Effect of flow on quasi-one-dimensional acoustic wave propagation in a variable area duct of finite length

The general equation for the velocity potential of quasi-one-dimensional acoustic wave motion in a variable area, finite duct with one-dimensional flow is derived by using a perturbation technique. The non-linear second-order partial differential equation is linearized and then solved, by either a power series expansion method or the Runge-Kutta fourth-order method, for harmonic time dependence. The boundary condition taken at the duct mouth is that of matching the impedance of the duct sound field to that of the radiation field at the duct opening. Three axial Mach number variations along the duct axis are considered and the results obtained are compared with those for the case of constant Mach number, to determine the influence of the axial velocity gradient on sound propagation. The effect of flow on the radiation impedance is also considered.     

Sound attenuation in acoustically lined curved ducts in the absence of fluid flow

A study has been made of the sound attenuation in a lined curved duct with rectangular cross-section. In this study, the derivation of the eigenvalue equation was based on the continuity of the normal component of the particle displacement and the matching of the acoustic pressure on the acoustic lining surface. The sound attenuation was calculated by using the acoustic energy expression for the waves propagating in a curved duct. For a given duct geometry and known acoustic lining impedances, a computer program was developed to solve for the eigenvalues and to obtain the sound attenuation of the propagating waves in the lined curved duct. It was found that in the case studied here the fundamental mode was least attenuated. The total sound attenuation was calculated on the assumption that the amplitudes for all propagating waves were equal at a given frequency. Effects of aspect ratio, bend angle and the acoustic impedance on the sound attenuation were investigated in the present work.     

The application of a semi-actuator disk model to sound transmission calculations in turbomachinery,part I: The single blade row

The work described in this paper generalizes the semi-actuator disk model of a blade row of Kaji and Okazaki [1, 2] to include a three-dimensional incident sound field. This field is equivalent, for example, to a radially fluctuating spinning mode in an annular compressor duct. The dependence of the earlier model upon the wavelength of the incident sound field is removed by adopting an average frequency approach. In this model, the acoustic energy entering the cascade of blades across one interface and leaving it across the other interface is summed over an infinite series of reflections within the cascade. This is equivalent, in effect, to the assumption of an incident delta function of pressure and thus gives an average frequency result. The validity of the model is demonstrated in a series of comparisons with data obtained from earlier models. The model is subsequently extended to include the effects of introducing cambered blades. In a companion paper (Part II), the model is applied to multiple blade rows.     

流管实验装置中声传播计算的模态方法

流管实验装置是测量有流动情况下航空发动机消声短舱内声衬声阻抗的主要装置。本文发展了一种解析的模态匹配方法进行在平均流有声衬条件下矩形流管中声传播的计算。用同伦方法求解特征值问题,并与用环绕积分求解的结果进行比较。声场通过轴向阻抗间断面的声压和声质点速度积分相等计算。第一个算例是无流动、硬壁、有限长、考虑端口反射的情况,并与北航流管实验台测量数据进行了对比;第二个算例为有流动情况下有限长声衬管道不考虑端口反射的声场计算,它与文献中NASA流管实验结果和CAA计算结果符合得很好。     

Effects of wall admittance changes on duct transmission and radiation of sound

This paper is concerned with the effect of changes in duct wall acoustic properties on the transmission of sound through ducts. Two special problems are considered. The first problem is that of a rectangular infinite-length duct with airflow and a single change in duct wall acoustic admittance. The second problem is that of an axisymmetric field in a finite circular duct without airflow and with an arbitrary number of duct wall acoustic admittance changes. Results for the first problem show the effect of wall admittance change and flow on the acoustic power transmission within the duct. Results for the second problem show the interactive effects of multiple duct liner sections on power radiated from a finite duct.     

Sound generation and transmission in flow ducts with axial temperature gradients

This article describes a one-dimensional, linearized, analysis of fundamental mode sound generation and propagation in rigid-walled flow ducts with axial temperature variation. An acoustic wave equation, including damping effects and volume sources, is derived and its solution (in the absence of sources) by a numerical technique and an approximate analytical method is discussed. The “forced” wave equation is then solved (the existence of an oscillating solution to the “unforced” equation being assumed) for sound generation by a side-branch volume source in an infinite duct, and the results are applied to a duct of finite length. Reasonably good agreement is obtained between measurements and predictions of the sound pressure field in a flow duct, away from the source region.     

