Method and system for reconstructing rotating sound sources in non-uniform flow fields

Abstract

The present invention provides a method and system for reconstructing a rotating sound source in a non-uniform flow field. The method divides an acoustic signal into signal segments, constructs a time-domain sound pressure matrix from the signal segments, obtains rotational modal coefficients based on the time-domain sound pressure matrix, reconstructs the signal in the signal segment based on the synchronous rotational modal coefficients to form a reconstructed signal, and extracts signal components of specific frequencies in the reconstructed signal to form a sound pressure vector. An acoustic inverse problem is constructed based on the sound pressure vector, and the acoustic inverse problem is solved to obtain a sound source intensity vector to be determined. Reconstruction of the rotating sound source is achieved based on the sound source intensity vector. In this way, sound source reconstruction can be performed under the condition of a non-uniform flow field, thereby expanding the application range and reliability of sound source reconstruction, reducing numerical calculation errors, and improving the accuracy of rotational sound source reconstruction. High-resolution reconstruction results can also be obtained in a noisy environment, thereby improving the practicality of sound source reconstruction.

Description

Technical Field

[0001] The present invention relates to the field of array signal processing technology, and more specifically, to a method for reconstructing a rotating sound source in a non-uniform flow field, and in particular to a method for reconstructing a rotating sound source based on array measurement under non-uniform flow conditions. Background Art

[0002] Rotational sound sources are widely present in industrial production processes, such as the meshing impact noise of planetary gears in planetary gearboxes and the aerodynamic noise caused by the rotating blades of axial-flow fans. The characteristics of these sound sources reflect the operating conditions of mechanical equipment. Therefore, reconstructing these sound sources can monitor the equipment's operating status in a non-contact manner and provide a reference for optimizing the aerodynamic performance of mechanical structures, thereby improving the efficiency and reliability of industrial production processes. A microphone array, composed of a series of microphones arranged in a specific pattern, can spatially sample the sound source. Appropriate array signal processing methods can be used to perform spatial filtering of the sound source, thereby reconstructing the sound source in the sound image and generating intuitive sound source identification results in the form of images. Commonly used techniques include beamforming and subspace methods, which can reconstruct stationary sound sources. However, due to the assumption of signal and environmental stationarity in sound propagation models, the identification results of rotational sound sources suffer from issues such as smearing and smearing, and the accuracy of sound source reconstruction is affected under non-uniform flow conditions. Therefore, these techniques alone cannot accurately identify rotational sound sources under non-uniform flow conditions.

[0003] To overcome the shortcomings of the aforementioned techniques' assumption of source stationarity, techniques such as virtual rotating arrays (RP Dougherty, BE Walker, Virtual rotating microphone imaging of broadband fan noise, 15th AIAA / CEAS Aeroacoustics Conference, 2009) and transient signal segmentation have been proposed. These techniques can, under certain conditions, overcome the adverse effects of the source's rotational motion on source reconstruction. However, they all assume that the source is located in an ideally homogeneous medium, without flow. In reality, rotating sound sources may be located in a flowing environment. For example, the rotation of components such as rotors and blades during the operation of rotating machinery can cause air flow. The influence of flow effects is not reflected in the sound propagation models of these techniques. This drawback can lead to significant discrepancies between the sound propagation characteristics used for source identification and the actual situation, making accurate reconstruction of rotating sound sources impossible. Furthermore, virtual rotating arrays interpolate discrete time-domain signals, resulting in potentially large numerical errors and limited accuracy. The length of the signal segments in transient signal segmentation affects the accuracy of the source reconstruction results, but selecting an appropriate segment length remains a challenge.

[0004] In order to introduce the influence of medium flow in the process of sound source reconstruction, the method of reconstructing rotating sound sources under uniform flow conditions (C. Ocker, W. Pannert, Imaging of broadband noise from rotating sources in uniform axial flow, Aiaa J, 55(4)(2017)1185-1193) has been proposed in recent years. This method approximates the sound propagation process under uniform flow field, thereby eliminating the interference caused by uniform flow field in the process of solving the acoustic inverse problem. However, the uniform flow field condition is an idealized assumption and is difficult to meet in practice. In the real world, non-uniform flow fields are more common. Therefore, the reconstruction of rotating sound sources under non-uniform flow conditions is of great significance. The existing technology has not yet considered the propagation law of rotating sound sources under non-uniform flow conditions. The imaging resolution of the beamforming and deconvolution methods used is low, and the accuracy of the results in strong noise environments is insufficient. These problems restrict the effectiveness of this technology in identifying rotating sound sources under actual conditions. Based on the above analysis, it can be seen that accurately reconstructing rotating sound sources under non-uniform flow conditions is still an unresolved problem.

