Method and system for identifying rotational motion sound source in non-uniform flow field
The signal is collected through the microphone array and the dynamic coordinate system conversion and acoustic inverse problems are solved, and the problem of rotating motion sound source identification in the non-uniform flow field is solved, and the identification accuracy and resolution are improved, which is suitable for actual conditions and noise environments.
Patent Information
- Application Number
- CN202210222795.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-07
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2042-03-07
AI Technical Summary
In the prior art, when identifying rotating motion sound sources in non-uniform flow fields, the signal segment length is difficult to determine, the modal decomposition efficiency is low, the uniform flow field model is not suitable for actual conditions, the deconvolution method has a low recognition resolution and is Poor accuracy under noise conditions.
The signal is collected using a microphone array, the angle compensation value is calculated according to the sound source speed, the signal is interpolated and converted to the dynamic coordinate system, and the acoustic inverse problem under non-uniform flow conditions is constructed by discrete Fourier transform, and the sound source position is solved by regularization parameters. and strength.
High resolution rotary motion sound source identification in the non-uniform flow field is realized, which eliminates the impact of sound source movement on signal stability, improves identification accuracy, and is suitable for actual non-uniform flow field and noise-containing conditions.
Smart Images

Figure CN115267670B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of acoustic signal processing, and more specifically, to a method and system for identifying a rotational motion sound source in a non-uniform flow field. Background Art
[0002] The sound source identification technology based on microphone array measurement has broad application prospects, such as noise source location, mechanical health monitoring, aerodynamic performance experiments, etc. This technology can generate intuitive sound images through array signal processing, reflecting the spatial information of the sound source. Commonly used sound source identification technologies include beamforming, deconvolution methods, and subspace methods. However, these technologies are based on the assumption of the stationarity of the sound source, assuming that the sound source is fixed, and do not take into account factors such as the movement of the sound source and the influence of the external flow field on the sound transmission medium. Therefore, when identifying the moving sound source in the non-uniform flow field, wrong results will be obtained, and the location and intensity of the sound source cannot be correctly identified.
[0003] Rotational sound sources are a typical type of motion sound source. They rotate around a fixed center and are widely found in rotating machinery such as wind turbines, axial flow fans, gearboxes, etc. Therefore, the identification of rotational sound sources is of great significance, but the above-mentioned technology alone cannot effectively identify rotational sound sources. In order to solve this problem, the instantaneous signal segmentation processing technology was proposed. This technology divides the multi-channel signal to be processed into a series of sub-segments. It is assumed that the sound source in each sub-segment is approximately stationary, and then conventional sound source identification methods are used for sound imaging. The shortcoming of this technology is that it is difficult to determine the appropriate signal segmentation length under actual conditions, especially when identifying high-speed sound sources. The assumption that the sound source is approximately stationary in this method is difficult to meet. Therefore, the use of this technology to identify rotational sound sources has great limitations. The modal decomposition method decomposes the acoustic signal generated by the rotating sound source into a superposition of a series of modes, and performs frequency shift on the signal components corresponding to each mode according to the rotation speed of the sound source in the modal domain, in an attempt to eliminate the signal non-stationarity caused by the rotation of the sound source. Since the modal decomposition uses special functions to approximate the transmission function of the sound field, the approximate accuracy is affected by the number of function expansion terms. Therefore, the efficiency of numerical calculation using this method is low, and it is not practical enough for practical scenarios such as mechanical health monitoring and aeroacoustic testing. Another technology for rotating sound source identification is the virtual rotating array method, which converts the signal collected by the microphone array into a moving coordinate system that rotates synchronously with the sound source, and identifies it in a moving coordinate system where the sound source and the microphone array are relatively stationary, so as to overcome the adverse effects of the sound source movement on the sound source identification. In addition, the prior art takes into account the situation when the sound source is in a uniform flow field, and can correct the transmission function according to the sound field modeling and decomposition theory under uniform flow conditions to eliminate the influence of the uniform flow field in the medium on the sound propagation. However, uniform flow field is an idealized mathematical model. In reality, there is almost no situation that fully satisfies this idealized condition. The flow in most real flow fields is non-uniform. Therefore, it is of great practical significance to identify the sound source of rotational motion in non-uniform flow fields. On the other hand, the existing technology is not applicable to the more extensive non-uniform flow fields in reality, and the widely used deconvolution method has the disadvantage of low identification resolution and poor sound source identification accuracy under noise conditions. Therefore, the existing rotation sound source identification technology still has many defects. In the presence of non-uniform flow fields, it cannot fully meet the need for accurate identification of rotation sound sources.
[0004] Therefore, it is urgent to eliminate the influence of sound source movement on signal smoothness, improve the recognizability of rotational motion sound sources, improve the accuracy of sound source identification under actual conditions, realize high-resolution rotational motion sound source identification, and develop a method and system for identifying rotational motion sound sources in non-uniform flow fields that can obtain accurate results even under noisy conditions. Summary of the invention
[0005] In view of the above problems, the purpose of the present invention is to provide a method for identifying rotational sound sources in non-uniform flow fields to solve the following problems in the prior art: one method is difficult to determine the appropriate signal segment length under actual conditions, especially when identifying high-speed sound sources, the assumption that the sound source is approximately stationary is difficult to satisfy in this method, so the first method has great limitations in identifying rotational sound sources; another method uses special functions to approximate the transfer function of the sound field in modal decomposition, and the approximation accuracy is affected by the number of function expansion terms. The efficiency of numerical calculation using this method is low in the second method, and it is not practical enough for actual scenarios such as mechanical health monitoring and aerodynamic acoustic testing; the third method is that there is almost no situation in which uniform flow fields completely meet this idealized condition under actual conditions. The flow in most real flow fields is non-uniform, so it is not suitable for more extensive non-uniform flow fields in reality, and the widely used deconvolution method has the disadvantage of low identification resolution, and the sound source identification accuracy under noise conditions is poor.