An improved multimodal method for sound propagation in nonuniform lined ducts

An efficient method is proposed for modeling time harmonic acoustic propagation in a nonuniform lined duct without flow. The lining impedance is axially segmented uniform, but varies circumferentially. The sound pressure is expanded in term of rigid duct modes and an additional function that carries the information about the impedance boundary. The rigid duct modes and the additional function are known a priori so that calculations of the true liner modes, which are difficult, are avoided. By matching the pressure and axial velocity at the interface between different uniform segments, scattering matrices are obtained for each individual segment; these are then combined to construct a global scattering matrix for multiple segments. The present method is an improvement of the multimodal propagation method, developed in a previous paper [Bi et al., J. Sound Vib. 289, 1091-1111 (2006)]. The radial rate of convergence is improved from O(n(-2)), where n is the radial mode indices, to O(n(-4)). It is numerically shown that using the present method, acoustic propagation in the nonuniform lined intake of an aeroengine can be calculated by a personal computer for dimensionless frequency K up to 80, approaching the third blade passing frequency of turbofan noise.     

穿纤维微穿孔板吸声系数的有限元仿真

为拓宽微穿孔板的吸声频带,该文用有限元算法建立了典型微穿孔板和穿入不同数量金属纤维的微穿孔板模型,研究了两种微穿孔板的吸声系数、声阻抗和微孔内法向质点速度的空间分布,并进行了试验验证。有限元仿真和试验数据表明:穿入金属纤维可以拓宽微穿孔板的吸声频带,吸声系数也随纤维根数的增加而下降;吸声系数仿真结果与试验结果趋势一致,仿真模型可以有效模拟穿入纤维前后微穿孔板的吸声特性;穿入金属纤维导致黏滞效应引起的低质点速度区域增大,声阻增加,引起吸声系数的降低,而声抗变化不大。研究发现,有限元仿真方法适用于结构相对复杂的微穿孔结构的声学建模,能直观地体现微孔复杂结构的影响,值得继续深入研究和工程应用。     

Propagation and radiation of sound in a finite length duct

An analysis has been carried out to investigate sound transmission in an unflanged circular duct of finite length. A new formulation for calculating the generalized radiation impedance of the opening of a finite length duct with a spinning source inside is presented. The effect of interference between the duct apertures at its two ends is identified in the calculations of radiation impedance, reflection coefficient, and the far field radiation pattern. The results show that the degree of interference between the two ends reduces for high frequency waves and/or for long ducts. Further, the interference effects on the far field for spinning sound are more serious than those for non-spinning sound.     

On the acoustic modes in a cylindrical duct with an arbitrary wall impedance distribution

The present paper considers the propagation of sound in a cylindrical duct, with a wall section of finite length covered by an acoustic liner whose impedance is an arbitrary function of position. The cases of (i) uniform wall impedance, and wall impedance varying along the (ii) circumference or (iii) axis of the duct, or (iv) both simultaneously, are explicitly considered. It is shown that a nonuniform wall impedance couples modes with distinct azimuthal l or axial m wave numbers, so that their radial wave numbers k can no longer be calculated separately for each pair (m,l). The radial wave numbers are the roots of an infinite determinant, in the case when the wall impedance varies either (i) circumferentially or (ii) radially. If the wall impedance varies (iv) both radially and circumferentially, then the radial wave numbers are the roots of a doubly infinite determinant, i.e., an infinite determinant in which each term is an infinite determinant. The infinite determinants specifying the radial wave numbers are written explicitly for sound in a cylindrical nozzle with a uniform axial flow, in which case the radial eigenfunctions are Bessel functions; the method of calculation of the radial wave numbers applies equally well to a cylindrical nozzle with shear flow and/or swirling flows, with the Bessel functions replaced by other eigenfunctions. The radial wave numbers are calculated by truncation of the infinite determinants, for several values of the aspect ratio, defined as the ratio of length to diameter. It is shown that a nonuniform wall impedance will give rise to additional modes compared with a uniform wall impedance. The radial wave numbers specify the eigenfrequencies for the acoustic modes in the duct; the imaginary parts of the eigenfrequencies specify the decay of the sound field with time, and thus the effectiveness of the acoustic liner.     