[0005] Therefore, there is an urgent need for a method for reconstructing rotating sound sources in non-uniform flow fields that can expand the application scope and reliability of sound source reconstruction, reduce numerical calculation errors, improve the accuracy of rotational motion sound source reconstruction, and improve the practicality of sound source reconstruction. Summary of the Invention

[0006] In view of the above problems, the purpose of the present invention is to provide a method for reconstructing rotating sound sources in non-uniform flow fields, so as to solve the problems that the existing technology has not considered the propagation law of rotating sound sources under non-uniform flow conditions, the imaging resolution of the beamforming and deconvolution methods adopted is low, the results are not accurate enough in strong noise environments, and the effectiveness is low.

[0007] The present invention provides a method for reconstructing a rotating sound source in a non-uniform flow field, comprising:

[0008] The sound source of the rotational motion in the non-uniform flow field is collected by a preset microphone array to form an acoustic signal, and the acoustic signal is segmented to obtain signal segments, so as to construct a time-domain sound pressure matrix according to the signal segments;

[0009] Obtaining rotational modal coefficients according to the time-domain sound pressure matrix, and performing frequency-domain offset processing on the rotational modal coefficients according to the pre-acquired rotation speed of the sound source to obtain synchronous rotational modal coefficients of a coordinate system that rotates synchronously with the sound source;

[0010] Reconstructing the signal in the signal segment according to the synchronous rotation modal coefficient to form a reconstructed signal, and extracting a signal component of a specific frequency in the reconstructed signal to form a sound pressure vector;

[0011] An acoustic inverse problem is constructed based on the sound pressure vector, and the acoustic inverse problem is solved to obtain a sound source intensity vector to be determined, and the reconstruction of the rotational motion sound source is achieved according to the sound source intensity vector.

[0012] Preferably, the process of collecting a sound source of rotational motion in a non-uniform flow field by a preset microphone array to form an acoustic signal, segmenting the acoustic signal to obtain signal segments, and constructing a time-domain sound pressure matrix based on the signal segments includes:

[0013] Determining a sound source plane according to a motion path of the sound source rotating in the non-uniform flow field;

[0014] Collecting sound signals at a position parallel to the sound source plane using a microphone array, and establishing an original sound pressure matrix based on the collected sound signals;

[0015] Segmenting the acoustic signal collected by each channel in the microphone array into signal segments of preset length and preset overlap rate based on the original sound pressure matrix, and calculating the number of the signal segments;

[0016] A time-domain sound pressure matrix is ​​constructed from the signal segments based on the number of the signal segments.

[0017] Preferably, the process of obtaining the rotational modal coefficients according to the time-domain sound pressure matrix includes:

[0018] Acquire the time-domain sound pressure of the signal segment based on the time-domain sound pressure matrix;

[0019] Converting the time-domain sound pressure of the signal segment into the frequency domain by discrete Fourier transform to obtain a frequency-domain signal;

[0020] The frequency domain signal is converted to the modal domain through modal decomposition to obtain rotational modal coefficients.

[0021] Preferably, the process of constructing an acoustic inverse problem based on the sound pressure vector and solving the acoustic inverse problem to obtain the sound source intensity vector to be determined includes:

[0022] performing discretization processing on a reconstruction region of the reconstructed signal to form a reconstruction grid;

[0023] Performing Fourier series expansion on the positions of the microphones in the microphone array and the positions of the reconstruction grid to obtain series coefficients;

[0024] Establishing an acoustic inverse problem based on the series coefficients and the sound pressure vector;

[0025] The acoustic inverse problem is solved based on a predetermined regularization parameter to obtain a desired sound source intensity vector.

[0026] Preferably, the process of discretizing the reconstruction area of ​​the reconstructed signal to form a reconstruction grid; and performing Fourier series expansion on the positions of the microphones in the microphone array and the positions of the reconstruction grid to obtain series coefficients includes:

[0027] Discretizing the sound source reconstruction region of the reconstructed signal to divide it into reconstruction grids;

[0028] establishing a rectangular coordinate system and a cylindrical coordinate system based on the reconstructed grid;

[0029] Acquiring position coordinates of a reconstructed grid from the position coordinates of microphones in the microphone array according to a mapping relationship between the rectangular coordinate system and the cylindrical coordinate system;

[0030] Fourier series expansion is performed on the velocity potential of the position coordinates of the microphone and the position coordinates of the reconstruction grid to obtain microphone series coefficients and reconstruction grid series coefficients.

[0031] Preferably, the process of establishing an acoustic inverse problem based on the series coefficients and the sound pressure vector includes:

[0032] Formulate the inverse acoustic problem for nonuniform flow conditions in a rotating coordinate system:

[0033] p=Gq+e

[0034] Where q=[q1,q2,…,q N ] T is the sound source intensity vector to be determined, representing the sound source intensity of each element at the reconstructed grid point in the reconstructed grid; e=[e1,e2,…,e M ] T is a noise vector; G is a transmission matrix of size M×N obtained based on the microphone series coefficients and the reconstruction grid series coefficients.

[0035] Preferably, the elements included in the transmission matrix are formed based on equivalent wave numbers, harmonic terms, spherical Bessel functions of the first kind, spherical Hankel functions of the second kind, and associated Legendre functions.