[0006] The present invention provides a method for identifying a rotational motion sound source in a non-uniform flow field, comprising:
[0007] The microphone array is used to collect and store the sound source signals of the rotating motion in the non-uniform flow field;
[0008] Calculating angle compensation values of microphones in the microphone array according to the rotation speed of the sound source signal; obtaining signal interpolation weights according to the angle compensation values, performing linear interpolation on the sound signals collected by the microphones according to the signal interpolation weights to form interpolation signals, and converting the interpolation signals into a moving coordinate system that rotates synchronously with the sound source signal;
[0009] The interpolation signal in the moving coordinate system is converted into the frequency domain by using discrete Fourier transform to form a sound pressure vector, and an acoustic inverse problem of a rotating sound source under non-uniform flow conditions is constructed based on the sound pressure vector;
[0010] The acoustic inverse problem is solved based on a predetermined regularization parameter to obtain the position and intensity of the sound source signal, thereby realizing the identification of the rotational motion sound source under non-uniform flow conditions.
[0011] Preferably, the microphone array is a microphone array comprising M microphones; wherein,
[0012] The time domain signals p of the same length collected by each microphone in the microphone array m (t)(m=1,2,…,M) forms the sound pressure matrix P=[p1(t),p2(t),…,p M (t)] T ; m represents each microphone in the microphone array.
[0013] Preferably, the angle compensation value of the microphone is calculated according to the rotation speed of the sound source signal; and the process of obtaining the signal interpolation weight according to the angle compensation value includes:
[0014] Define the position directly above the microphone array as the initial position, and take the clockwise direction of the microphone array as the positive direction to calculate the angle of each microphone in the microphone array
[0015] The rotation angle of the sound source from the acquisition time t0 to the sampling time t1 is calculated according to the rotation angular velocity Λ of the sound source signal And calculate the angle compensation value θ of each microphone according to the angle and the rotation angle m =θ m0 +α; and, determining the microphone number required for signal interpolation according to the periodicity of the rotation angle:
[0016]
[0017]
[0018] in, It means the largest integer not greater than x. It means taking the smallest integer not less than x, and mod means taking the remainder;
[0019] The signal interpolation weight ρ2 of the signal collected by the microphone at each sampling moment is calculated based on the angle compensation value and the rotation angle:
[0020]
[0021] ρ2=1-ρ1.
[0022] Preferably, in the process of performing linear interpolation on the acoustic signal collected by the microphone based on the signal interpolation weight to form an interpolation signal, and converting the interpolation signal into a moving coordinate system that rotates synchronously with the sound source signal,
[0023] The interpolation signal in the moving coordinate system is: p mV (t) = ρ1p m2 (t)+ρ2p m1 (t);
[0024] Among them, p m1 (t) is the signal collected by microphone m1, p m2 (t) is the signal collected by microphone m2.
[0025] Preferably, the process of converting the interpolation signal in the moving coordinate system into the frequency domain by using discrete Fourier transform to form the sound pressure vector includes:
[0026] The interpolation signal is converted into a frequency domain signal using discrete Fourier transform:
[0027]
[0028] According to the preselected frequency f c The corresponding frequency components in the frequency domain signal are extracted to form a sound pressure vector:
[0029] Preferably, the process of constructing the inverse acoustic problem of the rotating sound source under non-uniform flow conditions based on the sound pressure vector includes:
[0030] Determine a sound source reconstruction area based on the microphone array and discretize it to form a reconstruction network;
[0031] Performing series expansion on the velocity potential at the position of the microphone and the position of the reconstruction grid to obtain a Fourier expansion;
[0032] Based on the Fourier expansion, the equivalent wave number and harmonic coefficients are calculated, and a transmission matrix is established to form an acoustic inverse problem of a rotating sound source under non-uniform flow conditions.
[0033] Preferably, the process of determining the sound source reconstruction area based on the microphone array and discretizing it to form a reconstruction network; and performing series expansion on the velocity potential at the position of the microphone and the position of the reconstruction grid to obtain the Fourier expansion includes:
[0034] A coordinate system is created based on the microphone array, and the center of the microphone array is used as the origin of the coordinate system to obtain the position coordinates r of each microphone. m =(X m cosθ m sinφ m ,X m cosθ m cosφ m ,X m sinθ m );
[0035] Determine the position coordinates r of the reconstruction network based on the position coordinates of each microphone n =(X' n cosθ' n sinφ' n ,X' n cosθ' n cosφ' n ,X' n sinθ'n )(n=1,2,…,N), and perform series expansion on the velocity potential at the microphone location and the reconstruction grid location to obtain the microphone Fourier expansion and the reconstruction network Fourier expansion respectively; wherein,
[0036] The microphone Fourier expansion is:
[0037]
[0038] Among them, the series coefficient Is an imaginary unit.
[0039] The Fourier expansion of the reconstructed network is:
[0040]
[0041] Among them, the series coefficient
[0042] Preferably, in the process of calculating the equivalent wave number and the harmonic coefficient based on the Fourier expansion, establishing the transmission matrix, and forming the acoustic inverse problem of the rotating sound source under the non-uniform flow condition, the following steps are included:
[0043] Based on the Fourier expansion, the acoustic inverse problem under non-uniform flow conditions is established in the pre-acquired rotating coordinate system:
[0044] p=Gq+k
[0045] Where q=[q1,q2,…,q N ] T is a sound source intensity vector, each element of which represents the sound source intensity of each reconstructed grid point in the reconstructed grid; k = [k1, k2, ..., k M ] T is the noise vector; the transmission matrix G of size M×N contains the following elements:
[0046]
[0047] The equivalent wave number is
[0048] The harmonic coefficient is
[0049] ξ n () is the nth order first kind spherical Bessel function;
[0050] is the n-order second-kind spherical Hankel function;
[0051] is the polynomial product term;
[0052] is the |α|, nth-order associated Legendre function.
