Vibration noise intelligent control method and system using frequency spectrum characteristics

Abstract

The invention relates to the related technical field of vibration noise control, in particular to an intelligent vibration noise control method and system using frequency spectrum characteristics, and the method comprises the steps: collecting rail transit vibration noise signals, extracting time-frequency domain mixed characteristics, carrying out multi-source separation, generating a noise source contribution degree matrix, configuring a strategy set, and monitoring changes to generate feedback. And the active damper is adjusted in real time for noise reduction. The technical problems that the noise suppression effect is limited and the noise reduction requirement of a complex time-varying scene cannot be accurately matched due to the fact that vibration noise control is difficult to adapt to a dynamically changing vibration noise environment are solved, time-frequency domain mixed features are extracted, overlapped noise sources are identified and separated more accurately, complex characteristics of vibration and noise are comprehensively considered, and the noise reduction effect is improved. And the active damper is dynamically adjusted so as to adapt to changes of vibration and noise, and the accuracy and effectiveness of dynamic adjustment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of vibration and noise control, and particularly to an intelligent vibration and noise control method and system using spectral features. Background Art

[0002] With the acceleration of the urbanization process and the increasing development of the transportation network, as an important part of urban public transportation, the vibration and noise problems generated by the operation of rail transit systems are becoming increasingly prominent. These vibrations and noises not only affect the quality of life of surrounding residents, but also cause damage to surrounding buildings and infrastructure.

[0003] At present, there are various control methods for rail transit vibration and noise, but most of them have limitations. Conventional passive control methods, such as installing sound insulation walls and vibration damping pads, although they can reduce noise and vibration to a certain extent, still have difficulty adapting to complex and changing traffic environments and noise source characteristics. Vibration and noise control mostly relies on single spectral features or time-domain features, making it difficult to comprehensively and accurately describe the complex characteristics of vibrations and noises, and the control effect is limited.

[0004] In summary, there are technical problems in the prior art that vibration and noise control are difficult to adapt to dynamic vibration and noise environments, resulting in limited noise suppression effects and inability to accurately match the noise reduction requirements of complex time-varying scenarios. Summary of the Invention

[0005] The present application provides an intelligent vibration and noise control system using spectral features, aiming to solve the technical problems in the prior art that vibration and noise control are difficult to adapt to dynamic vibration and noise environments, resulting in limited noise suppression effects and inability to accurately match the noise reduction requirements of complex time-varying scenarios.

[0006] In view of the above problems, the technical solution of the present application is as follows: On the one hand, the present application provides an intelligent vibration and noise control method using spectral features. The method includes: collecting vibration and noise signals of rail transit and extracting time-frequency domain hybrid features; performing multi-source separation on the vibration and noise signals to generate a noise source contribution matrix, where the noise source contribution matrix includes wheel-rail contact noise contribution values, aerodynamic noise contribution values, and environmental reflection superposition coefficients; configuring an adaptive control strategy set including active damper frequency band adjustment parameters, acoustic absorption material layout optimization schemes, dynamic stiffness thresholds of vibration isolators, and phase compensation strategies according to the time-frequency domain hybrid features and the noise source contribution matrix; simultaneously, monitoring the noise synchronous attenuation rate and the change in the vibration transfer function to generate a dynamic feedback signal; and dynamically adjusting the active damper based on the adaptive control strategy set in combination with the dynamic feedback signal, where the active damper is used to actively suppress vibration and noise.

[0007] On the other hand, the present application provides an intelligent vibration and noise control system using spectral features. The system includes: a signal acquisition module for acquiring vibration and noise signals of rail transit and extracting time-frequency domain hybrid features; a multi-source separation module for performing multi-source separation on the vibration and noise signals to generate a noise source contribution matrix, where the noise source contribution matrix includes a wheel-rail contact noise contribution value, an aerodynamic noise contribution value, and an environmental reflection superposition coefficient; an adaptive configuration module for configuring an adaptive control strategy set including active damper frequency band adjustment parameters, acoustic absorption material layout optimization schemes, vibration isolator stiffness dynamic thresholds, and phase compensation strategies according to the time-frequency domain hybrid features and the noise source contribution matrix; a monitoring module for monitoring the noise synchronous attenuation rate and the change in the vibration transfer function simultaneously to generate a dynamic feedback signal; and a dynamic adjustment module for dynamically adjusting the active damper based on the adaptive control strategy set in combination with the dynamic feedback signal, where the active damper is used to suppress vibration and noise.

[0008] In summary, one or more technical solutions provided in the present application extract time-frequency domain hybrid features, more accurately identify and separate overlapping noise sources, comprehensively consider the complex characteristics of vibration and noise, dynamically adjust the active damper, and thus adapt to changes in vibration and noise, achieving the technical effects of improving the accuracy and effectiveness of dynamic adjustment. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 is a schematic flow chart of an intelligent vibration and noise control method using spectral features provided by the present application; Figure 2 is a schematic structural diagram of an intelligent vibration and noise control system using spectral features provided by the present application.