[0036] Preferably, an optimization algorithm is used to solve the acoustic inverse problem based on a predetermined regularization parameter to obtain the desired sound source intensity vector.

[0037] Preferably, the optimization algorithm adopts coordinate descent method.

[0038] On the other hand, the present invention also provides a non-uniform flow field rotation sound source reconstruction system, which implements the non-uniform flow field rotation sound source reconstruction method as described above, comprising:

[0039] a signal segmentation unit, configured to collect a sound source of rotational motion in a non-uniform flow field through a preset microphone array to form an acoustic signal, and segment the acoustic signal to obtain signal segments, so as to construct a time-domain sound pressure matrix based on the signal segments;

[0040] a frequency domain offset unit, configured to obtain rotational modal coefficients according to the time domain sound pressure matrix, and perform frequency domain offset processing on the rotational modal coefficients according to the pre-acquired rotation speed of the sound source to obtain synchronous rotational modal coefficients of a coordinate system that rotates synchronously with the sound source;

[0041] a signal reconstruction unit, configured to reconstruct the signal in the signal segment according to the synchronous rotation modal coefficient to form a reconstructed signal, and extract a signal component of a specific frequency in the reconstructed signal to form a sound pressure vector;

[0042] A sound source reconstruction unit is used to construct an acoustic inverse problem based on the sound pressure vector, solve the acoustic inverse problem to obtain a sound source intensity vector to be determined, and reconstruct the rotational motion sound source according to the sound source intensity vector.

[0043] From the above technical solution, it can be seen that the method for reconstructing the rotating sound source of the non-uniform flow field provided by the present invention first collects the sound source of the rotating motion in the non-uniform flow field through a preset microphone array to form an acoustic signal, and divides the acoustic signal to obtain signal segments, and constructs a time domain sound pressure matrix from the signal segments, and then obtains the rotational modal coefficients according to the time domain sound pressure matrix, and performs frequency domain offset processing on the rotational modal coefficients according to the rotation speed of the pre-acquired sound source to obtain the synchronous rotational modal coefficients of the coordinate system that rotates synchronously with the sound source; then reconstructs the signal in the signal segment according to the synchronous rotational modal coefficients to form a reconstructed signal, and extracts the signal components of specific frequencies in the reconstructed signal to form a sound pressure vector; constructs an acoustic inverse problem based on the sound pressure vector, and solves the acoustic inverse problem to obtain the sound to be obtained. Source intensity vector, the reconstruction of the rotational motion sound source is realized according to the sound source intensity vector. In this way, considering the influence of the non-uniform flow field, the velocity potential of the sound transmission medium is approximately calculated, so that the sound source can be reconstructed under the condition of the non-uniform flow field, overcoming the defect that the existing technology is not suitable for non-uniform flow field conditions, expanding the application scope and reliability of sound source reconstruction, processing the signals collected by the microphone array, overcoming the signal non-stationarity problem caused by the sound source rotation for sound source identification in the modal domain, making the numerical calculation error small, which is conducive to improving the accuracy of the rotational motion sound source reconstruction; and using sparse constraints to solve the acoustic inverse problem, it is possible to reconstruct the rotational motion sound source under non-uniform flow conditions, and obtain high-resolution reconstruction results in a noisy environment, thereby improving the practicability of the sound source reconstruction. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] By referring to the following description in conjunction with the accompanying drawings, and with a more complete understanding of the present invention, other objects and results of the present invention will become more apparent and easier to understand. In the accompanying drawings:

[0045] Figure 1 Flow chart of a method for reconstructing a rotating sound source in a non-uniform flow field according to an embodiment of the present invention;

[0046] Figure 2 Schematic diagram of a model of an acoustic inverse problem in a method for reconstructing a rotating sound source in a non-uniform flow field according to an embodiment of the present invention;

[0047] Figure 3 is a numerical simulation condition in a method for reconstructing a rotating sound source in a non-uniform flow field according to an embodiment of the present invention;

[0048] Figure 4 4 is a system block diagram of a method system for reconstructing a rotating sound source in a non-uniform flow field according to an embodiment of the present invention. DETAILED DESCRIPTION

[0049] Currently, in real-world conditions, the sound source of rotational motion may be located in a fluid environment. For example, the rotation of components such as rotors and blades during the operation of rotating machinery can cause air movement. However, the influence of this flow effect is not reflected in the sound propagation models used in the aforementioned technologies. This deficiency can lead to significant discrepancies between the sound propagation characteristics used for sound source identification and the actual situation, making it impossible to accurately reconstruct the sound source of rotational motion. Furthermore, the virtual rotation array interpolates discrete time-domain signals, resulting in potentially large numerical errors and difficulty in ensuring accuracy. Furthermore, the length of the signal segments used in instantaneous signal segmentation affects the accuracy of the sound source reconstruction results.