[0053] Preferably, the process of solving the acoustic inverse problem based on a predetermined regularization parameter to obtain the position and intensity of the sound source signal and realizing the identification of the rotational motion sound source under non-uniform flow conditions includes:
[0054] Determine the regularization parameter λ>0, and rewrite the acoustic inverse problem by the 1-norm sparse regularization method to form a standardized inverse problem
[0055] The standardized inverse problem is solved using a CVX optimization toolbox based on a MATLAB platform to obtain the value of a sound source intensity vector q; wherein the non-zero element index in the value of the sound source intensity vector q represents the location of the sound source signal, and the value of the element represents the intensity of the sound source signal.
[0056] On the other hand, the present invention also provides a system for identifying a rotational motion sound source in a non-uniform flow field, which implements the method for identifying a rotational motion sound source in a non-uniform flow field as described above, comprising:
[0057] A signal acquisition unit, used to collect and store the sound source signal of the rotational motion in the non-uniform flow field using a microphone array;
[0058] A signal compensation unit, configured to calculate an angle compensation value of a microphone in the microphone array according to a rotation speed of the sound source signal; obtain a signal interpolation weight according to the angle compensation value, perform linear interpolation on the sound signal collected by the microphone according to the signal interpolation weight to form an interpolation signal, and convert the interpolation signal into a moving coordinate system that rotates synchronously with the sound source signal;
[0059] An inverse problem construction unit, used for converting the interpolation signal in the moving coordinate system into the frequency domain by discrete Fourier transform to form a sound pressure vector, and constructing an acoustic inverse problem of a rotating sound source under non-uniform flow conditions based on the sound pressure vector;
[0060] The inverse problem solving unit is used to solve the acoustic inverse problem based on a predetermined regularization parameter to obtain the position and intensity of the sound source signal, so as to realize the identification of the rotational motion sound source under non-uniform flow conditions.
[0061] It can be seen from the above technical scheme that the method for identifying the rotational sound source in the non-uniform flow field provided by the present invention first uses a microphone array to collect and store the sound source signal of the rotational motion in the non-uniform flow field, and then calculates the angle compensation value of the microphone according to the rotation speed of the sound source signal; obtains the signal interpolation weight according to the angle compensation value, linearly interpolates the sound signal collected by the microphone according to the signal interpolation weight to form an interpolation signal, and converts the interpolation signal to a moving coordinate system that rotates synchronously with the sound source signal, and then uses discrete Fourier transform to convert the interpolation signal in the moving coordinate system to the frequency domain to form a sound pressure vector, and constructs the acoustic inverse problem of the rotational sound source under non-uniform flow conditions based on the sound pressure vector, and then solves the acoustic inverse problem based on a predetermined regularization parameter to obtain the position and intensity of the sound source signal, thereby realizing the identification of the rotational sound source under non-uniform flow conditions. ; In this way, the signals collected by the microphones in the array are converted into synchronous motion by interpolation, and numerical calculations can be easily performed on discrete signals, eliminating the influence of the movement of the sound source on the signal stability, which is beneficial to the identification of the rotational motion sound source, and also creates conditions for the application of the existing sound source identification method under stable conditions; and the potential function of the non-uniform flow field is decomposed by special functions and series expansion, so as to decouple the sound source term and other terms in the acoustic inverse problem, solve the problem that the existing technology is not suitable for sound source identification under non-uniform flow field conditions, and improve the accuracy of sound source identification under actual conditions; at the same time, it can be used for observation signals in non-uniform flow fields that are widely present in real conditions, and the acoustic inverse problem is solved by a sparse regularization method, so as to achieve high-resolution rotational motion sound source identification, and obtain accurate results under noisy conditions. BRIEF DESCRIPTION OF THE DRAWINGS
[0062] By referring to the following description in conjunction with the accompanying drawings, and with a more comprehensive understanding of the present invention, other objects and results of the present invention will become more clear and easy to understand. In the accompanying drawings:
[0063] Figure 1 A flowchart of a method for identifying a rotational motion sound source in a non-uniform flow field according to an embodiment of the present invention;
[0064] Figure 2 A flowchart of a specific implementation of a method for identifying a rotational motion sound source in a non-uniform flow field according to an embodiment of the present invention;
[0065] Figure 3 A schematic diagram of an acoustic inverse problem model in a method for identifying a rotational sound source in a non-uniform flow field according to an embodiment of the present invention;
[0066] Figure 4 The numerical simulation conditions in the specific implementation of the method for identifying the rotational motion sound source in the non-uniform flow field according to the embodiment of the present invention;
[0067] Figure 5 The numerical simulation results in a specific implementation of the method for identifying a rotational motion sound source in a non-uniform flow field according to an embodiment of the present invention;
[0068] Figure 6 4 is a block diagram of a system for identifying a rotational sound source in a non-uniform flow field according to an embodiment of the present invention. DETAILED DESCRIPTION
[0069] At present, the existing technology has the following problems: one method is difficult to determine the appropriate signal segment length under actual conditions, especially when identifying high-speed moving sound sources, the assumption that the sound source is approximately stationary is difficult to meet in this method, so the first method has great limitations in identifying rotational sound sources; another method uses special functions to approximate the transfer function of the sound field in modal decomposition, and the approximation accuracy is affected by the number of function expansion terms. The efficiency of numerical calculation using this method is low in the second method, and it is not practical for actual scenarios such as mechanical health monitoring and aerodynamic acoustic testing; the third method is that there is almost no situation in which the uniform flow field completely meets this idealized condition under actual conditions. The flow of most real flow fields is non-uniform, so it is not suitable for more extensive non-uniform flow fields in reality, and the widely used deconvolution method has the disadvantage of low identification resolution, and the sound source identification accuracy under noisy conditions is poor.
[0070] In view of the above problems, the present invention provides a method for identifying a rotational 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.