[0010] Description of reference numerals: signal acquisition module M100, multi-source separation module M200, adaptive configuration module M300, monitoring module M400, dynamic adjustment module M500. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0011] Embodiment 1 The present application will be specifically described below with reference to the accompanying drawings. As Figure 1 shown, the present application provides an intelligent vibration and noise control method using spectral features. The method includes: S1: Acquire vibration and noise signals of rail transit and extract time-frequency domain hybrid features; S2: Perform multi-source separation on the vibration and noise signals to generate a noise source contribution matrix, where the noise source contribution matrix includes a wheel-rail contact noise contribution value, an aerodynamic noise contribution value, and an environmental reflection superposition coefficient.

[0012] Specifically, collecting vibration and noise signals means obtaining the vibration and noise data generated by rail transit in real time through sensors (such as accelerometers, microphones, etc.); the time-frequency domain hybrid features refer to the characteristic parameters extracted simultaneously in the time domain and the frequency domain, which are used to comprehensively describe the complex characteristics of vibration and noise. The time-domain features include transient impact peaks, background noise baselines, etc.; the frequency-domain features include spectral energy distribution, harmonic distortion rate; extracting features means extracting key characteristic parameters from the original signal through signal processing algorithms (such as Fourier transform, wavelet transform, etc.).

[0013] In the acquisition stage, high-precision sensors (such as triaxial accelerometers and array microphones) are used to obtain vibration and noise signals in real time. The signal acquisition frequency is usually set from 10 kHz to 20 kHz to cover the common vibration and noise frequency bands in rail transit (such as vehicle engine vibration from 100 Hz to 500 Hz, wheel-rail contact noise from 500 Hz to 2 kHz). Further, when extracting the time-frequency domain hybrid features, the short-time Fourier transform (STFT) or continuous wavelet transform (CWT) is used to analyze the signal. For example, the spectral energy distribution is obtained by calculating the power spectral density (PSD) of the signal, the harmonic distortion rate is quantified by calculating the total harmonic distortion (THD), the transient impact peak is extracted by detecting the mutation points of the signal envelope, and the background noise baseline is estimated by the average energy in the low-frequency band (0 Hz to 50 Hz); through the above steps, the dynamic characteristics of vibration and noise are comprehensively described, providing a data basis for subsequent noise source separation and control strategies.

[0014] Multi-source separation means decomposing the mixed vibration and noise signals into multiple independent noise sources through signal processing algorithms (such as independent component analysis ICA, blind source separation BSS, etc.); the noise source contribution matrix is used to represent the contribution ratio of each noise source to the total noise. The parameters in the matrix include the contribution value of wheel-rail contact noise, the contribution value of aerodynamic noise, and the environmental reflection superposition coefficient; the contribution value refers to the proportion or intensity of a certain noise source in the total noise, usually calculated by energy or power.

[0015] Multi-source separation uses the independent component analysis (ICA) algorithm to decompose the collected vibration noise signal into multiple independent noise sources. Further, in the rail transit scenario, the contribution value of wheel-rail contact noise is calculated by the energy ratio in the middle frequency band (500 Hz to 2 kHz), the contribution value of aerodynamic noise is calculated by the energy ratio in the high frequency band (2 kHz to 10 kHz), the environmental reflection superposition coefficient is estimated by analyzing the coherence of the signal and the phase delay of the reflection path. The generation process of the noise source contribution matrix includes: performing energy normalization processing on the characteristic frequency band of each noise source; calculating the energy ratio of each noise source to form a matrix; introducing the environmental reflection superposition coefficient to correct the superposition effect of the reflection path on the noise. In the above steps, by generating the noise source contribution matrix, the main noise sources and their contribution ratios can be accurately identified, providing a quantitative basis for the subsequent adaptive control strategy.

[0016] S3: Configure an adaptive control strategy set including the frequency band adjustment parameters of the active damper, the optimized layout scheme of the sound-absorbing material, the dynamic threshold of the stiffness of the vibration isolator, and the phase compensation strategy according to the time-frequency domain hybrid feature and the noise source contribution matrix; S4: At the same time, monitor the synchronous attenuation rate of the noise and the change amount of the vibration transfer function to generate a dynamic feedback signal; S5: Based on the adaptive control strategy set, combined with the dynamic feedback signal, dynamically adjust the active damper, and the active damper is used to actively suppress vibration noise.

[0017] Specifically, the frequency band adjustment parameter of the active damper refers to the adjustment parameter of the active damper in a specific frequency band, which is used to suppress vibration noise in a specific frequency range; the optimized layout scheme of the sound-absorbing material refers to the layout scheme of the sound-absorbing material in space, and the sound-absorbing effect is improved by optimizing the layout; the dynamic threshold of the stiffness of the vibration isolator refers to the stiffness adjustment range of the vibration isolator in a dynamic environment, which is used to adapt to different vibration loads; the phase compensation strategy refers to compensating for the propagation delay of vibration noise through phase adjustment to ensure the synchronization of the control signal and the noise signal; the adaptive control strategy set refers to a set of dynamically adjusted control strategies that can be optimized in real time according to the noise characteristics and environmental changes.

