Interference modeling and multi-source cooperative suppression method and system for hybrid dynamic sound source

Abstract

The invention relates to an interference modeling and multi-source cooperative suppression method and system for a hybrid dynamic sound source. The method comprises the following steps: deploying a distributed microphone array and collecting noise signals of a low-altitude aircraft; processing the noise signal by using the three-dimensional dynamic sound field model, and outputting acoustic information; performing time-frequency analysis on the noise signal by using the acoustic information to generate frequency and time information; performing spatial distribution modeling on the noise signals to obtain spatial distribution characteristics of a three-dimensional dynamic sound field; performing Doppler correction on the noise signal by using the frequency and time information so as to compensate the frequency deviation of the noise signal in real time, and generating a Doppler correction signal; constructing a dynamic correlation model, separating the input Doppler correction signal, and outputting a separation result; and according to the spatial distribution characteristics and the separation result of the three-dimensional dynamic sound field, generating a corresponding collaborative suppression strategy. Compared with the prior art, the system has the advantages of being high in integration level and high in noise analysis precision.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of noise control and signal processing, in particular to a mixed dynamic sound source interference modeling and multi-source collaborative suppression method. BACKGROUND

[0002] With the rapid development of low-altitude economy, the noise pollution problem of low-altitude aircraft such as unmanned aerial vehicles and electric vertical take-off and landing aircraft has become increasingly prominent, and its influence on urban environment, residents' life and specific application scenarios (such as emergency rescue and precision instrument operation) has become a key factor restricting the development of the industry. The core of the present technical field is to realize the suppression and elimination of noise interference through accurate analysis of the characteristics of the sound source and effective control of the propagation path, involving the cross-fusion of multiple technical directions such as acoustic measurement, signal processing and active control.

[0003] Traditional noise control technology mainly adopts a single processing method based on the assumption of a static sound field. In terms of sound field modeling, it usually relies on single-point or small-array measurement at fixed positions, and uses Fourier transform and other steady-state analysis methods, which lack the ability to capture the non-stationary characteristics of dynamic sound sources. In terms of noise separation, it mostly uses fixed-parameter blind source separation algorithms or simple beamforming techniques, which are difficult to handle the time-varying mixing problem caused by moving sound sources. In terms of suppression means, passive control (such as sound barriers) or simple active noise control strategies are often used separately, and the control parameters are mostly set statically, which cannot adapt to the dynamic changes of sound source position and radiation characteristics.

[0004] The invention with publication number CN120428168B discloses a three-dimensional sound event precise identification and positioning method based on AI and multi-microphone array, which includes: acquiring multi-channel sound signals; performing frame windowing processing on the multi-channel sound signals to obtain time-frequency domain information; generating mel feature information; performing time-space joint feature extraction to generate initial event category and three-dimensional direction coding; generating multi-angle virtual sound field data by combining the attenuation relationship of sound source direction and microphone angle through simulation algorithm; inputting the real-time collected multi-channel sound signals into the optimized three-dimensional sound detection model for reasoning, and outputting the optimized event category and three-dimensional direction coding; performing dynamic calibration to generate a real-time calibrated three-dimensional sound detection model to solve the problems of insufficient multi-dimensional distribution characteristic capture ability of three-dimensional sound field, limited model generalization ability caused by single data training and static environment assumption, and insufficient real-time adaptability under dynamic noise interference and long-term performance decay. However, this scheme does not involve further analysis of multi-source noise under the motion state of the aircraft, and the analysis accuracy of the noise is low.

[0005] In summary, in the prior art, the sound field modeling method based on the steady-state assumption cannot accurately describe the Doppler effect and the space-time variation of the sound field caused by the movement of the aircraft, resulting in insufficient model accuracy; the separation algorithm with fixed parameters cannot cope with the time-varying mixing characteristics of the moving sound source signal, causing incomplete sound source separation and trajectory tracking lag; the isolated application and parameter fixation of passive and active control measures result in limited suppression effect and poor adaptability when facing dynamically changing multi-source noise. These defects ultimately lead to a series of problems such as modeling distortion, inaccurate separation, and unsatisfactory suppression effect in traditional technology when dealing with dynamic mixed sound sources such as low-altitude aircraft. SUMMARY

[0006] The purpose of the present application is to overcome the defects of the prior art and provide a mixed dynamic sound source interference modeling and multi-source collaborative suppression method to solve the key problems of insufficient modeling accuracy of dynamic multi-source sound field of low-altitude aircraft, incomplete noise component separation, and limited control effect of traditional suppression methods due to the difficulty in adapting to the dynamic changes of sound sources in related technology.

[0007] The purpose of the present application can be achieved by the following technical solutions: The present application provides a mixed dynamic sound source interference modeling and multi-source collaborative suppression method, comprising: Deploying a distributed microphone array and collecting low-altitude aircraft noise signals; processing the noise signals using a pre-constructed three-dimensional dynamic sound field model and outputting acoustic information; Constructing a dynamic compensation sound field model to perform time-frequency analysis, spatial distribution modeling, and Doppler correction on the noise signals: using the acoustic information, performing time-frequency analysis on the noise signals to generate frequency and time information; performing spatial distribution modeling on the noise signals to obtain the spatial distribution characteristics of the three-dimensional dynamic sound field; using the frequency and time information to perform Doppler correction on the noise signals to compensate for the frequency shift of the noise signals in real time and generate a Doppler correction signal; Constructing a dynamic correlation model to separate the input Doppler correction signal and output the separation result; generating a corresponding collaborative suppression strategy according to the spatial distribution characteristics of the three-dimensional dynamic sound field and the separation result.