[0050] In order to solve the above problems, the present invention provides a method for reconstructing a rotating sound source in a non-uniform flow field. The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0051] In order to illustrate the non-uniform flow field rotation sound source reconstruction method and system provided by the present invention, Figure 1-Figure 3 The method for reconstructing the rotating sound source of the non-uniform flow field according to the embodiment of the present invention is exemplarily illustrated; Figure 4 The non-uniform flow field rotating sound source reconstruction system according to the embodiment of the present invention is exemplarily illustrated.

[0052] The following description of exemplary embodiments is merely illustrative in nature and is in no way intended to limit the present invention, its application, or uses. Techniques and devices known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques and devices should be considered part of the specification.

[0053] like Figure 1 、 Figure 2 、 Figure 3 As shown in the figure, the present invention provides a method for reconstructing a rotating sound source in a non-uniform flow field, comprising:

[0054] S1: collecting a sound source of rotational motion in a non-uniform flow field through a preset microphone array to form an acoustic signal, and segmenting the acoustic signal to obtain signal segments, so as to construct a time-domain sound pressure matrix based on the signal segments;

[0055] S2: Obtaining rotational modal coefficients according to the time-domain sound pressure matrix, and performing frequency-domain offset processing on the rotational modal coefficients according to the pre-acquired rotation speed of the sound source to obtain synchronous rotational modal coefficients of a coordinate system that rotates synchronously with the sound source;

[0056] S3: reconstructing the signal in the signal segment according to the synchronous rotation modal coefficient to form a reconstructed signal, and extracting a signal component of a specific frequency in the reconstructed signal to form a sound pressure vector;

[0057] S4: constructing an acoustic inverse problem based on the sound pressure vector, solving the acoustic inverse problem to obtain a desired sound source intensity vector, and reconstructing the rotational motion sound source according to the sound source intensity vector.

[0058] exist Figure 1 、 Figure 2 、 Figure 3 , in the embodiment shown together, step S1 is a process of collecting a sound source of rotational motion in a non-uniform flow field through a preset microphone array to form an acoustic signal, and segmenting the acoustic signal to obtain signal segments, so as to construct a time domain sound pressure matrix according to the signal segments; wherein,

[0059] The process of collecting a sound source of rotational motion in a non-uniform flow field by a preset microphone array to form an acoustic signal, segmenting the acoustic signal to obtain signal segments, and constructing a time-domain sound pressure matrix based on the signal segments includes:

[0060] S11: determining a sound source plane according to a motion path of the sound source rotating in the non-uniform flow field;

[0061] S12: collecting sound signals at a position parallel to the sound source plane through a microphone array, and establishing an original sound pressure matrix according to the collected sound signals;

[0062] S13: dividing the acoustic signal collected by each channel in the microphone array into signal segments with a preset length and a preset overlap rate based on the original sound pressure matrix, and calculating the number of the signal segments;

[0063] S14: Constructing a time-domain sound pressure matrix from the signal segments based on the number of the signal segments.

[0064] In a specific embodiment, step S1 is to collect signals using a microphone array at a position parallel to the sound source plane according to the motion path of the rotating sound source in the non-uniform flow field, establish an original sound pressure matrix, and divide the signals collected by each channel into multiple signal segments by overlapping segmentation to construct a time domain sound pressure matrix; wherein, first, step (1-a) is performed: according to the motion path of the rotating sound source in the non-uniform flow field, the plane where the sound source is located is determined, and a microphone array containing M array elements is used at a position parallel to the sound source plane to collect signals, and obtain an acoustic signal p of length A collected by each channel om (m=1,2,…,M), establish the original sound pressure matrix P=[p o1 ,p o2 ,…,p oM ] T Then proceed to step (1-b): divide the signals collected by each channel of the array into multiple signal segments with a length of L and an overlap rate of B, and calculate the number of signal segments in represents the smallest integer not less than x; after segmentation, the cth (c=1,2,…,C)th signal segment collected by the mth microphone is Indicates p om The L(1-B)(c-1)+1th to L(1-B)(c-1)+Lth elements in the segmented signal are obtained; then step (1-c): the segmented signal segment p c,m Construct the time domain sound pressure matrix:

[0065]

[0066] exist Figure 1 、 Figure 2 、 Figure 3 In the embodiment shown in the figure, step S2 is a process of obtaining rotational modal coefficients according to the time-domain sound pressure matrix, and performing frequency-domain offset processing on the rotational modal coefficients according to the pre-acquired rotation speed of the sound source to obtain synchronous rotational modal coefficients of a coordinate system that rotates synchronously with the sound source; wherein,

[0067] The process of obtaining the rotational modal coefficients according to the time-domain sound pressure matrix includes:

[0068] S211: Acquire the time-domain sound pressure of the signal segment based on the time-domain sound pressure matrix;

[0069] S212: Converting the time-domain sound pressure of the signal segment into the frequency domain by discrete Fourier transform to obtain a frequency-domain signal;

[0070] S213: Convert the frequency domain signal to the modal domain through modal decomposition to obtain rotational modal coefficients.