[0018] Determine the frequency band range to be suppressed according to the spectral energy distribution and the noise source contribution matrix. Specifically, if the contribution value of wheel-rail contact noise is relatively high (such as 45%), then preferentially adjust the suppression intensity of the active damper in the frequency band from 500 Hz to 2 kHz; combine the heat map of the spatial distribution of noise sources to optimize the layout position of the sound-absorbing material, and increase the coverage density of the sound-absorbing material in the area with a high noise source contribution value (such as the wheel-rail contact noise area); according to the change amount of the vibration transfer function, dynamically adjust the stiffness of the vibration isolator. For example, when the change amount of the vibration transfer function exceeds 5%, trigger the dynamic stiffness adjustment mechanism and adjust the stiffness threshold from 10 kN / m to 15 kN / m; by monitoring the noise synchronous attenuation rate and the change amount of the vibration transfer function, adjust the phase compensation parameter in real time. For example, if the phase delay exceeds 1 ms, then adjust the phase compensation value to ensure the synchronization of the control signal and the noise signal. Through the above steps, configure an adaptive control strategy set to dynamically adapt to the characteristics of different noise sources and improve the flexibility and adaptability of the control system.

[0019] The noise synchronous attenuation rate refers to the attenuation consistency of noise under different paths or different control strategies; the change amount of the vibration transfer function refers to the change of the transfer characteristics of the vibration signal in the propagation path, usually represented by the frequency response function; the dynamic feedback signal refers to the noise and vibration change data monitored in real time, which is used to dynamically adjust the control strategy.

[0020] Real-time monitor the noise attenuation situation through multi-point sound pressure level sensors and calculate the attenuation rate consistency under different paths; through acceleration sensors and transfer function analysis tools, real-time monitor the change of the vibration transfer function. The role of the dynamic feedback signal is to sense the changes of noise and vibration in real time and provide a basis for adjusting the adaptive control strategy. For example, if the noise attenuation rate of a certain section of rail transit is lower than 10% of the overall average value, then trigger the first dynamic feedback signal. At the same time, if the amplitude change of the transfer function exceeds 10% or the phase change exceeds 5°, then generate the second dynamic feedback signal. Then, trigger the optimization of the frequency band adjustment parameters of the active damper to ensure the noise reduction effect.

[0021] Active vibration and noise suppression refers to the real-time suppression of vibration and noise through active control means (such as active dampers); according to the dynamic feedback signal, real-time adjust the frequency band adjustment parameters and phase compensation strategy of the active damper; optimize the layout of the sound-absorbing material and the stiffness of the vibration isolator through the dynamic feedback signal; when it is monitored that the noise attenuation effect of a certain frequency band decreases, increase the suppression intensity of that frequency band; when it is monitored that the change amount of the vibration transfer function in a certain area is large, then adjust the dynamic threshold of the stiffness of the vibration isolator to ensure the isolation effect; this dynamic adjustment mechanism can significantly improve the suppression effect of the active damper.

[0022] Furthermore, the method further includes: The time-frequency domain hybrid features include spectral energy distribution, harmonic distortion rate, transient impact peak, and background noise baseline; based on the time-frequency domain hybrid features, combined with direction-of-arrival estimation and coherence analysis, a heat map of the spatial distribution of noise sources is generated; independent component extraction is performed on the overlapping noise sources in the heat map of the spatial distribution of noise sources, and vibration sensor data is correlated to verify the multi-source separation result.

[0023] Specifically, the spectral energy distribution refers to the energy distribution of the vibration noise signal in different frequency bands, usually calculated by the power spectral density (PSD); the harmonic distortion rate refers to the ratio of the harmonic components to the fundamental component in the signal, used to quantify the distortion degree of the signal; the transient impact peak refers to the sudden high-amplitude short-time impact in the signal; the background noise baseline refers to the static level of the environmental noise.

[0024] The spectral energy distribution is calculated by the short-time Fourier transform (STFT) or wavelet transform (WT), which can identify the frequency bands of the main noise sources; the harmonic distortion rate is quantified by calculating the total harmonic distortion (THD), which can identify the harmonic distortion in the signal; the transient impact peak is extracted by signal envelope detection, which can locate the sudden noise events. Further, in rail transit, the vehicle engine vibration is mainly concentrated in the range of 100 Hz to 500 Hz, the wheel-rail contact noise is concentrated in the range of 500 Hz to 2 kHz, and the aerodynamic noise is concentrated in the range of 2 kHz to 10 kHz; if the THD exceeds 5%, it indicates that there is significant harmonic distortion in the signal, which may be caused by mechanical vibration or electrical interference.

[0025] It has been verified that in rail transit, the transient impact peak of wheel-rail contact usually appears when the train passes through the track joint, and the amplitude can reach more than 10 times that of the background noise. Further, the background noise baseline is estimated by the average energy in the low-frequency band (0 Hz to 50 Hz), which can evaluate the static level of the environmental noise. For example, in urban rail transit, the background noise baseline is usually between 30 dB and 40 dB; the spectral energy distribution, harmonic distortion rate, transient impact peak, and background noise baseline in the time-frequency domain hybrid features are used to comprehensively describe the dynamic characteristics of the vibration noise, providing a data basis for subsequent noise source separation and control strategies.

