Siyang Zhong

Compressive sensing beamforming based on covariance for acoustic imaging with noisy measurements

Compressive sensing, a newly emerging method from information technology, is applied to array beamforming and associated acoustic applications. A compressive sensing beamforming method (CSB-II) is developed based on sampling covariance matrix, assuming spatially sparse and incoherent signals, and then examined using both simulations and aeroacoustic measurements. The simulation results clearly show that the proposed CSB-II method is robust to sensing noise. In addition, aeroacoustic tests of a landing gear model demonstrate the good performance in terms of resolution and sidelobe rejection.

A controllable canonical form implementation of time domain impedance boundary conditions for broadband aeroacoustic computation

A new method, which can be effectively and efficiently applied in the simulations of broadband noise problems, is proposed for time domain impedance boundary condition implementations by using the so-called controllable canonical form that is well known in linear system. Usually, the impedance boundary condition can be defined in frequency domain as a rational polynomial function with poles in the negative half of the complex plane to guarantee stability; otherwise, causality might be violated in the corresponding time domain implementation. To address this issue, various methodologies have been proposed previously that usually lead to complicated polynomials, whose numerical implementations are often indirect and intricate. The proposed method with a controllable canonical form, on the other hand, directly transforms the frequency domain transfer function (a quotient of rational polynomials) to an equivalent state space model, which consists of a series of first-order ordinary differential equations that can be numerically implemented in a straightforward way. The proposed method is demonstrated by using two benchmark problems: a two-dimensional Gaussian pulse propagating in a uniform flow with a lined wall and the test cases from the NASA Langley grazing incidence tube experiments. Good agreements demonstrate the potential of the proposed computational method.

Analysis of scattering from an acoustic cloak in a moving fluid

This work develops a theoretical framework for acoustic cloak scattering analysis in a low speed non-stationary fluid that is simply described as a potential flow. The equivalent sound source induced by the moving fluid local to the cloak is analytically constructed and is then estimated using Born approximation. The far-field scattering can thereafter be obtained using the associated Greens function of the convected wave equation. The results demonstrate that the proposed analytical approach, which might be helpful in the design and evaluation of cloaking systems, effectively elucidates key characteristics of the relevant physics. In addition, it can be seen that, in a moving fluid, the so-called convected cloaking design achieves better cloaking performance than the classical cloaking design.

Experimental evaluation of flow-induced noise in level flight of the pigeon (Columba livia)

The experimental method employed in an anechoic wind tunnel to characterize flow-induced noise of the pigeon (Columba livia) during level flight is described in this letter. A live pigeon was managed to maintain a steady level flight at the wind tunnel test flow of 15 m/s. A microphone array was fabricated, and the conventional beamforming method was then adopted to yield the corresponding narrowband acoustic images and broadband sound pressure spectral results. The results justified the experimental method developed in this work. It can be seen that the flight noise of the pigeon is mostly from the wing tips. In addition, the spectral waveform of the pigeon flight suggests a slope of -20 dB/dec between 500 Hz and 5 kHz.

A generalized sound extrapolation method for turbulent flows

Sound extrapolation methods are often used to compute acoustic far-field directivities using near-field flow data in aeroacoustics applications. The results may be erroneous if the volume integrals are neglected (to save computational cost), while non-acoustic fluctuations are collected on the integration surfaces. In this work, we develop a new sound extrapolation method based on an acoustic analogy using Taylor’s hypothesis (Taylor 1938 Proc. R. Soc. Lon. A 164, 476–490. (doi:10.1098/rspa.1938.0032)). Typically, a convection operator is used to filter out the acoustically inefficient components in the turbulent flows, and an acoustics dominant indirect variable Dcp′ is solved. The sound pressure p′ at the far field is computed from Dcp′ based on the asymptotic properties of the Green’s function. Validations results for benchmark problems with well-defined sources match well with the exact solutions. For aeroacoustics applications: the sound predictions by the aerofoil–gust interaction are close to those by an earlier method specially developed to remove the effect of vortical fluctuations (Zhong & Zhang 2017 J. Fluid Mech. 820, 424–450. (doi:10.1017/jfm.2017.219)); for the case of vortex shedding noise from a cylinder, the off-body predictions by the proposed method match well with the on-body Ffowcs-Williams and Hawkings result; different integration surfaces yield close predictions (of both spectra and far-field directivities) for a co-flowing jet case using an established direct numerical simulation database. The results suggest that the method may be a potential candidate for sound projection in aeroacoustics applications.

Airfoil-Gust Interactions in Transonic Flow

Leading edge noise is a significant broadband noise source in aircraft engines, and is the primary broadband noise mechanism in outlet guide vane noise in turbofans, and broadband rotor wake interaction noise in contra-rotating open rotor engines. Previous authors have studied the effects of various aspects relating to this noise source, including airfoil geometry effects, cascade effects, and Mach number effects. However, previous literature has not addressed the effects on the noise due to locally supersonic regions that might be present in the mean flow around the rotor blades. The current work uses computational aeroacoustic methods to investigate the effects of locally supersonic regions on the noise due to airfoil- gust interactions. An established computational aeroacoustics code has been extended to give stable predictions in supersonic regions with a localized artificial diffusivity method. Initial results of a NACA 0012 airfoil in M = 0.8 flow interacting with oncoming vortical waves are shown, alongside results for a NACA 0006 airfoil in M = 0.5 flow at a 6 ? angle of attack. The changes to the noise and the underlying mechanisms are discussed for both cases, including additional noise sources caused by the supersonic region.

Artificial damping methods for stable computations with linearized Euler equations

In this work, new methods are developed to facilitate stable and accurate numerical solutions of linearized Euler equations, which are often used in solving problems in computational aeroacoustics. Solutions of LEE can suffer from numerical Kelvin-Helmholtz instabilities in the presence of a sheared mean flow. Various methods have been exploited to address this problem; each has its advantages and disadvantages. In this work, two new methods that use artificial damping terms (ADT) are introduced. The first method is constructed to damp the vortical components generated during the computation while the second one is proposed by revisiting the effect of viscosity in the Navier-Stokes equations. An adaptive method is also used to improve the proposed new methods. These methods are tested on two benchmark cases: a) acoustic wave refraction through a strongly sheared jet, and b) mode radiation from a semi-infinite duct with jet. It is found that numerical instabilities can be successfully suppressed with little side effect on the acoustic wave computations.

On the frequency domain formulation of the generalized sound extrapolation method

A frequency domain formulation of the generalized sound extrapolation method is proposed. The Fourier transform is performed to the flow variables to prepare the source terms as the input of the sound extrapolation solver, and the inhomogeneous convected Helmholtz equation is solved using the Greens functions in the frequency domain. The proposed formulations are applied to typical aeroacoustics problems including two dimensional (2D) and three dimensional monopoles in uniform mean flows, flat plate-gust interaction noise and the sound produced by a co-flowing jet. The formulations can yield close agreements with the predictions by the time domain solution. Substantial time saving can be achieved compared with a time domain solver, as the estimation of emission/retarded time, which needs to conduct interpolation of the time sequence of the flow variables, is not required. Also, the computation in the 2D space is much easier since the Greens function is used directly. The results suggest that the frequency formulation can potentially be used for sound projection computation in aeroacoustics.