Volume 85, Issue S1, May 1989
Index of content:
- PROGRAM OF THE 117TH MEETING OF THE ACOUSTICAL SOCIETY OF AMERICA
- Tutorial on Acoustical Oceanography
85(1989); http://dx.doi.org/10.1121/1.2026856View Description Hide Description
Sound propagation is the preeminent technique for sensing, identifying, and communicating under the oceansurface. However, the extraordinary spatial and temporal variability of the ocean has frustrated underwater acousticians for decades. The solution, of course, was to learn more about the causes of this variability. The traditional oceanographic instruments do this rather crudely, and probably the best technique is to use the acoustical behavior to characterize the medium. Several years ago, ocean acoustics gave birth to acoustical oceanography. The acoustical oceanographer inverts the problem; he uses the seemingly capricious nature of sound propagation to learn about the ocean. The many successes of this young science range from the identification and counting of physical and biological inhomogeneities, such as microbubbles,zooplankton, and fish, to the remote sensing of distant rainfall and sea surface roughness, deep sea mountains, rocks and sediments, as well as the shape and strength of immense churning ocean eddies, hundreds of kilometers in extent.
- Session A. Noise I: Noise of Air‐Moving Devices
- Invited Papers
85(1989); http://dx.doi.org/10.1121/1.2026857View Description Hide Description
Fans used in the cooling of mechanical and electrical equipment are often the major sources of noise. This paper reviews the past and currentnoise control activities in the fan industry and related trade and professional organizations. Technology trends are discussed. Research needs and priorities are identified. The impact of fan design (blade, housing, and motor) and fan manufacturing (materials and processes) on fan noise and system acoustical quality are discussed. This paper illustrates with examples that fan noise control is more a challenge for design, application, and manufacturing than an investigation of fundamental aeroacoustic mechanisms. The approach is multi‐disciplinary and comprehensive in its treatment of each integration level—from bearings and windings to motor, housing, and end‐use equipment, thereby integrating noise control into fan design and manufacturing to achieve total quality for tomorrow's market.
85(1989); http://dx.doi.org/10.1121/1.2026900View Description Hide Description
The predominant acoustic noise sources of electronic systems are very often due to convection cooling devices. A wide range of capacities of air‐moving devices (AMDs) are used in computers and business equipment, typically varying in total flow rate from 0.01 to 1.0 m3/s, and in static pressure drop from 10 to 1000 Pa. This review paper will cover several aspects of air‐moving device noise: (1) basic mechanisms of aerodynamic noise generation, especially in the context of air‐moving device noise, (2) measurement methods of characterizing aeroacoustic noise sources on airfoils immersed in turbulence, (3) important parameters in assessing aeroacoustic performance, (4) standardized measurement methods for determining noise emission levels of AMDs; (5) use of sound power levels of AMDS for predicting overall system noise levels; and (6) nonstandardized measurement methods for quantifying air‐moving devices based on sound intensity techniques.
85(1989); http://dx.doi.org/10.1121/1.2026901View Description Hide Description
Measurement of fan noise requires extremely careful control of a multiple of variables—including fan operating point and test system configuration—to obtain useful results. Frequently, fan noise measurements are contaminated by unrecognized fan or test system changes and spurious noise sources. A new fan test system was designed to eliminate extraneous noise sources and to provide good control of the fan and the test system. This system uses round sheet metal ducts to connect the test fan to inlet and discharge anechoic terminations. Rotating microphone probes, one in the inlet and one in the discharge, are used to measure induct SPL while the fan's air performance is being determined. Two test systems, one with a mean diameters of 8 in. and the other of 16 in., have been built. Measurements have been made on both axial and centrifugal fans of sizes from 6 to 20 in. Representative results of these measurements and data from recent studies of microphone probe design and surface microphone systems use for in‐duct noise measurements will be discussed. Data will also be presented on the influence of tip clearance on axial fan noise.
85(1989); http://dx.doi.org/10.1121/1.2026902View Description Hide Description
Active noise control presents unique advantages for the reduction of noise from air‐moving devices. These include excellent low‐frequency performance, minimal flow restriction, and ease of installation. In this paper emphasis will be placed on the use of active noise control for silencing discharge noise from fans. A recently developed digital system using adaptive signal processing will be described. Results will be presented from a number of centrifugal and vaneaxial fans. Application guidelines have been developed to ensure that the system performance is maximized. These guidelines will be presented using results from actual field testing.