[0008] Further, the three-dimensional dynamic sound field model specifically comprises: wherein, is the sound speed distribution of the environment medium, is the sound source excitation term, is the low-altitude aircraft noise signal, is the Laplacian operator of the sound pressure at the spatial position r and time t, used to describe the distribution variation rate of the sound pressure in three-dimensional space.

[0009] Further, the time-frequency analysis specifically includes: The noise signal is decomposed in time-frequency domain by using adaptive S transform, and the calculation formula is: Wherein, is a time-frequency spectrum, is a microphone collected signal, is a frequency, is a time, is a local time window, is the spatial position of the mth microphone, is an imaginary unit, is an adaptive Gaussian window function, is a Fourier basis function; integral is to weight and superimpose the signal segments on the whole time axis, to obtain the time-frequency spectrum value under the current time t, frequency ; is the instantaneous frequency at time t, is an adjustment coefficient.

[0010] Further, the spatial distribution modeling specifically uses a spherical harmonic function decomposition model to perform high-order approximation on the spatial propagation characteristics of the sound field, to represent the directivity and distance decay law of sound pressure in three-dimensional space, and the spherical harmonic function decomposition model is: Wherein, is a sound pressure, is a dynamic coefficient, represents a spherical harmonic function base, is a wave number, is a spatial position, is a spatial position, is a propagation phase and attenuation term of sound wave from the sound source position to the spatial position.

[0011] Further, the Doppler correction specifically includes: for the Doppler effect caused by high-speed motion of the aircraft, a Doppler correction function is defined to compensate for the real-time frequency offset of the signal, and the calculation formula is: Wherein, is a Doppler correction function, is a radial velocity of the aircraft relative to the microphone, is the spatial position of the mth microphone The time spectrum after Doppler correction The velocity of sound in the ambient medium at the real-time spatial location of the sound source.

[0012] Furthermore, the process of constructing a dynamic association model includes: The time delay information of the noise signal is extracted using a generalized cross-correlation function; the multi-source signal matrix is ​​decomposed into independent source vectors using a time-varying independent vector analysis algorithm combined with a time-maximizing correlation function; and a state vector is established for each separated independent source vector using an interactive multi-model Kalman filter; and the state vector is associated with the aircraft trajectory parameters and spatial position to form the dynamic association model.

[0013] Furthermore, the generalized cross-correlation function is: in, This is the time-frequency signal after Doppler correction. For time delay, Indicates conjugate. For microphone Complex conjugate of the spectrum during signal acquisition. The phase compensation factor is the time delay. Let i be the spatial location of the i-th microphone. yes The complex conjugate; Let j be the spatial location of the j-th microphone; The time-varying independent vector analysis algorithm includes: in, It is a multi-source signal matrix. For independent source vectors, For independent source vectors, The separated sound source; The function for maximizing time correlation is: in, For the unmixing matrix, As a time delay constraint weight, This represents the expectation operation; For the separated first The time-frequency spectrum of an independent source signal: n is the index of the independent sound source. =1,2,...,N, where N is the number of sound sources; The state vector is represented as follows: in, the nth independent source signal after separation, is a time derivative thereof.

[0014] Further, the synergistic suppression strategy specifically comprises: According to the spatial distribution characteristics and separation results of the three-dimensional dynamic sound field, the synergistic action area and division logic of the active noise control and the passive noise control are determined, and a linkage triggering mechanism of the active noise control and the passive noise control is established, the gain coefficient of the active noise control and the opening and adjusting state of the passive noise control are dynamically adjusted according to the noise time-frequency signal intensity after Doppler correction, and adaptive synergy of the active noise control and the passive noise control is realized. The active noise control comprises generating secondary sound waves opposite in phase to the noise source to realize active noise reduction, and the passive noise control comprises optimizing the geometric layout and sound-absorbing material distribution of the sound barrier to realize passive noise reduction.

[0015] Further, the method further comprises: The finite difference time domain simulation method is used to verify the noise suppression effect of the synergistic suppression strategy, and the synergistic suppression strategy is continuously optimized through numerical simulation. An experimental platform comprising multiple low-altitude aircraft, a microphone array, an active noise control device and a passive noise control device is constructed to simulate a dynamic noise environment in a real scene, and the performance of the dynamic compensation sound field model, the separation effect of the Doppler modified signal and the noise suppression effect of the synergistic suppression strategy are verified through experiments.

[0016] The application also provides a mixed dynamic sound source interference modeling and multi-source synergistic suppression system, comprising a memory and a processor, the memory stores a computer program, and the processor calls the computer program to execute the steps of the method as described above.