[0071] In order to illustrate the method and system for identifying the rotational sound source in a non-uniform flow field provided by the present invention, Figure 1-Figure 5 The method for identifying the sound source of rotational motion in a non-uniform flow field according to an embodiment of the present invention is exemplarily indicated; Figure 6 The system for identifying a rotational sound source in a non-uniform flow field according to an embodiment of the present invention is exemplarily illustrated.
[0072] The following description of exemplary embodiments is merely illustrative in nature and is not intended to limit the present invention and its application or use. Technologies and devices known to ordinary technicians in the relevant field may not be discussed in detail, but in appropriate cases, the technologies and devices should be considered as part of the specification.
[0073] like Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 As shown in the figure, the present invention provides a method for identifying a rotational sound source in a non-uniform flow field, comprising:
[0074] S1: Use microphone array to collect and store the sound source signal of rotational motion in non-uniform flow field;
[0075] S2: Calculating angle compensation values of microphones in the microphone array according to the rotation speed of the sound source signal; obtaining signal interpolation weights according to the angle compensation values, performing linear interpolation on the sound signals collected by the microphones according to the signal interpolation weights to form interpolation signals, and converting the interpolation signals into a moving coordinate system that rotates synchronously with the sound source signal;
[0076] S3: using discrete Fourier transform to convert the interpolation signal in the moving coordinate system into the frequency domain to form a sound pressure vector, and constructing an acoustic inverse problem of a rotating sound source under non-uniform flow conditions based on the sound pressure vector;
[0077] S4: Solving the acoustic inverse problem based on a predetermined regularization parameter to obtain the position and intensity of the sound source signal, thereby realizing the identification of the rotational motion sound source under non-uniform flow conditions.
[0078] exist Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 In the embodiment shown together, step S1 is a process of collecting and storing the sound source signal of the rotational motion in the non-uniform flow field using a microphone array. In this embodiment, the microphone array includes M microphones, wherein the microphone array collects and stores the sound source signal of the rotational motion in the non-uniform flow field; wherein,
[0079] The time domain signals p of the same length collected by each microphone in the microphone array m (t)(m=1,2,…,M) forms the sound pressure matrix P=[p1(t),p2(t),…,p M (t)] T ; m represents each microphone in the microphone array.
[0080] More specifically, step S1 is to use a microphone array including M microphones to collect and store the sound source signal of the rotational motion in the non-uniform flow field, and the time domain signals p of the same length collected by each microphone are m (t)(m=1,2,…,M) forms the sound pressure matrix P=[p1(t),p2(t),…,p M (t)] T .
[0081] exist Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5In the embodiment shown together, step S2 is a process of calculating the angle compensation value of the microphone according to the rotation speed of the sound source signal; obtaining the signal interpolation weight according to the angle compensation value, performing linear interpolation on the sound signal collected by the microphone according to the signal interpolation weight to form an interpolation signal, and converting the interpolation signal into a moving coordinate system that rotates synchronously with the sound source signal; wherein,
[0082] The process of calculating the angle compensation value of the microphone according to the rotation speed of the sound source signal; and obtaining the signal interpolation weight according to the angle compensation value includes:
[0083] S21: define the position directly above the microphone array as the initial position, and take the clockwise direction of the microphone array as the positive direction to calculate the angle of each microphone in the microphone array
[0084] S22: Calculate the rotation angle of the sound source from the acquisition time t0 to the sampling time t1 according to the rotation angular velocity Λ of the sound source signal And calculate the angle compensation value θ of each microphone according to the angle and the rotation angle m =θ m0 +α; and, determining the microphone number required for signal interpolation according to the periodicity of the rotation angle:
[0085]
[0086]
[0087] in, It means the largest integer not greater than x. It means taking the smallest integer not less than x, and mod means taking the remainder;
[0088] S23: Calculate the signal interpolation weight of the microphone collected signal at each sampling moment based on the angle compensation value and the rotation angle:
[0089]
[0090] ρ2=1-ρ1.
[0091] In the process of performing linear interpolation on the acoustic signal collected by the microphone based on the signal interpolation weight to form an interpolation signal, and converting the interpolation signal into a moving coordinate system that rotates synchronously with the sound source signal,
[0092] The interpolation signal in the moving coordinate system is: p mV (t) = ρ1p m2 (t)+ρ2p m1 (t);
[0093] Among them, p m1 (t) is the signal collected by microphone m1, p m2 (t) is the signal collected by microphone m2.
[0094] Specifically, step S2 is a process of calculating microphone angle compensation values and signal interpolation weights according to the rotation speed of the sound source, performing linear interpolation on the collected sound signals, and converting the signals collected by each microphone into a moving coordinate system that rotates synchronously with the sound source.
[0095] More specifically, step S2 needs to perform the following steps:
[0096] Step (2-a): Define the position directly above the microphone array as the initial position, the clockwise direction as the positive direction, and calculate the angle of each microphone.
[0097] Step (2-b): At each signal sampling moment, calculate the rotation angle of the sound source from the start time t0 of the acquisition to the sampling time t1 according to the rotation angular velocity Λ of the sound source Calculate the angle compensation value θ for each microphone m =θ m0 +α. The number of microphones required for signal interpolation is determined based on the periodicity of the rotation angle:
[0098]
[0099]
[0100] in, It means the largest integer not greater than x. It means to find the smallest integer not less than x, and mod means to find the remainder.
[0101] Step (2-c): Calculate the interpolation weight ρ2 of the microphone signal at each sampling moment:
[0102]
[0103] ρ2=1-ρ1 (4)
[0104] Perform linear interpolation on the sound signal and convert the signal collected by each microphone into a moving coordinate system that rotates synchronously with the sound source:
[0105] p mV (t) = ρ1p m2 (t)+ρ2p m1 (t) (5)
[0106] Among them, p m1 (t) is the signal collected by microphone m1, pm2 (t) is the signal collected by microphone m2.