- Contributed Papers
85(1989); http://dx.doi.org/10.1121/1.2026951View Description Hide Description
Jet impingement on a large flat normal plate produces a noisy turbulent divergent wall jet. As in the more general case, the imageflow in the plate can be postulated and the “three sound‐pressures theorem” applied (neglecting dissipation effects): For long wavelengths (“compact” flow), the aerodynamic sound sources must be reducible to lateral quadrupoles alone, with maxima normal to and in that plane. Additional (usually weak ??) shear dipoles associated with the viscous traction forces lie in the plane of the plate. Specifically, the large scale axisymmetric unsteady flow associated with the flowresonances that sometimes occur resembles a periodic series of ring vortices that approach the plate and then spread out, symmetrically. Regardless of the details of this symmetrical unsteady flow, the associated source field must be periodic annular distributions of the lateral quadrupoles and traction dipoles; in the long‐wavelength (compact) low‐speed case these degenerate into symmetrical point quadrupoles, one lateral and one longitudinal, yielding “U 8 laws” for simple similarity. Dissipation results in “U 8‐type” monopoles due to heat addition fluctuations.
85(1989); http://dx.doi.org/10.1121/1.2026952View Description Hide Description
For high‐speed flows the source region is often not compact and the low‐speed arguments are invalid: Then the motion of the flow boundary may provide insight as to the nature of the sound sources. (i) When a “two‐dimensional” jet impinges on a large flat plate, the sinuous jet puffs from side to side at the plate, approximated by simple sources separated by the high impedance jet. (ii) The high‐speed edge tone is similar. (iii) For small plates sound will be generated by the axially symmetric modulation of the strength of the unstable stand‐off shock wave. Then, (a) changes in flow yield the total annular simple source strength or (b) the nearly conical boundary of the flow enveloping the plate appears to vibrate sympathetically providing an alternative source model. Some results of an alternate approach are given, in which the boundary is considered to be generating the evanescent waves, partially scattered at impingement or radiating at initiation.
85(1989); http://dx.doi.org/10.1121/1.2026953View Description Hide Description
Experimental aeroacoustic investigations of the reduction of intense noise radiated by improperly expanded turbulent jet flows are reported. The noise intensity varies with the strength of the repetitive shock structure inherently present in improperly expanded jet flows. To reduce shock‐associated noise levels, the needed weakening of the repetitive shock structure is achieved when improperly expanded jet flows issue (1) from a plug nozzle with either a contoured plug or a conical perforated plug operated at a range of above‐critical pressure ratios; (2) from coaxial nozzles of different configurations operated in the inverted pressure mode. Farfield acoustic results and the supporting optical records of these improperly expanded jet flows are presented. Striking changes in sound pressure level spectra, overall sound pressure level directivity, and substantial reductions in the intensity of shock‐associated noise are reported. Suppression of shock‐associated noise from improperly expanded turbulent jet flows issuing from a plug nozzle operated at off‐design pressure ratios is shown to be more effective than that for jet flows from an equivalent contoured convergent‐divergent nozzle operated in the over‐ and underexpanded modes.
85(1989); http://dx.doi.org/10.1121/1.2026954View Description Hide Description
Cylindrical shells, such as industrial piping system components, are efficiently excited by broadband internal noise at discrete frequencies below the ring frequency. Two types of excitation are recognized: finite length pipe resonances that are visible in short pipes, and coincident excitation that has been studied for long or anechoically terminated pipes. Both mechanisms occur to varying degrees in pipes of any length, and both require that the acoustic and structural wavenumbers be closely (or exactly) matched. Because ducts possess an infinite number of potential coincidence frequencies, coincidence transmission (i.e., precisely matched wavenumbers) is dominant in pipes. In the limit for long shells, both mechanisms are damping controlled and approach the same levels. Experimental data for short and intermediate shells show that coincidence transmission (“infinite shell”) can be the sole cause of vibration over wide frequency bands where the density of resonant modes is low. A simple theory is used to predict the frequency and amplitude of response peaks caused by both mechanisms. [Work supported by NSF.]
85(1989); http://dx.doi.org/10.1121/1.2026955View Description Hide Description
Based on Naval Underwater Systems Center [communications with E. Payne of NUSC] and Soviet [Sov. Phys. Acoust.30(5), 394–397 (1984)] water‐tunnel data, the Corcos wall‐pressure model was modified to include fluid‐injection effects. Boundary‐layer parameters were used to model the wall‐pressure frequency spectral density and the longitudinal cross‐spectral density. Using predictions from a Martin Marietta boundary‐layer code, favorable comparisons were made to the NUSC measured boundary‐layer profiles and wall‐pressure frequency spectra. Excellent agreement was obtained for both the effect of injection rate and the reduced effectiveness of fluid injection with distance from the injection source. Calculated wavenumber spectra showed that at low frequencies (less than about 200 Hz) spectral levels increased with injection, while at higher frequencies spectral levels decreased. The predicted convective peaks shifted to higher wavenumbers due to the decrease in convective velocity with increasing injection rate. The rms wall pressures are shown to increase for all fluid‐injection rates. The model predicts that the large flow‐noise reductions (over 20 dB) obtained in water‐tunnel experiments will not be achieved in full‐scale applications.