[0017] Compared with the prior art, the application has the following advantages: (1) The application designs a joint optimization framework of generalized cross-correlation, time-varying independent vector analysis and interactive multiple model Kalman filtering: first, the time delay information of the multi-channel signal is extracted through generalized cross-correlation, then the blind source separation of the multi-source signal is realized in the time-frequency domain by using time-varying independent vector analysis, and finally the dynamic characteristics of each sound source are tracked by using interactive multiple model Kalman filtering, and a dynamic correlation model of sound source-environment-propagation is established. The accurate separation of multi-source noise in a multi-source dynamic sound field is realized, different sound source components in mixed noise can be effectively separated, the source and distribution of noise can be accurately tracked and analyzed, and a corresponding cooperative suppression strategy can be designed; the high-precision analysis of low-altitude aircraft noise is realized, the problem that the traditional method is difficult to process dynamic multi-source sound field is solved, accurate acoustic basis is provided for subsequent noise suppression, and the accuracy and reliability of noise analysis are significantly improved.

[0018] (2) The present application designs active and passive cooperative suppression strategies, designs active noise control based on three-dimensional sound field information and multi-source separation results, generates secondary sound waves opposite in phase to noise sources, combines the optimization of sound barrier geometric layout and sound absorption material distribution, forms an active and passive combined suppression system, realizes multi-dimensional control of dynamic noise, breaks through the limitations of single suppression method, and can effectively control different frequency band noises in complex environments, significantly improving the comprehensive effect of noise suppression.

[0019] (3) The present application can effectively capture the instantaneous frequency characteristics of noise signals in high-speed motion state by using adaptive S transform for time-frequency analysis and dynamically adjusting the time window width of non-stationary noise signals; at the same time, the spatial modeling of three-dimensional sound field is combined with spherical harmonic function, which accurately represents the propagation law of sound pressure with space angle and distance; and by introducing Doppler real-time compensation mechanism, the frequency offset caused by the movement of the aircraft is eliminated, which significantly improves the fidelity of dynamic sound field reconstruction.

[0020] (4) The present application builds a sound source-environment-propagation dynamic correlation model, so that the system can accurately separate and track different sound source components such as engine noise and aerodynamic noise in complex environment, provides high-precision, real-time acoustic data support for subsequent cooperative noise suppression, and effectively solves the technical bottleneck of insufficient dynamic multi-source sound field analysis capability of traditional methods. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 The flow chart of the interference modeling and multi-source cooperative suppression method of a mixed dynamic sound source provided in the embodiment of the present application; Figure 2 The overall step flow chart of the interference modeling and multi-source cooperative suppression method of a mixed dynamic sound source provided in the embodiment of the present application; Figure 3 The adaptive S transform time-frequency analysis flow chart of the interference modeling and multi-source cooperative suppression method of a mixed dynamic sound source provided in the embodiment of the present application; Figure 4 The spherical harmonic function sound field spatial modeling schematic diagram of the interference modeling and multi-source cooperative suppression method of a mixed dynamic sound source provided in the embodiment of the present application; Figure 5 A multi-source noise separation and joint optimization framework diagram of a mixed dynamic sound source interference modeling and multi-source collaborative suppression method provided in an embodiment of the present application; Figure 6 A timing diagram of multi-algorithm collaborative work in a joint optimization framework of a mixed dynamic sound source interference modeling and multi-source collaborative suppression method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings can be arranged and designed in various different configurations.

[0023] Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative labor are within the scope of protection of the present application.

[0024] It should be noted that: similar reference numbers and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0025] Embodiment 1 As shown in Figure 1 and Figure 2 , the present embodiment provides a mixed dynamic sound source interference modeling and multi-source collaborative suppression method, which comprises the steps of: S1: deploying a distributed microphone array and collecting low-altitude aircraft noise signals; processing the noise signals using a pre-constructed three-dimensional dynamic sound field model, and outputting acoustic information; Specifically, Dynamic sound field signal acquisition and preliminary modeling: real-time acquisition of low-altitude aircraft noise signals and construction of a three-dimensional dynamic sound field model to provide basic acoustic data for subsequent noise analysis and suppression. Deploy multiple groups of distributed microphone arrays in the monitoring area, each group containing microphones, with array positions being , to achieve three-dimensional coverage of aircraft noise propagation; through beamforming technology, the spatial directivity of the collected signals is enhanced, and the noise source position is estimated using sound source positioning technology Real-time coordinates. Based on the multi-physical field sound propagation equation: A three-dimensional dynamic sound field model of low-altitude aircraft noise is established, wherein the sound velocity distribution of the environment medium, the sound source excitation term, the above model can reflect the influence of aircraft attitude change , environmental medium characteristics and multi-source sound field superposition on sound pressure distribution, and provide accurate acoustic basis for subsequent noise separation and cooperative suppression; S2: Construct a dynamic compensation sound field model to perform time-frequency analysis, spatial distribution modeling and Doppler correction on the noise signal: use acoustic information to perform time-frequency analysis on the noise signal to generate frequency and time information; the spatial distribution of the noise signal is modeled to obtain the spatial distribution characteristics of the three-dimensional dynamic sound field; the frequency and time information is used to correct the noise signal to compensate for the real-time frequency shift of the noise signal, and generate a Doppler correction signal; Specifically, High-precision characterization and dynamic compensation of sound field: time-frequency characteristic analysis and spatial distribution modeling of the three-dimensional dynamic sound field model, specifically including: using adaptive S transform to perform time-frequency decomposition on the noise signal to capture the non-stationary characteristics of the dynamic sound source; using a spherical harmonic function decomposition model to approximate the spatial propagation characteristics of the sound field to high order to represent the directivity and distance decay law of sound pressure in three-dimensional space; S201: As Figure 3 and Figure 4 shown, based on the three-dimensional dynamic sound field model established in step 1, high-precision time-frequency analysis and spatial distribution modeling are performed on the noise signal to capture the non-stationary characteristics of the dynamic sound source and the spatial propagation law, and each microphone signal is generated by adaptive S transform to generate a time-frequency spectrum , wherein is the frequency, is the time, and the formula is: wherein, denotes the original time-domain signal of the m th microphone, is a time variable, is the imaginary unit; is an adaptive Gaussian window function, which changes with time to realize dynamic adjustment of the window width; is the Fourier basis function; the integral It involves weighted superposition of signal segments across the entire time axis to obtain the current time t and frequency. The time-frequency spectrum value below; in yes Instantaneous frequency at any given moment For adjustment coefficients; Through local time windows Adaptive resolution adjustment can effectively characterize the non-stationary noise characteristics generated by low-altitude aircraft at high speeds and provide frequency-time information for sound source separation and Doppler correction.