[0107] exist Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 In the embodiment shown together, step S3 is a process of converting the interpolation signal in the moving coordinate system into the frequency domain by using discrete Fourier transform to form a sound pressure vector, and constructing an acoustic inverse problem of a rotating sound source under non-uniform flow conditions based on the sound pressure vector; in this embodiment, the process of converting the interpolation signal in the moving coordinate system into the frequency domain by using discrete Fourier transform to form a sound pressure vector includes:
[0108] The interpolation signal is converted into a frequency domain signal using discrete Fourier transform:
[0109]
[0110] According to the preselected frequency f c The corresponding frequency components in the frequency domain signal are extracted to form a sound pressure vector:
[0111] In this embodiment, the process of constructing the inverse acoustic problem of the rotating sound source under non-uniform flow conditions based on the sound pressure vector includes:
[0112] Determine a sound source reconstruction area based on the microphone array and discretize it to form a reconstruction network;
[0113] Performing series expansion on the velocity potential at the position of the microphone and the position of the reconstruction grid to obtain a Fourier expansion;
[0114] Based on the Fourier expansion, the equivalent wave number and harmonic coefficients are calculated, and a transmission matrix is established to form an acoustic inverse problem of a rotating sound source under non-uniform flow conditions.
[0115] The process of determining a sound source reconstruction area based on the microphone array and discretizing it to form a reconstruction network; and performing series expansion on the velocity potential at the position of the microphone and the position of the reconstruction grid to obtain a Fourier expansion includes:
[0116] A coordinate system is created based on the microphone array, and the center of the microphone array is used as the origin of the coordinate system to obtain the position coordinates r of each microphone. m =(X m cosθ m sinφ m ,X m cosθ m cosφ m ,X m sinθm );
[0117] Determine the position coordinates r of the reconstruction network based on the position coordinates of each microphone n =(X' n cosθ' n sinφ' n ,X' n cosθ' n cosφ' n ,X' n sinθ' n )(n=1,2,…,N), and perform series expansion on the velocity potential at the microphone location and the reconstruction grid location to obtain the microphone Fourier expansion and the reconstruction network Fourier expansion respectively; wherein,
[0118] The microphone Fourier expansion is:
[0119]
[0120] Among them, the series coefficient Is an imaginary unit.
[0121] The Fourier expansion of the reconstructed network is:
[0122]
[0123] Among them, the series coefficient
[0124] In the process of calculating the equivalent wave number and harmonic coefficient based on the Fourier expansion, establishing the transmission matrix, and forming the acoustic inverse problem of the rotating sound source under the non-uniform flow condition, the following steps are included:
[0125] Based on the Fourier expansion, the acoustic inverse problem under non-uniform flow conditions is established in the moving coordinate system:
[0126] p=Gq+k
[0127] Where q=[q1,q2,…,q N ] T is a sound source intensity vector, each element of which represents the sound source intensity of each reconstructed grid point in the reconstructed grid; k = [k1, k2, ..., k M ] T is the noise vector; the transmission matrix G of size M×N contains the following elements:
[0128]
[0129] The equivalent wave number is
[0130] The harmonic coefficient is
[0131] ξ n () is the nth order first kind spherical Bessel function;
[0132] is the n-order second-kind spherical Hankel function;
[0133] is the polynomial product term;
[0134] is the |α|, nth-order associated Legendre function.
[0135] Specifically, step S3 first converts the interpolated signal into the frequency domain using discrete Fourier transform:
[0136]
[0137] According to the selected frequency f c , extract the corresponding frequency components in the frequency domain signal to form the sound pressure vector
[0138] Then, the sound source reconstruction area is determined and discretized, the velocity potential at the microphone location and the reconstruction grid location is expanded in series, the equivalent wave number and harmonic coefficient are calculated in the cylindrical coordinate system, and the transmission matrix is established to form the acoustic inverse problem of the rotating sound source under non-uniform flow conditions.
[0139] More specifically, step S3 performs the following steps:
[0140] Step (3-a): Determine the sound source reconstruction area and discretize it, divide the reconstruction area into N reconstruction grids, and establish Figure 2 The coordinate system shown in the figure has the origin of the coordinate system located at the center of the microphone, and the position coordinates of each microphone are r m =(X m cosθ m sinφ m ,X m cosθ m cosφ m ,X m sinθ m ), the position coordinates of the reconstructed grid are r' n =(X' n cosθ' n sinφ' n ,X' n cosθ' n cosφ' n ,X' n sinθ' n)(n=1,2,…,N). Perform Fourier series expansion on the velocity potential at the microphone location:
[0141]
[0142] Among them, the series coefficient Is an imaginary unit.
[0143] Perform Fourier series expansion on the velocity potential at the location of the reconstruction grid:
[0144]
[0145] Among them, the series coefficient
[0146] Step (3-b): Let the flow field Mach number be Ma, c0 = 1, c β =–j2πf c MaФ β ' / c(β≠0),d0=1,d β =j2πf c MaФ –β / c(β≠0), the inverse acoustic problem under non-uniform flow conditions is established in the rotating coordinate system:
[0147] p=Gq+k (9)
[0148] Where c is the speed of sound, q = [q1, q2, …, q N ] T is the sound source intensity vector, each element represents the sound source intensity of the reconstructed grid point, k = [k1, k2, …, k M ] T is the noise vector, and the transmission matrix G of size M×N contains the elements:
[0149]
[0150] Among them, the equivalent wave number Harmonic coefficient ξ n () is the nth order first kind spherical Bessel function, is a spherical Hankel function of the second kind of order n, and the polynomial product term is the |α|, nth-order associated Legendre function.
[0151] exist Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5In the embodiment shown together, step S4 is a process of solving the acoustic inverse problem based on a predetermined regularization parameter to obtain the position and intensity of the sound source signal, thereby realizing the identification of the rotational motion sound source under non-uniform flow conditions; wherein,
[0152] The process of solving the acoustic inverse problem based on a predetermined regularization parameter to obtain the position and intensity of the sound source signal and realizing the identification of the rotational motion sound source under non-uniform flow conditions includes:
[0153] Determine the regularization parameter λ>0, and rewrite the acoustic inverse problem by the 1-norm sparse regularization method to form a standardized inverse problem
[0154] The standardized inverse problem is solved using a CVX optimization toolbox based on a MATLAB platform to obtain the value of a sound source intensity vector q; wherein the non-zero element index in the value of the sound source intensity vector q represents the location of the sound source signal, and the value of the element represents the intensity of the sound source signal.
