C80, overall strength, and lateral fractions each measured in octave bands from 125 to 8000 Hz. Measurements were made at combinations of 3–5 source positions and 12 receiver positions. The results indicate that the shape of the hall influences the within‐hall variation of these acoustical quantities. The fan‐shaped hall is thought to have a less diffuse early sound field that causes larger seat‐to‐seat variations in acoustical quality and a systematic variation in acoustical characteristics from the front to the rear of the hall.
81(1987); http://dx.doi.org/10.1121/1.2024102View Description Hide Description
Much effort has been spent in investigating the relationship between the physical properties of reverberant sound fields and the parameters of sound qualities. A single form of parameter, which is a function of both space and time, that might help in establishing a relationship between its real‐time value and subjective criteria is sought. Based on a simplified model of spatially directional diffusibility, a general solution to the differential equation of transient diffusibility d(t) and the definiteness ratio D(t) is given. Computed and measured curves are consistent with each other; in particular, an explanation of the nonmonotonic V shape of d(t) curves by means of theoretical calculations and mathematical derivations is provided. It is suggested that d(t) and D(t) could serve as quantitative, predictive parameters in applications to architectural acoustics and the control of sound quality in acoustic engineering.
81(1987); http://dx.doi.org/10.1121/1.2024103View Description Hide Description
The proper design of an open plan office for speech privacy and for overall acoustical comfort considers many variables, all of which interact to produce the final acoustical environment. Using a loudspeaker to represent the talker, the sound spectrum that would be heard by a listener in an adjacent workstation is measured. From these measurements, the articulation index (AI) between adjacent workstations is calculated. Adjacent workstations are focused upon because it is generally more difficult to obtain good speech privacy for the occupants of these workstations than it is for workstations that are farther apart. Also, a PNC35 spectrum is used as the baseline masking spectrum. The PNC35 has an overall level of 43 dBA, and is representative of low‐to‐moderate levels of sound masking. A few trends emerge from this study. The sitting of workstations near hard room boundaries can be expected to have poorer speech privacy than corresponding workstations away from such room boundaries. Second, the acoustically hard screens whose NRC is 0.1 or less will have poorer speech privacy than will acoustically soft screens whose NRC is 0.5 or more.
81(1987); http://dx.doi.org/10.1121/1.2024104View Description Hide Description
The purpose of our experiment is to learn the influence of stiffening on the acoustic response of steel plates. The investigation is carried out on a rectangular plate that is clamped in an opening (1.20 × 2 m2) between a reverberant room and a test room. The sound transmission loss is determined by using two different methods: a method that is based on the sound pressuremeasurement and the two‐microphone intensity method. The mean radiation efficiency is obtained by measuring the panel vibration velocity with a regular mesh. The measurements are made over a frequency range from 250 to 10 000 Hz for a diffuse field excitation. The following series of samples is tested: a 1.5‐mm‐thick‐steel plate; a plate reinforced with uniformly spaced tubular stiffeners; a plate reinforced with uniformly spaced line stiffeners; and a corrugated panel (the space between the corrugations is the same as the space between the stiffeners). The acoustic effect of variation of bending stiffness is discussed, and all the experimental results provide comparison elements for computed results.
81(1987); http://dx.doi.org/10.1121/1.2024105View Description Hide Description
Several single number ratings, including STC, for rating the sound insulation properties of walls were correlated with the A‐weighted sound reductions of a group of gypsum board partitions for several types of sound source. None of the ratings fared well for the broad range of sound types. It is shown that a single rating method cannot deal with all sound sources. Two new ratings are proposed to supplement STC for coverage of indoor and outdoor noise reduction.
A source spectrum for simplified sound isolation tests specifically for results correlated to NIC ratings81(1987); http://dx.doi.org/10.1121/1.2024106View Description Hide Description
Using one‐third octave‐band sound transmission and reverberation time data of 54 party walls (ISO717, ISO140), the differences in dBA values between rooms were simulated as a function of wideband (90 Hz–3.5 kHz) source room spectra. An optimal source room spectrum was obtained by minimizing the standard deviation of (ΔdBA i − NIC i ) (i = 1.54) from the mean value. The rank correlation coefficient was seen to increase for decreasing standard deviation. The absorption data of the rooms were used to extend the results to the sound powerspectrum of the source. Results, examples, statistical data, and comparison with other workers will be given.
81(1987); http://dx.doi.org/10.1121/1.2024107View Description Hide Description
For reproducibility of simplified sound isolation tests [when a large approximately spherical, omnidirectional sound source is not used], it is appropriate to specify requirements for source placement. A scale model of a building with an exaggerated lopsided but reasonable quantity of absorption in one room (panel absorbers on wall surfaces around one corner) was used to investigate deviation of radiated power for the corner position from rigid wall conditions, the objective being to excite modes to an extent depending on their frequency and not mode number. The model and experiment will be discussed, and the results for this model as well as those for a full‐sized baffled loudspeaker will be given. The placement is of more importance in countries where the 100‐Hz, one‐third octave band is included. In light of annoyance from hi‐fi bass, the European range would seem more appropriate.