[0026] S202: To characterize the propagation characteristics in three-dimensional space, a higher-order approximation model of the spherical harmonic function is defined: in Describe the basis of spherical harmonic functions. For dynamic coefficients, For wavenumber, the above formula will convert sound pressure level to wavenumber. It decomposes the contribution of spatial angle and distance, realizing a high-order characterization of the directionality and distance attenuation of sound pressure in three-dimensional space, and providing accurate parameters for multi-source superposition and sound pressure prediction.

[0027] Specifically, Traditional sound field modeling typically targets fixed sound sources, while the above formula utilizes dynamic trajectories. and dynamic coefficient It directly adapts to the motion characteristics of the aircraft and can update the spatial distribution of the sound field in real time, using spherical harmonic functions. Describing the directional radiation characteristics of noise, such as the aerodynamic noise of an aircraft wing radiating to both sides, while the dynamic coefficient... and trajectory This reflects how noise changes over time, such as increased noise when an aircraft accelerates or changes in direction when turning; At the same time, the formula output (i.e., three-dimensional sound pressure distribution) is the spatial input for spectral analysis during adaptive S-transform; S203: To address the Doppler effect caused by the high-speed motion of an aircraft, a Doppler correction function is defined. Real-time compensation for signal frequency offset: in , The radial velocity of the aircraft relative to the microphone is used to eliminate the frequency shift caused by the aircraft's motion, while incorporating phase synchronization correction from multiple microphone arrays. , ensure the reconstructed three-dimensional sound field model Maintain accuracy and real-time under dynamic conditions, form a high-fidelity acoustic data basis that can be used for subsequent multi-source noise separation and joint optimization.

[0028] S3: Construct a dynamic correlation model, separate the input Doppler corrected signal, and output the separation result; According to the spatial distribution characteristics of the three-dimensional dynamic sound field and the separation result, the corresponding cooperative suppression strategy is generated.

[0029] S301: Multi-source noise separation and joint optimization framework: design a joint optimization framework, combine generalized cross-correlation GCC, time-varying independent vector analysis TV-IVA and interactive multi-model Kalman filter IMM-KF algorithm, to accurately separate multi-source noise in three-dimensional dynamic sound field; GCC is used to extract the time delay information of noise signal, TV-IVA separates signals of different sound sources based on time-varying independent component analysis theory, and IMM-KF processes the non-stationarity of dynamic sound sources through multi-model switching; Combine the separated sound source signal with the aircraft trajectory parameter to establish a dynamic correlation model of sound source-environment-propagation; Specifically, As Figure 5 and Figure 6 shown, based on the dynamic compensation sound field model obtained in step 2, a joint optimization framework is designed to realize accurate separation and correlation modeling of multi-source noise, first extract the time delay information of different microphone signals and and , the function is defined as: is the time correlation function of the signal collected by microphone and microphone : is the time delay, is the current time, which describes the similarity of the signal delay with signal, used to estimate the propagation delay of each sound source relative to the microphone, as a priori constraint for subsequent independent component analysis; where is the time-frequency signal after Doppler correction in step 2, is the time delay, denotes the conjugate; is the complex conjugate of the time-frequency spectrum of the signal collected by microphone : is the complex conjugate of ; Phase compensation factor for time delay: It represents the time delay, f is the frequency, and j is the imaginary unit; The core of the TV-IVA algorithm is to maximize the independence of different signals, and It is a quantitative indicator of signal independence—if two signals come from different sound sources, their This will be significantly reduced, and based on this, the mixed noise can be decomposed into independent single-source signals; The above formulas include time. and time delay It can capture correlations that change over time—for example, if the signal correlation between two microphones increases when an aircraft approaches. Increase; as the aircraft moves away, the correlation decreases, then It reduces noise and perfectly adapts to the time-varying, non-stationary characteristics of low-altitude aircraft noise.