[0155] Specifically, step S4 first determines that the regularization parameter λ>0, and uses the 1-norm sparse regularization method to rewrite the acoustic inverse problem into the following form:
[0156]
[0157] The above problem is solved by using the CVX optimization toolbox based on the MATLAB platform to obtain the sound source intensity vector q. The non-zero element index in the vector q represents the location of the sound source, and the value of the element represents the intensity of the sound source, thereby realizing the identification of the rotational motion sound source under non-uniform flow conditions.
[0158] The numerical simulation conditions are as follows Figure 4 As shown, the three frequencies f c =4000Hz point sound source rotates counterclockwise around O'(0m,0m,0.75m) at an angular velocity of Λ=80πrad / s in the x'-y' plane (viewed from the positive direction of the z axis). The rotation radius X s =0.2m, the sound source intensity is 80dB, the microphone array parallel to the sound source rotation plane contains M = 52 uniformly distributed microphones, the array center is located at the coordinate origin O (0m, 0m, 0m), the array radius X m=0.6m, an infinitely long rigid cylinder with a radius of a = 0.2m is placed between the sound source and the array, and the center of the cross section in the yz plane is located at O" (0m, 0m, 0.4m). The uniform potential flow in the xz plane flows along the angle α = 20°. Due to the interference of the cylinder on the airflow, a non-uniform flow field is formed near the cylinder. The reconstruction area is selected as a rectangular area distributed in the plane where the sound source is located, with a range of -0.5m≤x≤0.5m and -0.5m≤y≤0.5m, and is divided into N = 1681 uniformly distributed reconstruction grids. Gaussian noise is added to the signal collected by the microphone to make the signal-to-noise ratio Where n is the noise, and the regularization parameter λ is set to 0.5 based on experience.
[0159] In the method for identifying the rotational sound source in a non-uniform flow field according to the embodiment of the present invention, the sound source identification result is as follows: Figure 5 As shown, the rectangular box represents the real position of the sound source. It can be seen that the three sound sources can be identified at the real position, the sound source positioning error is very small, and the intensity estimation value of the three sound sources 78.1dB is close to the real intensity, with a relative error of 1.9%. The above results show that the technical solution in the method for identifying the rotational motion sound source in the non-uniform flow field in the embodiment of the present invention can achieve accurate positioning and sound source intensity estimation of the rotational motion sound source in the non-uniform flow field, and can also obtain very accurate identification results under the condition of noise. Compared with the prior art, the method for identifying the rotational motion sound source in the non-uniform flow field in the embodiment of the present invention has made great progress in the accuracy of identifying the rotational motion sound source in the non-uniform flow field.
[0160] In summary, the method for identifying the rotational sound source in a non-uniform flow field provided by the present invention first uses a microphone array to collect and store the sound source signal of the rotational motion in the non-uniform flow field, and then calculates the angle compensation value of the microphone according to the rotation speed of the sound source signal; obtains the signal interpolation weight according to the angle compensation value, linearly interpolates the sound signal collected by the microphone according to the signal interpolation weight to form an interpolation signal, and converts the interpolation signal to a moving coordinate system that rotates synchronously with the sound source signal, and then uses discrete Fourier transform to convert the interpolation signal in the moving coordinate system to the frequency domain to form a sound pressure vector, and constructs an acoustic inverse problem of the rotational sound source under non-uniform flow conditions based on the sound pressure vector, and then solves the acoustic inverse problem based on a predetermined regularization parameter to obtain the position and intensity of the sound source signal, thereby realizing the identification of the rotational sound source under non-uniform flow conditions; in this way , the signals collected by the microphones in the array are converted into synchronous motion by interpolation, and numerical calculations can be easily performed on discrete signals, eliminating the influence of the movement of the sound source on the signal stability, which is beneficial to the identification of the rotational motion sound source, and also creates conditions for the application of the existing sound source identification method under stable conditions; and the potential function of the non-uniform flow field is decomposed by special functions and series expansion, so as to decouple the sound source term and other terms in the acoustic inverse problem, solve the problem that the existing technology is not suitable for sound source identification under non-uniform flow field conditions, and improve the accuracy of sound source identification under actual conditions; at the same time, it can be used for observation signals in non-uniform flow fields that are widely present in real conditions, and the acoustic inverse problem is solved by the sparse regularization method, so as to achieve high-resolution rotational motion sound source identification, and obtain accurate results under noisy conditions.
[0161] like Figure 6 As shown, the present invention also provides a system 100 for identifying a rotational motion sound source in a non-uniform flow field, which implements the method for identifying a rotational motion sound source in a non-uniform flow field as described above, comprising:
[0162] The signal collection unit 101 is used to collect and store the sound source signal of the rotational motion in the non-uniform flow field using a microphone array;
[0163] The signal compensation unit 102 is used to calculate the angle compensation value of the microphone in the microphone array according to the rotation speed of the sound source signal; obtain the signal interpolation weight according to the angle compensation value, perform linear interpolation on the sound signal collected by the microphone according to the signal interpolation weight to form an interpolation signal, and convert the interpolation signal into a moving coordinate system that rotates synchronously with the sound source signal;
[0164] An inverse problem construction unit 103 is used to convert the interpolation signal in the moving coordinate system into the frequency domain by discrete Fourier transform to form a sound pressure vector, and construct an acoustic inverse problem of a rotating sound source under non-uniform flow conditions based on the sound pressure vector;
[0165] The inverse problem solving unit 104 is used to solve the acoustic inverse problem based on a predetermined regularization parameter to obtain the position and intensity of the sound source signal, so as to identify the rotational motion sound source under non-uniform flow conditions.