[0071] Specifically, in this embodiment, discrete Fourier transform is first used to convert each signal segment into the frequency domain The signal is then converted to the modal domain through modal decomposition to obtain the rotational modal coefficients:

[0072]

[0073] in, is the wave number, c is the speed of sound, exp() is the exponential function, Is an imaginary unit.

[0074] In the process of performing frequency domain offset processing on the rotational modal coefficients according to the pre-acquired rotation speed of the sound source to obtain the synchronous rotational modal coefficients of the coordinate system rotating synchronously with the sound source, the rotational modal coefficients are frequency offset according to the sound source rotation speed Ω to obtain the rotational modal coefficients in the coordinate system rotating synchronously with the sound source d

[0075] exist Figure 1、 Figure 2 、 Figure 3 In the embodiment shown together, step S3 is a process of reconstructing the signal in the signal segment according to the synchronous rotation modal coefficient to form a reconstructed signal, and extracting the signal component of a specific frequency in the reconstructed signal to form a sound pressure vector.

[0076] Specifically, the rotational modal coefficients obtained in step S2 are used to reconstruct the signals in each signal segment to obtain a reconstructed signal in a coordinate system that rotates synchronously with the sound source, and the signal components of specific frequencies in the reconstructed signal are extracted to form a sound pressure vector.

[0077] First, the sound source signal is reconstructed using the rotational modal coefficients in the coordinate system that rotates synchronously with the sound source, and the reconstructed signal in the coordinate system that rotates synchronously with the sound source is obtained.

[0078] Then extract the frequency f in the cth signal segment c The corresponding channel signal components constitute the sound pressure vector

[0079] exist Figure 1 、 Figure 2 、 Figure 3 In the embodiment shown together, step S4 is a process of constructing an acoustic inverse problem based on the sound pressure vector, solving the acoustic inverse problem to obtain the sound source intensity vector to be determined, and reconstructing the rotational motion sound source according to the sound source intensity vector; wherein,

[0080] The process of constructing an acoustic inverse problem based on the sound pressure vector and solving the acoustic inverse problem to obtain the sound source intensity vector to be determined includes:

[0081] S41: performing discretization processing on the reconstruction area of ​​the reconstructed signal to form a reconstruction grid;

[0082] S42: Performing Fourier series expansion on the positions of the microphones in the microphone array and the positions of the reconstruction grid to obtain series coefficients;

[0083] S43: establishing an acoustic inverse problem based on the series coefficient and the sound pressure vector;

[0084] S44: Solve the acoustic inverse problem based on a predetermined regularization parameter to obtain a desired sound source intensity vector.

[0085] The process of discretizing the reconstruction area of ​​the reconstructed signal to form a reconstruction grid; and performing Fourier series expansion on the positions of the microphones in the microphone array and the positions of the reconstruction grid to obtain series coefficients includes:

[0086] S421: Discretize the sound source reconstruction area of ​​the reconstructed signal to divide it into reconstruction grids;

[0087] S422: Establishing a rectangular coordinate system and a cylindrical coordinate system based on the reconstructed grid;

[0088] S423: Obtaining position coordinates of a reconstructed grid from the position coordinates of microphones in the microphone array according to a mapping relationship between the rectangular coordinate system and the cylindrical coordinate system;

[0089] S424: Performing Fourier series expansion on the velocity potential of the position coordinates of the microphone and the position coordinates of the reconstructed grid to obtain microphone series coefficients and reconstructed grid series coefficients.

[0090] The process of establishing an acoustic inverse problem based on the series coefficients and the sound pressure vector includes:

[0091] S431: Establishing the inverse acoustic problem under non-uniform flow conditions in a rotating coordinate system:

[0092] p=Gq+e

[0093] Where q=[q1,q2,…,q N ] T is the sound source intensity vector to be determined, representing the sound source intensity of each element at the reconstructed grid point in the reconstructed grid; e=[e1,e2,…,e M ] T is a noise vector; G is a transmission matrix of size M×N obtained based on the microphone series coefficients and the reconstruction grid series coefficients.

[0094] The elements contained in the transmission matrix are formed based on equivalent wave numbers, harmonic terms, spherical Bessel functions of the first kind, spherical Hankel functions of the second kind, and associated Legendre functions.

[0095] In this embodiment, an optimization algorithm is used to solve the acoustic inverse problem based on a predetermined regularization parameter to obtain the desired sound source intensity vector; in a specific embodiment, the optimization algorithm uses a coordinate descent method.