[0030] Using time-varying independent vector analysis (TV-IVA), the multi-source signal matrix is: Decomposed into independent source vectors The above decomposition is achieved by maximizing the time correlation function: In the formula, For the unmixing matrix, As a time delay constraint weight, This indicates the expected operation.

[0031] Specifically, The above optimization function achieves accurate separation of multi-source noise by maximizing the non-Gaussianity of a single source and minimizing the correlation between signals. The specific logic is as follows: This refers to utilizing the stronger non-Gaussianity of independent signals—mixed noise is the superposition of multiple signals, while the separated independent source signals have higher non-Gaussianity. By maximizing this aspect, the separation matrix is ​​forced to... Output more independent single-source signal country This refers to adding a time correlation penalty term—if two signals come from different sound sources, their time correlation is reduced. It should be as small as possible. By subtracting this part, the independence of different signals is further strengthened, and pseudo-independent signals with high correlation are avoided. In the above formula, The time-varying separation matrix has the dimension of "number of sound sources × number of microphones" and varies with time. Dynamic updates are used to update the mixed signal matrix. Converted into independent source signals; The time-spectrum matrix of the mixed signal: f is the frequency, t is the time, and each column corresponds to the time-spectrum of the mixed noise collected by a microphone; For the separated first The time spectrum of an independent source signal: n is the index of the independent sound source. =1,2,...,N, where N is the number of sound sources; As a time delay constraint weight, This represents the expectation operation; The above algorithm continuously adjusts the separation matrix using methods such as gradient descent. ,make The value of continues to increase until j stops changing—at which point we obtain . The corresponding output signal is the noise signal of the separated independent low-altitude aircraft; Preferred, To further address the dynamic characteristics caused by changes in the aircraft's trajectory, an Interactive Multi-Model Kalman Filter (IMM-KF) was employed for each separated sound source. Establish state vector It accurately tracks non-stationary noise by switching between multiple models of probability, while simultaneously controlling the state of the sound source. With aircraft trajectory parameters and spatial location This correlation forms a dynamic correlation model of sound source-environment-propagation, which guides active noise control in generating secondary sound waves that are out of phase with the noise source. And optimization of the geometric layout of sound barriers and the distribution of sound-absorbing materials in passive noise control.

[0032] S302: Design and Simulation Verification of Collaborative Suppression Strategy: Based on three-dimensional dynamic sound field information, a collaborative suppression strategy combining active noise control and passive noise control is set up.

[0033] Specifically, The propagation characteristics of noise in complex environments are simulated using the finite difference time-domain (FDTD) simulation method. The effectiveness of sound field modeling and cooperative suppression algorithms is verified. Furthermore, the parameters of the active control algorithm and the material layout rules of the passive control strategy are optimized through numerical simulation.

[0034] Specifically, Based on the three-dimensional dynamic sound field, time-frequency characteristics, and multi-source separation results from steps 1-3, an active + passive collaborative suppression strategy was set up. Its effectiveness was verified and key parameters were optimized through FDTD simulation. The specific experimental steps are as follows: Step 4.1, Collaborative suppression strategy architecture building, according to the spatial distribution characteristics of three-dimensional dynamic sound field and the separation results of multi-source noise, the collaborative action area and division logic of active noise control ANC and passive noise control PNC are determined. For the high frequency band and transient strong noise components of low altitude aircraft noise, an active suppression module based on secondary sound source array is designed: according to the real-time position and trajectory parameters of sound source obtained by IMM-KF tracking, the phase, amplitude and emission delay of secondary sound wave are calculated, so that the secondary sound wave and noise source form destructive interference in the target area; for low frequency and diffuse noise components, a passive suppression module is designed, combined with the sound field directionality law obtained by spherical harmonic function decomposition, the geometric layout of sound barrier and the distribution scheme of sound absorption material are determined. At the same time, the linkage trigger mechanism of active-passive suppression is established, according to the Doppler corrected noise time-frequency signal intensity, the gain coefficient of active control and the opening / closing / adjustment state of passive barrier are dynamically adjusted, realizing the adaptive collaboration of the two.

[0035] Step 4.2, simulation model building, based on finite difference time domain FDTD method, numerical simulation model of noise propagation and suppression is built, parameters of multi-physical field sound propagation equation in step 1, time-frequency characteristics and Doppler correction data of step 2, and multi-source noise separation results of step 3 are imported into FDTD simulation platform, three-dimensional simulation scene matching real low altitude aircraft noise environment is constructed, spatial position of microphone array, dynamic trajectory of aircraft, geometric parameters and material properties of active secondary sound source array and passive sound barrier are accurately restored in the model, at the same time, time step and spatial grid accuracy of simulation are set: time step is determined according to time-frequency resolution of adaptive S transform, to ensure that non-stationary characteristics of noise can be captured; spatial grid accuracy is set according to high order approximation order of spherical harmonic function, to ensure spatial calculation accuracy of sound field propagation.