[0166] The specific implementation method is not specifically limited and will not be described in detail here. Please refer to the specific embodiment of the method for identifying the rotational motion sound source in a non-uniform flow field.
[0167] As described above, the system for identifying the rotational sound source in the non-uniform flow field provided by the present invention first collects and stores the sound source signal of the rotational motion in the non-uniform flow field by using a microphone array through the signal acquisition unit 101, and then calculates the angle compensation value of the microphone according to the rotation speed of the sound source signal through the signal compensation unit 102; obtains the signal interpolation weight according to the angle compensation value, linearly interpolates the sound signal collected by the microphone according to the signal interpolation weight to form an interpolation signal, and converts the interpolation signal to a moving coordinate system that rotates synchronously with the sound source signal, and then converts the interpolation signal in the moving coordinate system to the frequency domain by using a discrete Fourier transform through the inverse problem construction unit 103 to form a sound pressure vector, and constructs an acoustic inverse problem of the rotating sound source under the non-uniform flow condition based on the sound pressure vector, and then solves the acoustic inverse problem based on a predetermined regularization parameter through the inverse problem solving unit 104 to obtain the position and intensity of the sound source signal. The invention realizes the identification of the rotational motion sound source under the condition of non-uniform flow. Thus, the signals collected by the microphones in the array are converted into synchronous motion by interpolation. The discrete signals can be conveniently calculated numerically, and the influence of the motion of the sound source on the signal stability is eliminated, which is beneficial to the identification of the rotational motion sound source and creates conditions for the application of the existing sound source identification method under stable conditions. The potential function of the non-uniform flow field is decomposed by special functions and series expansion, so as to decouple the sound source term and other terms in the acoustic inverse problem, solve the problem that the existing technology is not suitable for the sound source identification under the condition of non-uniform flow field, and improve the accuracy of the sound source identification under the actual conditions. At the same time, the invention can be used for the observation signals in the non-uniform flow field which are widely present in the actual conditions, and the acoustic inverse problem is solved by the sparse regularization method, so as to realize the high-resolution rotational motion sound source identification and obtain accurate results under the condition of noise.
[0168] As described above with reference to the accompanying drawings, the system and method for identifying the sound source of rotational motion in a non-uniform flow field proposed by the present invention are described by way of example. However, those skilled in the art should understand that various improvements can be made to the system and method for identifying the sound source of rotational motion in a non-uniform flow field proposed by the present invention without departing from the content of the present invention. Therefore, the protection scope of the present invention should be determined by the content of the attached claims.
Claims
1. A method for identifying a rotational sound source in a non-uniform flow field, characterized in that: include: The microphone array is used to collect and store the sound source signals of the rotating motion in the non-uniform flow field; Calculating angle compensation values of microphones in the microphone array according to the rotation speed of the sound source signal; obtaining signal interpolation weights according to the angle compensation values, performing linear interpolation on the sound source signal collected by the microphone according to the signal interpolation weights to form an interpolation signal, and converting the interpolation signal into a moving coordinate system that rotates synchronously with the sound source signal; The interpolation signal in the moving coordinate system is converted into the frequency domain by discrete Fourier transform to form a sound pressure vector, and the acoustic inverse problem of the rotating sound source under non-uniform flow conditions is constructed based on the sound pressure vector; wherein, the process of constructing the acoustic inverse problem of the rotating sound source under non-uniform flow conditions based on the sound pressure vector includes: determining the sound source reconstruction area based on the microphone array and discretizing it to form a reconstruction grid; performing series expansion on the velocity potential at the position of the microphone and the position of the reconstruction grid to obtain the Fourier expansion; calculating the equivalent wave number and harmonic coefficient based on the Fourier expansion, establishing a transmission matrix, and forming the acoustic inverse problem of the rotating sound source under non-uniform flow conditions; wherein, The process of determining a sound source reconstruction area based on the microphone array and discretizing it to form a reconstruction grid; and performing series expansion on the velocity potential at the position of the microphone and the position of the reconstruction grid to obtain a Fourier expansion includes: A coordinate system is created based on the microphone array, and the center of the microphone array is used as the origin of the coordinate system to obtain the position coordinates r of each microphone. m ; Determine the position coordinates r' of the reconstruction grid based on the position coordinates of each microphone n , and the velocity potential at the microphone location and the reconstruction grid location is expanded in series to obtain the microphone Fourier expansion and the reconstruction grid Fourier expansion respectively; where, The microphone Fourier expansion is: Among them, the series coefficient is an imaginary unit; is the azimuth of the microphone in the spherical coordinate system; The Fourier expansion of the reconstructed grid is: Among them, the series coefficient To reconstruct the orientation of the network in the spherical coordinate system; In the process of calculating the equivalent wave number and harmonic coefficient based on the Fourier expansion, establishing the transmission matrix, and forming the acoustic inverse problem of the rotating sound source under the non-uniform flow condition, the following steps are included: Based on the Fourier expansion, the acoustic inverse problem under non-uniform flow conditions is established in the moving coordinate system: p=Gq'+k Where q'=[q1,q2,…,q N ] T is a sound source intensity vector, each element of which represents the sound source intensity of each reconstructed grid point in the reconstructed grid; k = [k1, k2, ..., k M ] T is the noise vector; the transmission matrix G of size M×N contains elements g m,n ; The acoustic inverse problem is solved based on a predetermined regularization parameter to obtain the position and intensity of the sound source signal, thereby realizing the identification of the rotational motion sound source under non-uniform flow conditions.
2. The method for identifying a rotational sound source in a non-uniform flow field according to claim 1, characterized in that: The microphone array is a microphone array comprising M microphones; wherein, The time domain signals p of the same length collected by each microphone in the microphone array m (t)(m=1,2,…,M) forms the sound pressure matrix P=[p1(t),p2(t),…,p M (t)] T ; m represents each microphone in the microphone array.