[0026] Direction of arrival (DOA) estimation refers to estimating the direction of a noise source through array signal processing techniques; coherence analysis refers to evaluating the correlation between different noise sources by calculating the coherence between signals; the heatmap of the spatial distribution of noise sources shows the spatial distribution of noise sources in the form of a heatmap; DOA estimation uses a high-resolution algorithm based on array microphones (such as the MUSIC algorithm) to accurately estimate the direction of noise sources; coherence analysis evaluates the correlation between noise sources by calculating the coherence matrix between different sensors; the heatmap of the spatial distribution of noise sources can visually display the spatial distribution of noise sources by visualizing the results of DOA estimation and coherence analysis. Further, if the coherence coefficient between two noise sources exceeds 0.7, it indicates that the two noise sources come from the same noise source or there is a coupling relationship. Correspondingly, the heatmap shows that the wheel-rail contact noise is mainly concentrated on both sides of the track, while the aerodynamic noise is distributed in front of the train running direction; the heatmap of the spatial distribution of noise sources is used to identify the spatial distribution characteristics of noise sources and provide a basis for subsequent optimization of the layout of sound-absorbing materials and the arrangement of vibration isolators.

[0027] Independent component extraction (ICA) refers to decomposing a mixed signal into independent noise sources; vibration sensor data association refers to associating the data of vibration sensors with the results of noise source separation to verify the accuracy of separation; independent component extraction uses the Fast Independent Component Analysis (FastICA) algorithm to separate overlapping noise sources into independent components; vibration sensor data association verifies the accuracy of the multi-source separation results by comparing the data of vibration sensors with the separated noise sources. For example, if a separated noise source is highly consistent with the wheel-rail contact vibration characteristics recorded by a vibration sensor, it indicates that the separation result is reliable; through this verification mechanism, the accuracy of the multi-source separation results can be ensured, providing reliable data support for subsequent adaptive control strategies.

[0028] Furthermore, for dynamic adjustment of the active damper, the method includes: Using the spectral energy distribution and harmonic distortion rate in the time-frequency domain hybrid features to predict the resonance frequency band and determine the target suppression frequency band of the active damper; determining the frequency band adjustment parameters of the active damper according to the target suppression frequency band.

[0029] Specifically, the spectral energy distribution refers to the energy distribution of vibration noise signals in different frequency bands, usually calculated by the power spectral density (PSD); the harmonic distortion rate refers to the ratio of the harmonic components to the fundamental component in a signal, used to quantify the distortion degree of the signal, usually calculated by the total harmonic distortion (THD); resonance frequency band prediction is to predict the possible resonance frequency band in vibration noise by analyzing the spectral energy distribution and harmonic distortion rate; the target suppression frequency band refers to the frequency band that the active damper needs to focus on suppressing, usually the resonance frequency band or the frequency band with a high noise contribution value.

[0030] The spectral energy distribution is calculated by short-time Fourier transform (STFT) or wavelet transform (WT), which can identify the frequency bands of the main noise sources; the harmonic distortion rate is quantified by calculating the total harmonic distortion (THD), which can identify the harmonic distortion in the signal; the resonance frequency band prediction identifies the frequency bands with concentrated energy and high harmonic distortion rate by analyzing the spectral energy distribution and the harmonic distortion rate; after determining the target suppression frequency band, the active damper will focus on suppressing the vibration noise in this frequency band to improve the control effect.

[0031] The frequency band adjustment parameters refer to the adjustment parameters of the active damper within a specific frequency band, including gain, phase, bandwidth, etc., which are used to optimize the suppression effect; the active damper refers to a device that suppresses vibration noise through active control means and can dynamically adjust the damping characteristics according to the control signal. According to the energy magnitude of the target suppression frequency band, the gain parameter of the active damper is set; according to the phase characteristics of the vibration noise, the phase compensation value of the active damper is adjusted to ensure that the control signal is synchronized with the noise signal; according to the width of the target suppression frequency band, the bandwidth parameter of the active damper is adjusted. By optimizing the frequency band adjustment parameters, the active damper can more accurately suppress the vibration noise in the target frequency band.

[0032] Furthermore, the method includes: Using the transient shock peak and the background noise baseline in the time-frequency domain hybrid features to determine the dynamic adaptation range of the porosity of the sound-absorbing material; adjusting the layout optimization scheme of the sound-absorbing material according to the dynamic adaptation range of the porosity of the sound-absorbing material.

[0033] Specifically, the transient shock peak refers to a sudden high-amplitude short-time shock in the vibration noise signal, which is usually extracted by signal envelope detection; the background noise baseline refers to the static level of the ambient noise; the dynamic adaptation range of the porosity refers to the adjustable range of the porosity of the sound-absorbing material in different noise environments, which is used to optimize the sound-absorbing effect; the transient shock peak is extracted by signal envelope detection, which can identify sudden noise events; the background noise baseline is estimated by the average energy in the low-frequency band (0 Hz to 50 Hz), which can evaluate the static level of the ambient noise; the dynamic adaptation range of the porosity is determined by analyzing the transient shock peak and the background noise baseline. For example, if the transient shock peak is high (such as more than 15 dB above the background noise baseline), the porosity of the sound-absorbing material needs to be increased to improve the sound-absorbing effect. By adjusting the porosity, the sound-absorbing material can more effectively absorb noises of different frequencies and intensities.