On the acoustic modes in a duct containing a parabolic shear flow

The exact solution of the acoustic wave equation in an unidirectional shear flow with a parabolic velocity profile is obtained, representing sound propagation in a plane, parallel walled duct, with two boundary layers over rigid or impedance walls. It is shown that there are four cases, depending on the critical level(s) where the Doppler shifted frequency vanishes: (i) for propagation upstream the critical levels are outside the duct (case II); (ii) for propagation downstream there may be two (case IV), one (case I) or no (case III) critical level inside the duct. The acoustic wave equation is transformed in each of the four cases to particular forms of the extended hypergeometric equation, which has power series solutions, some involving logarithmic singularities. In the cases where critical levels occur, at real or ‘imaginary’ distance, matching of two or three pairs of solutions, valid over regions each overlapping the next, is needed. The particular case of the parabolic velocity profile is used to address general properties of sound in unidirectional shear flows. For example, it is shown that for ducted shear flows, there exist a pair of even and odd eigenfunctions, in the absence of critical levels. It is also proved, in more than one instance, that there is no single set of eigenvalues and eigenfunctions valid across one or two shear layers. This leads to the general conjecture, considering the acoustics of shear flows in ducts, that critical levels separate regions with distinct sets of eigenvalues and eigenfunctions.     

Theory of the Greenspan viscometer

We present a detailed acoustic model of the Greenspan acoustic viscometer, a practical instrument for accurately measuring the viscosity eta of gases. As conceived by Greenspan, the viscometer is a Helmholtz resonator composed of two chambers coupled by a duct of radius rd. In the lowest order, eta=pi f rho(rd/Q)2, where f and Q are the frequency and quality factor of the isolated Greenspan mode, and rho is the gas density. In this level of approximation, the viscosity can be determined by measuring the duct radius and frequency response of the resonator. In the full acoustic model of the resonator, the duct is represented by a T-equivalent circuit, the chambers as lumped impedances, and the effects of the diverging fields at the duct ends by lumped end impedances with inertial and resistive components. The model accounts for contributions to 1/Q from thermal dissipation (primarily localized in the chambers) and from a capillary used for filling and evacuating the resonator. A robust, prototype instrument is being used for measuring the viscosity of reactive gases used in semiconductor processing. For well-characterized surrogate gases, the prototype viscometer generated values of eta that were within +/-0.8% of published reference values throughout the pressure range 0.2-3.2 MPa. Remarkably, we achieved this level of agreement by only slight adjustment of the numerically calculated inertial and resistive end effect parameters to improve the agreement with helium reference values. No other parameters were adjusted.     

Sound transmission characteristics of Tee-junctions and the associated length corrections

The sound transmission characteristics of a Tee-junction formed by a sidebranch and an infinitely long duct are investigated numerically using the finite element method. The associated corrections of the branch length and the upstream duct length are also discussed in detail. The types of branch resonance that result in strong or weak sound transmission across the junction are determined and their effects on the length corrections examined. Results suggest that the type of sidebranch, the branch width, the branch length, and the order and the form of the resonance affect more significantly the length corrections of the duct section. The excitation of nonplanar higher branch modes gives rise to rapid increase in the duct length corrections and also results in lower sound transmission.     

Acoustic characterization and prediction for fan-duct-plenum-room integrations

This paper investigates mutual influence of duct and room acoustics in the whole fan-duct-plenum-room integrations. Applying the parametric design language of finite element software ANSYS (APDL), dimensional and positional influence on system acoustics has been studied. Models with different room dimensions, duct lengths, duct cross-sections, duct locations, duct discharges and duct elbow were constructed, and their characteristics were compared qualitatively. Results show that small rooms, short ducts, large duct cross-sections and bell mouth duct discharges help to increase room sound pressure levels (SPLs); SPLs in ducts and plenums are sensitive to duct dimensions and duct discharge types but insensitive to duct locations and room dimensions; duct elbows have relatively indistinct acoustic influence in each component. Based on the calculation results, a semi-experimental method was proposed for simply and approximately evaluating indoor acoustic spectra of fan-duct-plenum-room integrations, then an example was used to demonstrate the prediction process. Finally, by adopting several ideal models, sound field constitutions, duct and room wall admittances and duct end reflection were explored quantitatively. This study may give a detailed understanding of fan-duct-plenum-room acoustics for researchers, also it might provide a new, simple and approximate prediction method for professionals to evaluate and improve fan-ducted acoustics.     