[0036] Step 4.3, suppression effect simulation test, in the FDTD simulation model built, the whole trajectory noise release process of low altitude aircraft from take-off, cruising to landing is simulated, active and passive collaborative suppression system is started synchronously, multi-dimensional suppression effect test is carried out. First, the sound pressure level attenuation of target area is detected, the sound pressure distribution cloud map when the suppression system is opened / closed is compared, and the noise attenuation amount of different frequency bands is calculated; secondly, the real-time performance of suppression system is analyzed, the response time from sound source position change to active secondary sound source parameter adjustment is counted, to verify whether it meets the real-time suppression demand of dynamic motion of aircraft; finally, the collaborative suppression effect of multi-source noise is evaluated, for different sound source components such as engine noise and aerodynamic noise, the noise reduction efficiency of single suppression method and collaborative suppression strategy is tested respectively, to determine the advantage scene of active-passive collaboration.

[0037] Step 4.4, parameter iteration optimization, according to the noise attenuation data obtained by simulation test, real-time index and multi-source suppression effect, carry out parameter iteration optimization of collaborative suppression strategy. For the active noise control module, taking the minimum sound pressure level of the target area as the objective function, adjusting the number, spatial arrangement and emission parameters (such as phase compensation coefficient, amplitude gain) of the secondary sound source, using gradient descent method to optimize the update rate of the demixing matrix , improve the accuracy of active suppression; for the passive noise control module, according to the reflection and diffraction law of noise in the sound field simulation, adjust the structural parameters of the sound barrier and the combination scheme of the sound absorption material, optimize the sound absorption efficiency of the material in different frequency bands. After each parameter adjustment, retest by FDTD simulation, compare the suppression effect before and after optimization, until each index meets the design requirements, and finally form the optimal collaborative suppression strategy parameter set suitable for the mixed dynamic sound source of low-altitude aircraft.

[0038] Preferably, The method further comprises: experimental platform construction and cross-platform verification, constructing an experimental platform including multiple low-altitude aircrafts, microphone arrays, active noise control devices and passive noise control devices, simulating a dynamic noise environment in a real scene, and verifying the noise control effect and real-time performance of the three-dimensional dynamic sound field modeling, multi-source noise separation and collaborative suppression algorithm through experiments.

[0039] Embodiment 2 The embodiment provides a transformer internal fault degree detection system, including a memory and a processor, the memory stores a computer program, and the processor calls the computer program to execute the steps of the method of any one of the above.

[0040] The working principle and beneficial effects of the present application are: The present application realizes high-precision analysis of low-altitude aircraft noise through accurate separation and modeling of multi-source dynamic sound field. The joint optimization framework of generalized cross-correlation, time-varying independent vector analysis and interactive multi-model Kalman filter can effectively separate different sound source components in mixed noise and establish a dynamic correlation model of sound source-environment-propagation. This technology solves the problem that traditional methods are difficult to handle dynamic multi-source sound field, provides accurate acoustic basis for subsequent noise suppression, and significantly improves the accuracy and reliability of noise analysis.

[0041] The present application realizes multi-dimensional control of dynamic noise through active and passive collaborative suppression strategy. Based on three-dimensional sound field information and multi-source separation results, the active noise control generates secondary sound waves with opposite phase to the noise source, and combines the optimization of sound barrier geometric layout and sound absorption material distribution to form an active and passive combined suppression system. This technology breaks through the limitations of single suppression method, can effectively control different frequency band noises in complex environment, and significantly improves the comprehensive effect of noise suppression.

[0042] In steps 2 and 3, the present application realizes accurate analysis and separation of low-altitude aircraft dynamic noise by constructing a multi-level, adaptive acoustic signal processing system. In step 2, adaptive S-transform is used for time-frequency analysis of non-stationary noise signals, effectively capturing the instantaneous frequency characteristics of noise signals under high-speed motion state by dynamically adjusting the time window width. Meanwhile, spatial modeling of three-dimensional sound field is combined with spherical harmonic function decomposition to accurately represent the propagation law of sound pressure with spatial angle and distance. The Doppler real-time compensation mechanism is introduced to eliminate the frequency shift caused by aircraft motion, significantly improving the fidelity of dynamic sound field reconstruction. In step 3, the GCC-TV-IVA-IMM-KF joint optimization framework is set up. First, the time delay information of multi-channel signals is extracted by generalized cross-correlation. Then, time-varying independent vector analysis is used to realize blind source separation of multi-source signals in the time-frequency domain. Finally, the dynamic characteristics of each sound source are tracked by interacting multiple model Kalman filter, establishing a dynamic correlation model of sound source-environment-propagation. The synergistic effect of these two steps enables the system to accurately separate and track different sound source components such as engine noise and aerodynamic noise in complex environments, providing high-precision, real-time acoustic data support for subsequent collaborative noise suppression, effectively solving the technical bottleneck of insufficient dynamic multi-source sound field analysis capability of traditional methods.