85(1989); http://dx.doi.org/10.1121/1.2026956View Description Hide Description
A methodology to predict the effects of pressure gradients on flow noise is presented. The method was developed from the pressure‐gradient data of Schloemer [J. Acoust. Soc. Am. 42, 93–113 (1966)] and of Burton [MIT Acoustics and Vibration Lab. Rep. 70208‐9 (1973)] and implemented with the Corcos wall‐pressure model. Predicted boundary‐layer parameters yield a wall‐pressure frequency spectrum for a particular flow condition, thus allowing one to predict conditions where data do not exist. For increasing pressure flows, rms wall pressures were predicted to be unaffected for Burton's test cases but increased for Schloemer's experiments. This apparent contradiction is in agreement with the experiments. Also, it is shown that erroneous conclusions result from a zero‐pressure‐gradient wall‐pressure model. Computations of wavenumber spectra in adverse gradients gave higher spectral levels except in the vicinity of convective peaks where levels are noticeably lower, especially at the high frequencies. For Burton's test cases, frequency‐spectral‐density calculations showed increases at low frequencies and reductions at high frequencies, although not as dramatic as Burton's data indicated for the most severe adverse‐pressure gradients. Wall‐pressure levels were predicted to increase at all frequencies for Schloemer's data. The model clearly shows that these differences are due to the way that the adverse‐pressure gradients are formed.
85(1989); http://dx.doi.org/10.1121/1.2026999View Description Hide Description
Computer cabinet designers choose fans based on their air flow characteristics but noise concerns often have a significant impact on the final cabinet design. Many physical and operating parameters can impact the fan noise. The noise emissions from a large number of fans have been measured using ANSI S12.11‐1987. Regression analysis is used to evaluate the correlation of the various parameters to the fan noise level in order to understand trade‐offs at the initial design stage.
- Session B. Physical Acoustics I: Nonlinear Acoustics
- Invited Paper
85(1989); http://dx.doi.org/10.1121/1.2027000View Description Hide Description
Early calculations by Bethe, Zel'dovich, and others have suggested the possibility of rarefaction shocks in fluids and solids. Recent experiments have demonstrated the existence of rarefaction shocks in rubber (Kolsky and Rader), quartz (Barker and Hollenbach), iron (Erkman, Ivanov, and Novikov), fluids at near‐critical states (Borisov et al.), and in evaporating mixtures (Chaves, Thompson et al.). Analogous discontinuities are found in second sound in superfluidhelium. The assertion that all rarefaction shocks violate the second law is untrue, both in theory and practice! Physical experiments in fluids leading to a single‐phase rarefaction shock, or to a complete rarefaction‐evaporation shock have, until now not been realized. Key ingredients for real rarefaction shocks include a negative nonlinearity parameter Γ, usually some degree of metastability, Chapman‐Jouguet conditions, and in the case of fluids, a large molar heat capacity (many molecular degrees of freedom). Examples and related phenomena are discussed.
- Contributed Papers
85(1989); http://dx.doi.org/10.1121/1.2027001View Description Hide Description
Periodic wavetrains propagating in fluids whose specific heats are large compared to the molal gas constant are examined. In the dense gas regime, the fundamental nonlinearity parameter 1 + B/2A of these fluids may become negative. The present study examines the nonclassical properties of the evolution of a sinusoidal wavetrain including the formation and propagation of expansion shocks and sonic shocks. Analytical solutions are presented for the inviscid problem and are compared to numerical solutions describing the dissipative evolution.
85(1989); http://dx.doi.org/10.1121/1.2027002View Description Hide Description
Approximate nonlinear equations describing the propagation of finite amplitude sound beams in liquids, isotropic solids, and crystals are suggested. The equations have been derived under the assumptions of small perturbation and slow variation of the wave shape, which are due to diffraction and nonlinearity. Diffraction is taken into account within the quasioptical approximation. Nonlinearity is considered up to quadratic terms for the cases of wave propagation in liquids and crystals, as well as for longitudinal wave propagation in solids. As for nonlinear propagation of transverse waves in an isotropic solid, cubic terms in the magnitude of the perturbation are retained.