[0096] More specifically, step S4 is a process of discretizing the sound source reconstruction area, performing a Fourier series expansion on the velocity potential at the microphone location and the reconstruction grid location, calculating the equivalent wave number and harmonic terms in the cylindrical coordinate system, establishing an inverse acoustic problem of a rotating sound source under non-uniform flow conditions, and solving it using a sparse regularization method to achieve accurate reconstruction of the rotating sound source under non-uniform flow conditions; wherein the specific steps are:

[0097] Step (4-a): Discretize the sound source reconstruction area and divide it into N reconstruction grids, and establish Figure 2 According to the relationship between the rectangular coordinate system and the cylindrical coordinate system, the position coordinate of the microphone in the microphone array whose center is at the origin of the coordinate is r m =(R m cosθ m sinφ m ,R m cosθ m cosφ m ,R m sinθ m ), the position coordinate of the reconstructed grid is r' n =(R' n cosθ' n sinφ' n ,R' n cosθ' n cosφ' n ,R' n sinθ' n )(n=1,2,…,N). Perform Fourier series expansion on the velocity potential at the microphone location:

[0098]

[0099] Among them, the series coefficient

[0100] Perform Fourier series expansion on the velocity potential at the location of the reconstructed grid:

[0101]

[0102] Among them, the series coefficient

[0103] Step (4-b): Let the flow field Mach number be Ma, c0=1, c β =–jkMaФ β '(β≠0),d0=1,d β =jkMaФ –β (β≠0), the inverse acoustic problem under non-uniform flow conditions is established in the rotating coordinate system:

[0104] p=Gq+e (5)

[0105] Where q=[q1,q2,…,q N ] T is the sound source intensity vector, each element represents the sound source intensity of the reconstructed grid point, e=[e1,e2,…,e M ] Tis the noise vector, and the transmission matrix G of size M×N contains the elements:

[0106]

[0107] Among them, the equivalent wave number Harmonic terms ξ n () is the nth order spherical Bessel function of the first kind, is the nth-order spherical Hankel function of the second kind, and the polynomial product term is the |α|, nth-order associated Legendre function.

[0108] Step (4-c): Determine the regularization parameter λ>0 and use the 1-norm sparse regularization method to solve the inverse problem:

[0109]

[0110] In the specific implementation process, the solution to the above inverse problem can be completed by using optimization algorithms such as the coordinate descent method, so as to solve the sound source intensity vector q and realize the reconstruction of the rotating motion sound source under non-uniform flow conditions. In this way, the sound source intensity vector is obtained, and the reconstruction of the rotating motion sound source is realized.

[0111] It can be seen that the non-uniform flow field rotation sound source reconstruction method in the embodiment of the present invention takes into account the influence of the non-uniform flow field in the modeling process of the acoustic inverse problem, and approximately calculates the velocity potential of the sound transmission medium, so that the sound source can be reconstructed under the condition of non-uniform flow field, overcoming the defect that the existing technology is not suitable for non-uniform flow field conditions, and expanding the application scope and reliability of sound source reconstruction; the signal collected by the microphone array is processed in the form of modal decomposition and reconstruction, and the signal non-stationarity problem caused by the sound source rotation for sound source identification is overcome in the modal domain. Compared with the existing rotation sound source reconstruction technology, it has a more complete theoretical basis and small numerical calculation error, which is conducive to improving the accuracy of rotational motion sound source reconstruction; sparse constraints are used to solve the acoustic inverse problem, which can reconstruct the rotational motion sound source under non-uniform flow conditions, and obtain high-resolution reconstruction results in a noisy environment, thereby improving the practicality of sound source reconstruction.

[0112] In summary, the method for reconstructing a rotating sound source in a non-uniform flow field provided by the present invention first collects the rotating sound source in the non-uniform flow field through a preset microphone array to form a sound signal, and divides the sound signal to obtain signal segments, and constructs a time domain sound pressure matrix from the signal segments, and then obtains the rotational modal coefficients according to the time domain sound pressure matrix, and performs frequency domain offset processing on the rotational modal coefficients according to the pre-acquired rotation speed of the sound source to obtain the synchronous rotational modal coefficients of the coordinate system that rotates synchronously with the sound source; then reconstructs the signal in the signal segment according to the synchronous rotational modal coefficients to form a reconstructed signal, and extracts the signal components of specific frequencies in the reconstructed signal to form a sound pressure vector; constructs an acoustic inverse problem based on the sound pressure vector, and solves the acoustic inverse problem to obtain the sound source intensity to be determined Vector, the reconstruction of the rotational sound source is realized according to the sound source intensity vector. In this way, the influence of the non-uniform flow field is taken into consideration, and the velocity potential of the sound transmission medium is approximately calculated, so that the sound source can be reconstructed under the condition of the non-uniform flow field, overcoming the defect that the existing technology is not suitable for non-uniform flow field conditions, expanding the application scope and reliability of the sound source reconstruction, processing the signals collected by the microphone array, overcoming the signal non-stationarity problem caused by the sound source rotation for sound source identification in the modal domain, making the numerical calculation error small, which is conducive to improving the accuracy of the rotational sound source reconstruction; and using sparse constraints to solve the acoustic inverse problem, it is possible to reconstruct the rotational sound source under non-uniform flow conditions, and obtain high-resolution reconstruction results in a noisy environment, thereby improving the practicability of the sound source reconstruction.