[0034] The layout optimization scheme of the sound-absorbing material refers to the layout scheme of the sound-absorbing material in space, which improves the sound-absorbing effect through optimization; the dynamic adaptation range of the porosity refers to the adjustable range of the porosity of the sound-absorbing material in different noise environments, which is used to optimize the sound-absorbing effect.

[0035] Identify the distribution of noise sources through the heat map of the spatial distribution of noise sources; determine the optimal layout positions of the sound-absorbing materials according to the distribution of noise sources and the dynamic adaptation range of porosity; optimize the layout scheme of the sound-absorbing materials through simulation and experimental verification. In the rail transit scenario, the heat map shows that the wheel-rail contact noise is mainly concentrated on both sides of the track, while the aerodynamic noise is distributed in front of the train running direction. In the area with higher wheel-rail contact noise, increase the coverage density of the sound-absorbing materials and adjust the porosity to a higher value (such as 30% to 40%). By adjusting the layout positions and porosity of the sound-absorbing materials, improve the sound pressure level attenuation and significantly enhance the sound absorption effect, reducing the impact of environmental noise on surrounding residents and infrastructure.

[0036] Furthermore, the method includes: Set the noise control effect evaluation indicators, which include the sound pressure level attenuation, the vibration acceleration reduction rate, and the energy consumption index of strategy execution; at the same time, optimize the parameters of the adaptive control strategy set, balance the noise reduction effect and the equipment life loss, and dynamically adjust the triggering conditions of the phase compensation strategy according to the historical control records.

[0037] Specifically, the sound pressure level attenuation refers to the difference in sound pressure level before and after noise control, usually in decibels (dB); the vibration acceleration reduction rate refers to the reduction ratio of vibration acceleration before and after vibration control, usually expressed as a percentage; the energy consumption index of strategy execution refers to the energy consumed during the execution of the noise control strategy, usually in watt-hours (Wh).

[0038] The sound pressure level attenuation is calculated by comparing the sound pressure levels before and after control. If the sound pressure level before control is 80 dB and the sound pressure level after control is 65 dB, then the sound pressure level attenuation is 15 dB; the vibration acceleration reduction rate is calculated by comparing the peak values of vibration acceleration before and after control. If the peak value of vibration acceleration before control is 0.5 m / s² and the peak value after control is 0.2 m / s², then the vibration acceleration reduction rate is 60%; the energy consumption index of strategy execution is obtained by monitoring and accumulating the energy consumption of the active damper and other equipment.

[0039] Parameter optimization refers to the process of finding the optimal parameters in the control strategy through optimization algorithms (such as genetic algorithms, particle swarm optimization, etc.); Equipment life loss refers to the reduction of the service life of the equipment due to factors such as wear and aging during operation; The triggering condition of the phase compensation strategy refers to the condition for starting the phase compensation mechanism, usually based on the real-time monitoring data of noise and vibration; Parameter optimization is carried out through genetic algorithm (GA) or particle swarm optimization (PSO) algorithm. The gain, phase and bandwidth parameters of the active damper are optimized by the genetic algorithm to maximize the sound pressure level attenuation and minimize the equipment life loss at the same time; In the above steps, the noise control effect is comprehensively quantified, and the adaptability and efficiency of vibration and noise control can be significantly improved according to the dynamic adjustment mechanism.

[0040] Furthermore, the spectral energy distribution and harmonic distortion rate in the time-frequency domain hybrid features are used for resonance frequency band prediction, and the method further includes: Based on the spectral non-stationary characteristics of the vibration and noise signal, a time-varying transfer function is set, and the time-varying transfer function is used to correct the phase delay parameter of the noise propagation path; By wavelet packet decomposition, the broadband random component and narrowband impulse component in the vibration and noise signal are separated to generate a frequency band priority weight coefficient; Based on the time-varying transfer function, combined with the frequency band priority weight coefficient, the center frequency tracking rate of the adaptive notch filter is adjusted to suppress the secondary reflection of periodic noise.

[0041] Specifically, the spectral non-stationary characteristic refers to the characteristic that the spectral characteristic of the vibration and noise signal changes with time; The time-varying transfer function refers to a transfer function that changes with time and is used to describe the dynamic characteristics of the noise propagation path; The phase delay parameter refers to the phase delay caused by the path difference during the propagation of the noise.

[0042] The spectral non-stationary characteristic is obtained through short-time Fourier transform (STFT) or wavelet transform (WT) analysis. The change in the train running speed will cause the non-stationary characteristic of the noise spectrum; The time-varying transfer function is updated in real time through recursive least squares (RLS) or adaptive filtering algorithm. If the train speed increases from 60 km / h to 100 km / h, the amplitude of the time-varying transfer function increases correspondingly, and then the phase delay increases; The phase delay parameter is calculated by measuring the phase difference of the noise signal. If the phase difference of the noise signal between two sensors is 10°, the phase delay parameter is 10°; By setting the time-varying transfer function, the phase delay parameter of the noise propagation path can be dynamically corrected, improving the accuracy of noise control. By correcting the phase delay parameter, the sound pressure level attenuation is increased from 20 dB to 25 dB.