85(1989); http://dx.doi.org/10.1121/1.2027058View Description Hide Description
The linear and nonlinear propagation of a pulsed sound beam generated by a real source in a fluid is considered. The source can be plane or weakly focusing. The investigation is based on a linear and quasilinear solution of the Khokhlov‐Zabolotskaya‐Kuznetsov nonlinear parabolic equation. Analytical and numerical results are presented. The evolution of the pulse as it propagates from the source into the farfield region is investigated for various pulse forms. The special case of a source with distribution exp(−x 2/a 2)r(t) (x radial distance, t time) is investigated in detail, with emphasis on the role of diffraction and absorption on the self‐demodulation of the pulse. The results are related to the problem of scattering of sound by sound. [Work supported by the IR&D program of ARL:UT and VISTA/STATOIL, Norway.]
85(1989); http://dx.doi.org/10.1121/1.2027059View Description Hide Description
The scattering of sound by sound in a lossless fluid was discussed at an earlier meeting [Berntsen et al., J. Acoust. Soc. Am. Suppl. 1 83, S4 (1988), and Darvennes and Hamilton, J. Acoust. Soc. Am. Suppl. 1 83, S4 (1988)]. Here, the effects of absorption are included. The Khokhlov‐Zabolotskaya‐Kuznetsov equation is used to derive farfield asymptotic results for the sum and difference frequency sound due to the noncollinear interaction of real sound beams radiated from displaced sources. There are two main contributions to the nonlinearly generatedsound in the farfield: the continuously pumped sound and the scatteredsound. Weak absorption affects neither the locations nor the relative amplitudes of the pumped and scattered difference frequency sound. Strong absorption attenuates the pumped difference frequency sound faster than the scattered difference frequency sound. The scattered sum frequency sound is always attenuated faster than the pumped sum frequency sound, and there may be shifts in the locations of the maxima. Numerical results are presented for the case of Gaussian primary beams. [Work supported by ONR (CMD and MFH), IR&D Program of ARL:UT, and VISTA/STATOIL (JNT and ST).]
85(1989); http://dx.doi.org/10.1121/1.2027060View Description Hide Description
A problem of continuous interest in underwater sound propagation is the prediction of the scatteringproperties of bubbles. Two important parameters in this problem are the bubble's radius and damping coefficient. A method of measuring these parameters, which is a modification of a bubble‐sizing technique developed by V. L. Newhouse and P. M. Shankar [J. Acoust. Soc. Am. 75, 1473–1477 (1984)] is described. The method exploits the nonlinear, mixing property of resonant bubbles simultaneously exposed to sound fields of different frequencies. The results of measurements of bubble radius and resonance damping coefficient are presented for bubbles having resonance frequencies less than approximately 100 kHz. Extension of the technique to smaller bubbles is discussed. [Work conducted for the Naval Coastal Systems Center and funded by the Naval Postgraduate School.]
85(1989); http://dx.doi.org/10.1121/1.2027061View Description Hide Description
Previous attempts to find linearized solutions of the equations for vapor bubble pulsations [e.g., see Finch and Neppiras, J. Acoust. Soc. Am. 53, 1402–1410 (1973)] assumed that oscillations would be the result of imposed sonic excitation. It was shown by Nicholas and Finch [12th Int. Cong. Acoust., Toronto, paper 14‐2 (1986)] that nonlinear solutions of the equations in fact often showed exponential expansions or collapses. In this paper it is shown that if the governing equations are linearized using a Taylor series expansion, without assuming either oscillatory or exponential behavior, then there are regimes of bubble radius in which one or the other of these two is a necessary result. The simple case of a bubble with an insoluble gas content in a nonconducting liquid is considered. There can then be shown to exist a critical radius, with a value given by 2σ/3γP, where σ is the surface tension, γ is the ratio of specific heats, and P is the gas pressure. Below this size the behavior is exponential, above it, oscillatory. The results can be interpreted in terms of the stability of an open‐loop control system.
Effects of nuclei and host fluid parameters on the threshold for cavitation produced by pulsed ultrasound85(1989); http://dx.doi.org/10.1121/1.2027062View Description Hide Description
An experimental apparatus has been developed to determine thresholds for cavitation produced in a fluid by short tone bursts of ultrasound at 0.76, 0.99, and 2.30 MHz [A. Atchley et al., Ultrasonics26, 280–285 (1988)]. A fluid jet was used to convect potential cavitation nuclei, such as 1‐μm polystyrene spheres, echo contrast spheres, and whole blood constituents, through the focal region of the insonifying transducer.Cavitation thresholds measured with this system in water and in a fluid with ten times the viscosity of water will be presented. Cavitation was detected by a passive acoustical technique that is sensitive to sound scattered from cavitation bubbles. Results from these experiments that permit the control of nuclei and host fluid properties will be compared to an approximate theory that predicts the onset of cavitation [C. K. Holland and R. E. Apfel, Trans. IEEE UFFC‐36(2) (1989)]. [Work supported by NIH, grant number 1RO1‐CA‐39374.]