[0113] like Figure 4 As shown, the present invention further provides a non-uniform flow field rotation sound source reconstruction system 100, which implements the non-uniform flow field rotation sound source reconstruction method as described above, including:

[0114] The signal segmentation unit 101 is configured to collect a sound source of rotational motion in a non-uniform flow field using a preset microphone array to form an acoustic signal, and segment the acoustic signal to obtain signal segments, so as to construct a time-domain sound pressure matrix based on the signal segments;

[0115] a frequency domain offset unit 102 for acquiring rotational modal coefficients based on the time domain sound pressure matrix, and performing frequency domain offset processing on the rotational modal coefficients based on the pre-acquired rotation speed of the sound source to acquire synchronous rotational modal coefficients of a coordinate system that rotates synchronously with the sound source;

[0116] a signal reconstruction unit 103, configured to reconstruct the signal in the signal segment according to the synchronous rotation modal coefficient to form a reconstructed signal, and extract a signal component of a specific frequency in the reconstructed signal to form a sound pressure vector;

[0117] The sound source reconstruction unit 104 is configured to construct an acoustic inverse problem based on the sound pressure vector, solve the acoustic inverse problem to obtain a desired sound source intensity vector, and reconstruct the rotational motion sound source according to the sound source intensity vector.

[0118] The specific implementation method is not particularly limited and will not be described in detail here. Please refer to the specific embodiment of the method for reconstructing a rotating sound source in a non-uniform flow field.

[0119] As described above, the non-uniform flow field rotating sound source reconstruction system provided by the present invention first uses the signal segmentation unit 101 to collect the rotating sound source in the non-uniform flow field through a preset microphone array to form an acoustic signal, and segments the acoustic signal to obtain signal segments, and constructs a time domain sound pressure matrix from the signal segments, and then obtains the rotational modal coefficients according to the time domain sound pressure matrix through the frequency domain offset unit 102, and performs frequency domain offset processing on the rotational modal coefficients according to the pre-acquired rotation speed of the sound source to obtain the synchronous rotational modal coefficients of the coordinate system that rotates synchronously with the sound source; then, based on the signal reconstruction unit 103, the signal in the signal segment is reconstructed according to the synchronous rotational modal coefficients to form a reconstructed signal, and the signal components of specific frequencies in the reconstructed signal are extracted to form a sound pressure vector; and then the acoustic inverse problem is constructed based on the sound pressure vector through the sound source reconstruction unit 104. , and solve the acoustic inverse problem to obtain the sound source intensity vector to be determined, and realize the reconstruction of the rotational motion sound source according to the sound source intensity vector. In this way, considering the influence of the non-uniform flow field, the velocity potential of the sound transmission medium is approximately calculated, so that the sound source can be reconstructed under the condition of the existence of a non-uniform flow field, overcoming the defect that the existing technology is not suitable for non-uniform flow field conditions, expanding the application scope and reliability of the sound source reconstruction, processing the signals collected by the microphone array, and overcoming the signal non-stationarity problem caused by the sound source rotation for sound source identification in the modal domain, so that the numerical calculation error is small, which is conducive to improving the accuracy of the rotational motion sound source reconstruction; and using sparse constraints to solve the acoustic inverse problem, it is possible to reconstruct the rotational motion sound source under non-uniform flow conditions, and obtain high-resolution reconstruction results in a noisy environment, thereby improving the practicality of the sound source reconstruction.

[0120] The above description uses the accompanying drawings to illustrate the method and system for reconstructing rotating sound sources in a non-uniform flow field according to the present invention. However, those skilled in the art will appreciate that various modifications may be made to the method and system for reconstructing rotating sound sources in a non-uniform flow field without departing from the scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A method for reconstructing rotating sound sources in a non-uniform flow field, characterized in that: include: The sound source of the rotational motion in the non-uniform flow field is collected by a preset microphone array to form an acoustic signal, and the acoustic signal is segmented to obtain signal segments, so as to construct a time-domain sound pressure matrix according to the signal segments; Obtaining rotational modal coefficients according to the time-domain sound pressure matrix, and performing frequency-domain offset processing on the rotational modal coefficients according to the pre-acquired rotation speed of the sound source to obtain synchronous rotational modal coefficients of a coordinate system that rotates synchronously with the sound source; Reconstructing the signal in the signal segment according to the synchronous rotation modal coefficient to form a reconstructed signal, and extracting a signal component of a specific frequency in the reconstructed signal to form a sound pressure vector; An acoustic inverse problem is constructed based on the sound pressure vector, and the acoustic inverse problem is solved to obtain a sound source intensity vector to be determined, and the reconstruction of the rotational motion sound source is achieved according to the sound source intensity vector.