[0043] Wavelet packet decomposition can decompose a signal into sub-bands of different frequency bands. The broadband random component refers to the random noise component with a wide frequency distribution in the signal, and the narrowband impulse component refers to the impulse noise component with a concentrated frequency and a high amplitude in the signal. The frequency band priority weight coefficient refers to the weight assigned according to the noise contribution degree of different frequency bands, which is used to optimize the control strategy; Wavelet packet decomposition decomposes the vibration noise signal into sub-bands of multiple frequency bands through multi-resolution analysis. Using the Daubechies wavelet to decompose the signal by 3 layers, 8 sub-bands of frequency bands can be obtained; The broadband random component and the narrowband impulse component are separated by analyzing the energy distribution and amplitude characteristics of each sub-band. If the energy distribution of a certain sub-band is relatively uniform and the amplitude is low, it is determined as the broadband random component; If the energy of a certain sub-band is concentrated and the amplitude is high, it is determined as the narrowband impulse component; The frequency band priority weight coefficient is calculated through the energy proportion and noise contribution degree of each sub-band. If the energy proportion of a certain frequency band is 30% and the noise contribution degree is high, a higher weight (such as 0.4) is assigned to it; By separating the broadband random component and the narrowband impulse component and generating the frequency band priority weight coefficient, the control strategy can be optimized more accurately.

[0044] The adaptive notch filter can dynamically adjust the center frequency to suppress the noise in a specific frequency band; The center frequency tracking rate refers to the speed at which the center frequency of the adaptive notch filter is adjusted; The secondary reflection of periodic noise refers to the secondary interference caused by the reflection of periodic noise in the propagation path. Further, the adaptive notch filter dynamically adjusts the center frequency through the least mean square (LMS) or recursive least squares (RLS) algorithm. If the center frequency of the detected periodic noise is 300 Hz, the center frequency of the adaptive notch filter will track and lock at 300 Hz; The center frequency tracking rate is adjusted through the frequency band priority weight coefficient and the time-varying transfer function. If the weight coefficient of a certain frequency band is 0.4 and the time-varying transfer function shows an increase in phase delay, the center frequency tracking rate is increased from 50 Hz / s to 100 Hz / s; By adjusting the center frequency tracking rate, the secondary reflection of periodic noise can be effectively suppressed and the noise control effect can be improved.

[0045] Furthermore, the method includes: Perform clustering analysis on the time-domain correlation of multi-source noise under the positioning of the noise propagation path to identify the coupling relationship between the dominant noise source and the secondary noise source; According to the coupling relationship between the dominant noise source and the secondary noise source, generate a cooperative control logic chain to synchronize the action timings of the active damper and the vibration isolator.

[0046] Specifically, noise propagation path localization refers to determining the specific path of noise propagation through signal processing techniques; time-domain correlation refers to the correlation of noise signals in the time domain, which is used to evaluate the synchronization between different noise sources. Cluster analysis is used to group similar data points. The dominant noise source refers to the noise source that contributes the most to the total noise, and the secondary noise source refers to the indirect noise source caused by the dominant noise source. The coupling relationship refers to the interaction relationship between different noise sources.

[0047] Noise propagation path localization is achieved through direction of arrival (DOA) estimation and time difference of arrival (TDOA) techniques. With a 16-channel microphone array and acceleration sensors, the position of the noise source can be accurately determined. Time-domain correlation is evaluated by calculating the cross-correlation function of different noise source signals. If the cross-correlation coefficient between two noise sources exceeds 0.7, it indicates a significant correlation between the two noise sources. Cluster analysis uses the K-means clustering or hierarchical clustering algorithm to group multi-source noise according to correlation. Specifically, through K-means clustering, the noise sources are divided into three groups: vehicle engine vibration, wheel-rail contact noise, and aerodynamic noise. The coupling relationship between the dominant noise source and the secondary noise source is determined by analyzing the energy contribution and correlation of each noise source. Further, if the energy contribution of wheel-rail contact noise is 45% and its correlation coefficient with aerodynamic noise is 0.8, then wheel-rail contact noise is the dominant noise source and aerodynamic noise is the secondary noise source. Identifying the dominant noise source and the secondary noise source, and through cluster analysis and coupling relationship identification, can more accurately optimize the control strategy.

[0048] The cooperative control logic chain refers to the control logic generated based on the coupling relationship of noise sources, which is used to synchronously adjust multiple control devices. The action timing refers to the time sequence and delay of the control device actions. Specifically, the cooperative control logic chain is generated through a state machine or Petri net to ensure the synchronization of the action timing of the active damper and the vibration isolator. When it is detected that the wheel-rail contact noise is the dominant noise source, the logic chain is generated: first start the active damper to suppress the wheel-rail contact noise, and then start the vibration isolator to block the vibration propagation after 20 ms. The action timing is dynamically adjusted through real-time monitoring and feedback mechanisms. If it is detected that the suppression effect of the active damper decreases, the start delay of the vibration isolator is shortened from 20 ms to 10 ms to ensure their coordinated operation. In the above steps, through the cooperative control logic chain, the noise control effect can be significantly improved.