[0043] Noun explanation: Beamforming technology: It is a spatial signal processing technology based on sensor array. By adjusting the amplitude and phase of the signals received by each sensor in the array (i.e. "beamforming"), the array forms a directional receiving or transmitting beam in space, enhancing the gain of signals from a specific direction and suppressing interference from other directions, achieving spatial selective perception and signal enhancement. In the field of acoustics, this technology is widely used in sound source positioning, noise suppression, speech enhancement, and sound field reconstruction. Its essence is to realize signal filtering and separation in the spatial domain, which can be divided into time-domain beamforming and frequency-domain beamforming according to the processing domain, and is often combined with adaptive algorithms to cope with dynamic sound field environments.

[0044] Sound source positioning technology: It is a technology that uses the sound wave signals received by an acoustic sensor array to analyze the time delay, phase difference, or amplitude difference between signals, and combines the array geometry and signal processing algorithms to estimate and track the position of one or more sound sources in space. Its core is to extract the direction or three-dimensional coordinates of the sound source from the mixed sound field. Common methods include geometric positioning based on time difference (TDOA), scanning positioning based on beamforming, and subspace-based methods based on high-resolution spectrum estimation. This technology is widely used in noise source identification, speech interaction, intelligent security, underwater detection, and acoustic imaging, and is the basis for realizing sound field perception and active control.

[0045] Multi-Physics Acoustic Propagation Equation: A governing equation that describes the propagation of sound waves in complex media with non-uniform, flow, thermodynamic non-equilibrium, or coupling with other physical fields (e.g., fluid, structure, electromagnetic field).

[0046] Adaptive S-Transform: A high-resolution time-frequency analysis technique for non-stationary signals, which introduces an adjustable time-frequency window function based on the traditional S-transform. It dynamically adjusts the window width according to the local characteristics of the signal (such as instantaneous frequency, energy distribution), thereby achieving an adaptive balance between time resolution and frequency resolution.

[0047] Generalized Cross-Correlation (GCC): A signal processing technique for high-precision estimation of the time difference between two signals. It is an enhanced version of the classic cross-correlation method, which suppresses noise and reverberation effects by weighting the cross-power spectrum in the frequency domain, thereby more clearly and robustly extracting the time delay peak.

[0048] Time-Varying Independent Vector Analysis (TV-IVA): An advanced blind source separation technique specifically designed for processing multi-channel observation signals of a mixed system that changes over time. It breaks through the assumption of a fixed and unchanging mixing matrix based on traditional independent vector analysis, and models the time-varying characteristics of the mixing system to achieve simultaneous tracking and separation of multiple source signals in a dynamic environment.

[0049] Interacting Multiple Model Kalman Filter (IMM-KF): An advanced state estimation method for tracking targets with multiple possible motion models. It runs multiple Kalman filters in parallel (each corresponding to a candidate motion model), and based on model matching probability, it interacts and fuses the estimation results of each filter in real time, thereby adaptively handling the uncertainty of target motion patterns and significantly improving tracking accuracy and robustness in complex maneuvering scenarios.

[0050] Spherical Harmonics: A set of orthogonal and complete basis functions defined on the unit sphere, used to describe the function distribution in the spherical direction, mainly used for spatial decomposition and reconstruction of three-dimensional sound fields.

[0051] Doppler Correction Function: A mathematical tool used to compensate for or model the frequency shift effect caused by relative motion. When there is relative radial motion between the sound source and the receiver, the received frequency will deviate from the transmitted frequency, i.e., the Doppler effect. This function eliminates or quantifies this frequency shift by stretching the time domain, adjusting the phase, or correcting the frequency domain, thereby restoring the original spectral characteristics of the signal or accurately estimating the relative motion parameters.

[0052] Finite-difference time-domain (FDTD) simulation method: a numerical calculation technique for directly solving the time-domain Maxwell's equations (or acoustic wave equation). It discretizes the continuous space and time into grids, replaces partial derivatives with difference approximations, and iteratively updates the field values at each grid point in time to simulate the whole process of electromagnetic wave or acoustic wave propagation, scattering and interaction in complex media.

[0053] The above merely describes preferred embodiments of the present application, and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0054] The preferred embodiments of the present application are described in detail above. It should be understood that those of ordinary skill in the art can make many modifications and changes without requiring creative efforts, based on the concept of the present application. Therefore, any technical solution obtained by logical analysis, reasoning or limited experiments based on the prior art according to the concept of the present application shall be within the protection scope defined by the claims.

Claims

1. A method of mixed dynamic acoustic source interference modeling and multi-source collaborative suppression, characterized in that, The method comprises the following steps: deploying a distributed microphone array and collecting low-altitude aircraft noise signals; processing the noise signals by using a pre-constructed three-dimensional dynamic sound field model, and outputting acoustic information; constructing a dynamic compensation sound field model to perform time-frequency analysis, spatial distribution modeling, and Doppler correction on the noise signals: using the acoustic information, performing time-frequency analysis on the noise signals to generate frequency and time information; performing spatial distribution modeling on the noise signals to obtain the spatial distribution characteristics of the three-dimensional dynamic sound field; using the frequency and time information, performing Doppler correction on the noise signals to compensate for the frequency shift of the noise signals in real time and generate a Doppler-corrected signal; constructing a dynamic correlation model to separate the input Doppler-corrected signal and output a separation result; generating a corresponding cooperative suppression strategy according to the spatial distribution characteristics of the three-dimensional dynamic sound field and the separation result.