2. The method for reconstructing a rotating sound source in a non-uniform flow field according to claim 1, wherein: The process of collecting a sound source of rotational motion in a non-uniform flow field by a preset microphone array to form an acoustic signal, segmenting the acoustic signal to obtain signal segments, and constructing a time-domain sound pressure matrix based on the signal segments includes: Determining a sound source plane according to a motion path of the sound source rotating in the non-uniform flow field; Collecting sound signals at a position parallel to the sound source plane using a microphone array, and establishing an original sound pressure matrix based on the collected sound signals; Segmenting the acoustic signal collected by each channel in the microphone array into signal segments of preset length and preset overlap rate based on the original sound pressure matrix, and calculating the number of the signal segments; A time-domain sound pressure matrix is ​​constructed from the signal segments based on the number of the signal segments.

3. The method for reconstructing a rotating sound source in a non-uniform flow field according to claim 2, wherein: The process of obtaining the rotational modal coefficients according to the time-domain sound pressure matrix includes: Acquire the time-domain sound pressure of the signal segment based on the time-domain sound pressure matrix; Converting the time-domain sound pressure of the signal segment into the frequency domain by discrete Fourier transform to obtain a frequency-domain signal; The frequency domain signal is converted to the modal domain through modal decomposition to obtain rotational modal coefficients.

4. The method for reconstructing a rotating sound source in a non-uniform flow field according to claim 1, wherein: The process of constructing an acoustic inverse problem based on the sound pressure vector and solving the acoustic inverse problem to obtain the sound source intensity vector to be determined includes: performing discretization processing on a reconstruction region of the reconstructed signal to form a reconstruction grid; Performing Fourier series expansion on the positions of the microphones in the microphone array and the positions of the reconstruction grid to obtain series coefficients; Establishing an acoustic inverse problem based on the series coefficients and the sound pressure vector; The acoustic inverse problem is solved based on a predetermined regularization parameter to obtain a desired sound source intensity vector.

5. The method for reconstructing a rotating sound source in a non-uniform flow field according to claim 4, wherein: performing discretization processing on a reconstruction region of the reconstructed signal to form a reconstruction grid; The process of performing Fourier series expansion on the positions of the microphones in the microphone array and the positions of the reconstruction grid to obtain series coefficients includes: Discretizing the sound source reconstruction region of the reconstructed signal to divide it into reconstruction grids; establishing a rectangular coordinate system and a cylindrical coordinate system based on the reconstructed grid; Acquiring position coordinates of a reconstructed grid from the position coordinates of microphones in the microphone array according to a mapping relationship between the rectangular coordinate system and the cylindrical coordinate system; Fourier series expansion is performed on the velocity potential of the position coordinates of the microphone and the position coordinates of the reconstruction grid to obtain microphone series coefficients and reconstruction grid series coefficients.

6. The method for reconstructing a rotating sound source in a non-uniform flow field according to claim 4, wherein: The process of establishing an acoustic inverse problem based on the series coefficients and the sound pressure vector includes: Formulate the inverse acoustic problem for nonuniform flow conditions in a rotating coordinate system: p=Gq+e Where q=[q1,q2,…,q N ] T is the sound source intensity vector to be determined, representing the sound source intensity of each element at the reconstructed grid point in the reconstructed grid; e=[e1,e2,…,e M ] T is a noise vector; G is a transmission matrix of size M×N obtained based on the microphone series coefficients and the reconstruction grid series coefficients.

7. The method for reconstructing a rotating sound source in a non-uniform flow field according to claim 6, wherein: The elements included in the transmission matrix are formed based on equivalent wave numbers, harmonic terms, spherical Bessel functions of the first kind, spherical Hankel functions of the second kind, and associated Legendre functions.

8. The method for reconstructing a rotating sound source in a non-uniform flow field according to claim 4, wherein: An optimization algorithm is used to solve the acoustic inverse problem based on a predetermined regularization parameter to obtain the desired sound source intensity vector.

9. The method for reconstructing a rotating sound source in a non-uniform flow field according to claim 8, wherein: The optimization algorithm adopts coordinate descent method.

10. A non-uniform flow field rotation sound source reconstruction system, characterized in that: The method for reconstructing a rotating sound source in a non-uniform flow field according to any one of claims 1 to 9 comprises: a signal segmentation unit, configured to collect a sound source of rotational motion in a non-uniform flow field through a preset microphone array to form an acoustic signal, and segment the acoustic signal to obtain signal segments, so as to construct a time-domain sound pressure matrix based on the signal segments; a frequency domain offset unit, configured to obtain rotational modal coefficients according to the time domain sound pressure matrix, and perform frequency domain offset processing on the rotational modal coefficients according to the pre-acquired rotation speed of the sound source to obtain synchronous rotational modal coefficients of a coordinate system that rotates synchronously with the sound source; a signal reconstruction unit, configured to reconstruct the signal in the signal segment according to the synchronous rotation modal coefficient to form a reconstructed signal, and extract a signal component of a specific frequency in the reconstructed signal to form a sound pressure vector; A sound source reconstruction unit is used to construct an acoustic inverse problem based on the sound pressure vector, solve the acoustic inverse problem to obtain a sound source intensity vector to be determined, and reconstruct the rotational motion sound source according to the sound source intensity vector.

Images