3. The method for identifying a rotational sound source in a non-uniform flow field according to claim 2, characterized in that: The process of calculating the angle compensation value of the microphone according to the rotation speed of the sound source signal; and obtaining the signal interpolation weight according to the angle compensation value includes: Define the position directly above the microphone array as the initial position, and take the clockwise direction of the microphone array as the positive direction to calculate the angle of each microphone in the microphone array The rotation angle of the sound source from the acquisition time t0 to the sampling time t1 is calculated according to the rotation angular velocity Λ of the sound source signal And calculate the angle compensation value θ of each microphone according to the angle and the rotation angle m =θ m0 +α'; and, determining the microphone number required for signal interpolation according to the periodicity of the rotation angle: in, It means the largest integer not greater than x. It means taking the smallest integer not less than x, and mod means taking the remainder; The signal interpolation weights ρ1 and ρ2 of the signal collected by the microphone at each sampling moment are calculated based on the angle compensation value and the rotation angle: ρ2=1-ρ1.
4. The method for identifying a rotational sound source in a non-uniform flow field according to claim 3, characterized in that: In the process of performing linear interpolation on the sound source signal collected by the microphone based on the signal interpolation weight to form an interpolation signal, and converting the interpolation signal into a moving coordinate system that rotates synchronously with the sound source signal, The interpolation signal in the moving coordinate system is: p mV (t) = ρ1p m2 (t)+ρ2p m1 (t); Among them, p m1 (t) is the signal collected by microphone m1, p m2 (t) is the signal collected by microphone m2.
5. The method for identifying a rotational sound source in a non-uniform flow field according to claim 4, characterized in that: The process of converting the interpolation signal in the moving coordinate system into the frequency domain by using discrete Fourier transform to form a sound pressure vector includes: The interpolation signal is converted into a frequency domain signal using discrete Fourier transform: According to the preselected frequency f c The corresponding frequency components in the frequency domain signal are extracted to form a sound pressure vector:
6. The method for identifying a rotational sound source in a non-uniform flow field according to claim 5, characterized in that: The process of solving the acoustic inverse problem based on a predetermined regularization parameter to obtain the position and intensity of the sound source signal and realizing the identification of the rotational motion sound source under non-uniform flow conditions includes: Determine the regularization parameter λ>0, and rewrite the acoustic inverse problem by the 1-norm sparse regularization method to form a standardized inverse problem The standardized inverse problem is solved using the CVX optimization toolbox based on the MATLAB platform to obtain the value of the sound source intensity vector q'; wherein the non-zero element index in the value of the sound source intensity vector q' represents the location of the sound source signal, and the value of the element represents the intensity of the sound source signal.
7. A system for identifying a rotating sound source in a non-uniform flow field, characterized in that: Implementing the method for identifying a rotational sound source in a non-uniform flow field as described in any one of claims 1 to 6, comprising: A signal acquisition unit, used to collect and store the sound source signal of the rotational motion in the non-uniform flow field using a microphone array; A signal compensation unit, configured to calculate an angle compensation value of a microphone in the microphone array according to a rotation speed of the sound source signal; obtain a signal interpolation weight according to the angle compensation value, perform linear interpolation on the sound source signal collected by the microphone according to the signal interpolation weight to form an interpolation signal, and convert the interpolation signal into a moving coordinate system that rotates synchronously with the sound source signal; An inverse problem construction unit is used to convert the interpolation signal in the moving coordinate system into the frequency domain by discrete Fourier transform to form a sound pressure vector, and construct an acoustic inverse problem of a rotating sound source under non-uniform flow conditions based on the sound pressure vector; wherein, the process of constructing an acoustic inverse problem of a rotating sound source under non-uniform flow conditions based on the sound pressure vector includes: determining a sound source reconstruction area based on the microphone array and discretizing it to form a reconstruction grid; performing series expansion on the velocity potential at the position of the microphone and the position of the reconstruction grid to obtain a Fourier expansion; calculating the equivalent wave number and harmonic coefficient based on the Fourier expansion, establishing a transmission matrix, and forming an acoustic inverse problem of a rotating sound source under non-uniform flow conditions; wherein, The process of determining a sound source reconstruction area based on the microphone array and discretizing it to form a reconstruction grid; and performing series expansion on the velocity potential at the position of the microphone and the position of the reconstruction grid to obtain a Fourier expansion includes: A coordinate system is created based on the microphone array, and the center of the microphone array is used as the origin of the coordinate system to obtain the position coordinates r of each microphone. m ; Determine the position coordinates r' of the reconstruction grid based on the position coordinates of each microphone n , and the velocity potential at the microphone location and the reconstruction grid location is expanded in series to obtain the microphone Fourier expansion and the reconstruction grid Fourier expansion respectively; where, The microphone Fourier expansion is: Among them, the series coefficient is an imaginary unit; is the azimuth of the microphone in the spherical coordinate system; The Fourier expansion of the reconstructed grid is: Among them, the series coefficient To reconstruct the orientation of the network in the spherical coordinate system; In the process of calculating the equivalent wave number and harmonic coefficient based on the Fourier expansion, establishing the transmission matrix, and forming the acoustic inverse problem of the rotating sound source under the non-uniform flow condition, the following steps are included: Based on the Fourier expansion, the acoustic inverse problem under non-uniform flow conditions is established in the moving coordinate system: p=Gq'+k Where q'=[q1,q2,…,q N ] T is a sound source intensity vector, each element of which represents the sound source intensity of each reconstructed grid point in the reconstructed grid; k = [k1, k2, ..., k M ] T is the noise vector; the transmission matrix G of size M×N contains elements g m,n ; The inverse problem solving unit is used to solve the acoustic inverse problem based on a predetermined regularization parameter to obtain the position and intensity of the sound source signal, so as to realize the identification of the rotational motion sound source under non-uniform flow conditions.