[0049] Furthermore, the method further includes: When the deviation of the sound pressure distribution uniformity is detected to exceed the deviation threshold, the path parameter recalibration mechanism is triggered; under the path parameter recalibration mechanism, the geographical data and meteorological parameters are fused to establish an optimized model for the noise propagation path, and the compensation value for the multi-path reflection superposition effect is determined; based on the compensation value for the multi-path reflection superposition effect, dynamic optimization is performed on the dynamic threshold fluctuation of the stiffness corresponding to the vibration isolator, and a stiffness adjustment gradient matching the terrain undulation is generated.

[0050] Specifically, the deviation of the sound pressure distribution uniformity refers to the distribution difference of the sound pressure level at different positions, which is used to evaluate the uniformity of the noise field; the deviation threshold is the critical value of the deviation of the sound pressure distribution uniformity that triggers the calibration mechanism; the path parameter recalibration mechanism refers to the process of recalibrating the noise propagation path parameters to optimize the noise control effect. The deviation of the sound pressure distribution uniformity is measured by a multi-point sound pressure level sensor; after the path parameter recalibration mechanism is triggered, the parameters of the noise propagation path will be re-evaluated. If the deviation of the sound pressure distribution uniformity is detected to exceed the threshold, the calibration mechanism will be started, and the noise source position and propagation path will be re-measured, thereby ensuring the accuracy and effectiveness of the noise control strategy.

[0051] The geographical data refers to geographical information such as terrain and building distribution, and the meteorological parameters refer to meteorological conditions such as wind speed, temperature, and humidity; the compensation value for the multi-path reflection superposition effect refers to the value for compensating the noise superposition effect caused by multi-path reflection. The fusion of geographical data and meteorological parameters is achieved through GIS (Geographic Information System) and meteorological sensors; the optimized model for the noise propagation path is established through ray tracing or finite element analysis. Through the ray tracing model, combined with geographical data and meteorological parameters, the noise propagation path is simulated, and the reflection and superposition effects are identified.

[0052] Preferably, geographical data such as terrain and building distribution are obtained from a Geographic Information System (GIS), which provides information on the physical environment for noise propagation. At the same time, meteorological parameters such as wind speed, temperature, and humidity are obtained from meteorological sensors, which affect the propagation speed and direction of sound waves. Through a data fusion algorithm, the geographical data and meteorological parameters are integrated into a unified framework to provide a comprehensive environmental description for the modeling of the noise propagation path; based on the fused data, an optimized model for the noise propagation path is established. This model usually uses ray tracing or finite element analysis methods to simulate the noise propagation path in a complex environment, considering the multi-path reflection and superposition effects, that is, the phenomenon of noise superposition after reflection on different surfaces (such as buildings and the ground). Through simulation, the main reflection paths and superposition areas can be identified.

[0053] Select a suitable noise propagation model, such as a ray tracing model or a finite element analysis model, and connect the noise propagation path optimization model; set the model parameters according to the fused data, including the sound source position, the characteristics of the propagation medium, the characteristics of the reflecting surface, etc. Further, use the established noise propagation path optimization model to obtain the compensation value for the multi-path reflection superposition effect. The compensation value for the multi-path reflection superposition effect is used to adjust the noise control strategy to offset the influence of reflection and superposition effects on noise measurement and control. In the noise propagation path optimization model, simulate the intensity and phase difference of the reflection path to determine the compensation value for the multi-path reflection superposition effect; apply the obtained compensation value for the multi-path reflection superposition effect to the dynamic optimization process of the dynamic threshold fluctuation of the stiffness of the vibration isolator. By adjusting the stiffness of the vibration isolator, more effectively block the noise after reflection and superposition; evaluate the optimized noise control effect through indicators such as the sound pressure level attenuation. If the performance indicators do not meet the expectations, it is necessary to re-adjust the model parameters or compensation value and perform iterative optimization.

[0054] The dynamic threshold fluctuation of the stiffness refers to the fluctuation range of the stiffness of the vibration isolator with time or environment changes; dynamic optimization refers to dynamically adjusting parameters through an optimization algorithm to achieve the best effect; the stiffness adjustment gradient refers to the adjustment rate of the stiffness of the vibration isolator with the terrain undulation changes. The compensation value for the multi-path reflection superposition effect is used to adjust the dynamic threshold of the stiffness of the vibration isolator; adopt an optimization algorithm (such as a genetic algorithm, a particle swarm optimization, etc.) to dynamically adjust the stiffness parameters of the vibration isolator to achieve the best noise control effect. Obtain the terrain undulation information through geographical data; generate a stiffness adjustment gradient matching the terrain according to the terrain undulation and the compensation value; if the terrain undulation is large, increase the stiffness adjustment gradient to improve the isolation effect; dynamically adjust the stiffness of the vibration isolator by real-time monitoring the sound pressure level and vibration data to ensure that the control strategy can adapt to environmental changes; evaluate the optimized noise control effect through indicators such as the sound pressure level attenuation. If it does not meet the expectations, re-adjust the model parameters or compensation value and perform iterative optimization; through multiple iterative adjustments, the vibration isolation effect can be significantly improved.