2. The method of claim 1, wherein, The three-dimensional dynamic sound field model specifically comprises: wherein is the ambient medium sound speed distribution, is the sound source excitation term, is the low flying aircraft noise signal, is the Laplacian of the sound pressure at spatial location r and time t, describing the rate of change of the sound pressure distribution in three dimensions.

3. The method of claim 1, wherein, The time-frequency analysis specifically comprises: using adaptive S transform to perform time-frequency decomposition on the noise signals, and the calculation formula is: wherein, is a time-frequency spectrum, is a microphone acquisition signal, is a frequency, is a time, is a local time window, is a spatial position of the mth microphone, is an imaginary unit, is an adaptive Gaussian window function, is a Fourier basis function; the integral is a weighted superposition of signal segments over the entire time axis, resulting in a time-frequency spectrum value at the current time t, frequency is the instantaneous frequency at the time t, is an adjustment coefficient.​ 4. The method of claim 1, wherein, The spatial distribution modeling specifically uses a spherical harmonic function decomposition model to perform high-order approximation on the spatial propagation characteristics of the sound field to represent the directivity and distance attenuation law of the sound pressure in three-dimensional space, and the spherical harmonic function decomposition model is: wherein is the sound pressure, is the dynamic coefficient, denotes a spherical harmonic basis, is the wave number, is the spatial position, is the spatial position, is the propagation phase and attenuation term of the sound wave from the sound source position to the spatial position.

5. The method of claim 1, wherein, The Doppler correction specifically comprises: for the Doppler effect caused by the high-speed motion of the aircraft, defining a Doppler correction function to compensate for the frequency shift of the signal in real time, and the calculation formula is: wherein, is a Doppler correction function, is a radial velocity of the vehicle relative to the microphone, is a time-frequency spectrum of the mth microphone Doppler corrected at spatial position is an ambient medium sound speed at the real-time spatial position of the sound source.​ 6. The method of claim 1, wherein, The process of constructing the dynamic correlation model comprises: using a generalized cross-correlation function to extract the time delay information of the noise signal; using a time-varying independent vector analysis algorithm, combining a maximum time correlation function to decompose a multi-source signal matrix into independent source vectors; and using an interacting multiple model Kalman filter to establish a state vector for each separated independent source vector; and correlating the state vector with the aircraft trajectory parameters and the spatial position to form the dynamic correlation model.

7. The method of claim 6, wherein, The generalized cross-correlation function is: wherein is the Doppler corrected time-frequency signal, is the time delay, denotes the conjugate, is the microphone collects the complex conjugate of the time-frequency spectrum of the signal, is the time-delayed phase compensation factor, is the spatial position of the i-th microphone, is the complex conjugate of is the spatial position of the j-th microphone; The time-varying independent vector analysis algorithm comprises: wherein, is a multi-source signal matrix, is an independent source vector, is an independent source vector, is a separated sound source; The maximum time correlation function is: wherein, is an unmixing matrix, is a delay constraint weight, denotes an expectation operation; is a time-frequency spectrum of the separated th independent source signal: n is an index of an independent sound source, = 1, 2,..., N, N is the number of sound sources; The state vector is represented as: wherein, is the n-th independent source signal after separation, is the time derivative thereof.

8. The method of claim 1, wherein, The cooperative suppression strategy specifically comprises: According to the spatial distribution characteristics of the three-dimensional dynamic sound field and the separation result, the cooperative action area and the division logic of active noise control and passive noise control are determined; at the same time, a linkage triggering mechanism of the active noise control and the passive noise control is established, the gain coefficient of the active noise control and the opening and adjustment state of the passive noise control are dynamically adjusted according to the intensity of the Doppler-corrected noise time-frequency signal, and adaptive cooperation of the active noise control and the passive noise control is realized. The active noise control comprises: generating a secondary sound wave opposite in phase to the noise source to achieve active noise reduction; and the passive noise control comprises: optimizing the geometric layout and sound-absorbing material distribution of the sound barrier to achieve passive noise reduction.

9. The method of claim 1, wherein, The method further comprises: using a finite difference time domain simulation method to verify the noise suppression effect of the cooperative suppression strategy, and continuously optimizing the cooperative suppression strategy through numerical simulation. An experimental platform is constructed, which comprises a plurality of low-altitude flying vehicles, a microphone array, an active noise control device and a passive noise control device, a dynamic noise environment in a real scene is simulated, and performance of the dynamic compensation sound field model, separation effect of the Doppler correction signal and noise suppression effect of the cooperative suppression strategy are verified through experiments.

10. A system for mixed dynamic acoustic source interference modeling and multi-source coordinated suppression, the system comprising: The computer program is stored in the memory and comprises a plurality of instructions, which are executed by the processor to perform the steps of the method according to any one of claims 1 to 9.

Images