Vibration amplitude and induced temperature limitation of high power air-borne ultrasonic transducers

The acoustic impedances of matching layers, their internal loss and vibration amplitude are the most important and influential parameters in the performance of high power airborne ultrasonic transducers. In this paper, the optimum acoustic impedances of the transducer matching layers were determined by using a genetic algorithm, the powerful tool for optimizating domain. The analytical results showed that the vibration amplitude increases significantly for low acoustic impedance matching layers. This enhancement is maximum and approximately 200 times higher for the last matching layer where it has the same interface with the air than the vibration amplitude of the source, lead zirconate titanate-pizo electric while transferring the 1 kW is desirable. This large amplitude increases both mechanical failure and temperature of the matching layers due to the internal loss of the matching layers. It has analytically shown that the temperature in last matching layer with having the maximum vibration amplitude is high enough to melt or burn the matching layers. To verify suggested approach, the effect of the amplitude of vibration on the induced temperature has been investigated experimentally. The experimental results displayed good agreement with the theoretical predictions.     

Influence of acoustic leakage at fluid–solid interface on interferometric detection of surface acoustic wave

When using laser interferometer to detect surface acoustic wave at fluid–solid interface, there are two factors which will cause the optical path length variation of the probe laser beam: interface deformation, and refractive index changes in fluid induced by acoustic leakage. Influence of acoustic leakage on laser interferometric detection for surface acoustic wave is researched here. A metal plate immersed in an infinite fluid is used as a physical model. Interface deformation due to laser-induced acoustic wave and pressure in fluid due to acoustic leakage are computed for select cases by finite element method. The optical path length variation caused by the two factors are calculated respectively and compared. The results show that the influence of acoustic leakage increases with the increasing acoustic impedance matching of fluid and solid, the peak-to-peak of influence degree increases linearly with the increasing acoustic impedance of fluid, and that decreasing the distance between the interferometer and interface can effectively reduce the influence of acoustic leakage.     

On whistling of pipes with a corrugated segment: Experiment and theory

Corrugated pipes are commonly used because of their local rigidity combined with global flexibility. The flow through such a pipe can induce strong whistling tones, which is an environmental nuisance and can be a threat to the mechanical integrity of the system. This paper considers the use of a composite pipe: a shorter corrugated pipe segment embedded between smooth pipe segments. Such a pipe retains some flexibility, while the acoustical damping in the smooth pipe reduces whistling tones. Whistling is the result of coherent vortex shedding at the cavities in the wall. This vortex shedding is synchronized by longitudinal acoustic waves traveling along the pipe. The acoustic waves trigger the vortex shedding, which reinforces the acoustic field for a critical range of the Strouhal number values. A linear theory for plane wave propagation and the sound production is proposed, which allows a prediction of the Mach number at the threshold of whistling in such pipes. A semi-empirical approach is chosen to determine the sound source in this model. This source corresponds to a fluctuating force acting on the fluid as a consequence of the vortex shedding. The functional form of the Strouhal number dependency of the dimensionless sound source amplitude is based on numerical simulations. The magnitude of the source and the Strouhal number range in which it can drive whistling are determined by matching the model to results for a specific corrugated pipe segment length. This semi-empirical source model is then applied to composite pipes with different corrugated segment lengths. In addition, the effect of inlet acoustical convective losses due to flow separation is considered. The Mach number at the threshold of whistling is predicted within a factor 2.     

The attenuation of the higher-order cross-section modes in a duct with a thin porous layer

A numerical method for sound propagation of higher-order cross-sectional modes in a duct of arbitrary cross-section and boundary conditions with nonzero, complex acoustic admittance has been considered. This method assumes that the cross-section of the duct is uniform and that the duct is of a considerable length so that the longitudinal modes can be neglected. The problem is reduced to a two-dimensional (2D) finite element (FE) solution, from which a set of cross-sectional eigen-values and eigen-functions are determined. This result is used to obtain the modal frequencies, velocities and the attenuation coefficients. The 2D FE solution is then extended to three-dimensional via the normal mode decomposition technique. The numerical solution is validated against experimental data for sound propagation in a pipe with inner walls partially covered by coarse sand or granulated rubber. The values of the eigen-frequencies calculated from the proposed numerical model are validated against those predicted by the standard analytical solution for both a circular and rectangular pipe with rigid walls. It is shown that the considered numerical method is useful for predicting the sound pressure distribution, attenuation, and eigen-frequencies in a duct with acoustically nonrigid boundary conditions. The purpose of this work is to pave the way for the development of an efficient inverse problem solution for the remote characterization of the acoustic boundary conditions in natural and artificial waveguides.     

Sound transmission across duct constrictions with and without tapered sections

The sound power transmission loss across duct constrictions with linearly tapered sections is studied with the finite element method. Results show that the acoustic energy distributions of transmitted waves at high frequency depend critically on the exit configuration of the constriction. The corresponding strengths of these waves are very much affected by the entrance setup of the constriction. The difference between inlet and outlet impedance of a constriction leads to weaker resonant sound transmission.     

Determination of acoustic impedances of multi matching layers for narrowband ultrasonic airborne transducers at frequencies <2.5 MHz - Application of a genetic algorithm

The effective ultrasonic energy radiation into the air of piezoelectric transducers requires using multilayer matching systems with accurately selected acoustic impedances and the thickness of particular layers. One major problem of ultrasonic transducers, radiating acoustic energy into air, is to find the proper acoustic impedances of one or more matching layers. This work aims at developing an original solution to the acoustic impedance mismatch between transducer and air. If the acoustic impedance defences between transducer and air be more, then finding best matching layer(s) is harder. Therefore we consider PZT (lead zirconate titanate piezo electric) transducer and air that has huge acoustic impedance deference. The vibration source energy (PZT), which is used to generate the incident wave, consumes a part of the mechanical energy and converts it to an electrical one in theoretical calculation. After calculating matching layers, we consider the energy source as layer to design a transducer. However, this part of the mechanical energy will be neglected during the mathematical work. This approximation is correct only if the transducer is open-circuit. Since the possibilities of choosing material with required acoustic impedance are limited (the counted values cannot always be realized and applied in practice) it is necessary to correct the differences between theoretical values and the possibilities of practical application of given acoustic impedances. Such a correction can be done by manipulating other parameters of matching layers (e.g. by changing their thickness). The efficiency of the energy transmission from the piezoceramic transducer through different layers with different thickness and different attenuation enabling a compensation of non-ideal real values by changing their thickness was computer analyzed (base on genetic algorithm). Firstly, three theoretical solutions were investigated. Namely, Chebyshev, Desilets and Souquet theories. However, the obtained acoustic impedances do not necessarily correspond to a nowadays available material. Consequently, the values of the acoustic impedances are switched to the nearest values in a large material database. The switched values of the acoustic impedances do not generally give efficient transmission coefficients. Therefore, we proposed, in a second step, the use of a genetic algorithm (GA) to select the best acoustic impedances for matching layers from the material database for a narrow band ultrasonic transducer that work at frequency below the 2.5 MHz by considering attenuation. However this bank is rich, the results get better. So the accuracy of the propose method increase by using a lot of materials with exact data for acoustic impedance and their attenuation, especially in high frequency. This yields highly more efficient transmission coefficient. In fact by using increasing number of layer we can increase our chance to find the best sets of materials with valuable both in acoustic impedance and low attenuation. Precisely, the transmission coefficient is almost equal to unity for the all studied cases. Finally the effect of thickness on transmission coefficient is investigated for different layers. The results showed that the transmission coefficient for air media is a function of thickness and sensitive to it even for small variation in thickness. In fact, the sensitivity increases when the differences of acoustic impedances to be high (difference between PZT and air).