[0055] In summary, the beneficial effects of the embodiments of the present application are: By adopting the method of collecting vibration and noise signals of rail transit, extracting time-frequency domain hybrid features; separating multiple sources of the vibration and noise signals to generate a noise source contribution matrix; configuring an adaptive control strategy set according to the time-frequency domain hybrid features and the noise source contribution matrix; simultaneously, monitoring the synchronous attenuation rate of noise and the change amount of vibration transfer function to generate a dynamic feedback signal; and dynamically adjusting an active damper based on the adaptive control strategy set and in combination with the dynamic feedback signal, where the active damper is used to suppress vibration and noise. The present application provides an intelligent control method and system for vibration and noise using spectral features, extracts time-frequency domain hybrid features, more accurately identifies and separates overlapping noise sources, comprehensively considers the complex characteristics of vibration and noise, dynamically adjusts the active damper, and thus adapts to the changes of vibration and noise, achieving the technical effects of improving the accuracy and effectiveness of dynamic adjustment.

[0056] Embodiment 2 Based on the same inventive concept as an intelligent control method for vibration and noise using spectral features in the foregoing embodiment, as Figure 2 shown, an embodiment of the present application provides an intelligent control system for vibration and noise using spectral features, where the system includes: A signal acquisition module M100, which is used to collect vibration and noise signals of rail transit and extract time-frequency domain hybrid features.

[0057] A multi-source separation module M200, which is used to perform multi-source separation on the vibration and noise signals to generate a noise source contribution matrix, and the noise source contribution matrix includes a wheel-rail contact noise contribution value, an aerodynamic noise contribution value, and an environmental reflection superposition coefficient.

[0058] An adaptive configuration module M300, which is used to configure an adaptive control strategy set including active damper frequency band adjustment parameters, an optimized layout plan of sound-absorbing materials, a dynamic threshold of the stiffness of a vibration isolator, and a phase compensation strategy according to the time-frequency domain hybrid features and the noise source contribution matrix.

[0059] A monitoring module M400, which is used to, simultaneously, monitor the synchronous attenuation rate of noise and the change amount of vibration transfer function to generate a dynamic feedback signal.

[0060] A dynamic adjustment module M500, which is used to dynamically adjust an active damper based on the adaptive control strategy set and in combination with the dynamic feedback signal, where the active damper is used to suppress vibration and noise.

[0061] Further, the signal acquisition module M100 is used to execute the following method: The time-frequency domain hybrid features include spectral energy distribution, harmonic distortion rate, transient impact peak, and background noise baseline; based on the time-frequency domain hybrid features, combined with direction-of-arrival estimation and coherence analysis, a heat map of the spatial distribution of noise sources is generated; independent component extraction is performed on the overlapping noise sources in the heat map of the spatial distribution of noise sources, and vibration sensor data is correlated to verify the multi-source separation result.

[0062] Further, the dynamic adjustment module M500 is used to execute the following method: Use the spectral energy distribution and harmonic distortion rate in the time-frequency domain hybrid features to predict the resonance frequency band, and determine the target suppression frequency band of the active damper; according to the target suppression frequency band, determine the frequency band adjustment parameters of the active damper.

[0063] Further, the adaptive configuration module M300 is also used to execute the following method: Use the transient impact peak and background noise baseline in the time-frequency domain hybrid features to determine the dynamic adaptation range of the porosity of the sound-absorbing material; according to the dynamic adaptation range of the porosity of the sound-absorbing material, adjust the layout optimization scheme of the sound-absorbing material.

[0064] Further, the adaptive configuration module M300 is also used to execute the following method: Set noise control effect evaluation indicators, and the noise control effect evaluation indicators include sound pressure level attenuation, vibration acceleration reduction rate, and strategy execution energy consumption indicators; at the same time, perform parameter optimization on the adaptive control strategy set, balance the noise reduction effect and equipment life loss, and dynamically adjust the trigger conditions of the phase compensation strategy according to historical control records.

[0065] Further, the dynamic adjustment module M500 is also used to execute the following method: Based on the spectral non-stationary characteristics of the vibration noise signal, set a time-varying transfer function, and the time-varying transfer function is used to correct the phase delay parameter of the noise propagation path; perform wavelet packet decomposition to separate the broadband random component and narrowband impact component in the vibration noise signal, and generate a frequency band priority weight coefficient; based on the time-varying transfer function, combined with the frequency band priority weight coefficient, adjust the center frequency tracking rate of the adaptive notch filter to suppress the secondary reflection of periodic noise.

[0066] Further, the dynamic adjustment module M500 is also used to execute the following method: Perform clustering analysis on the time-domain correlation of multi-source noise under the noise propagation path positioning to identify the coupling relationship between the dominant noise source and the secondary noise source; according to the coupling relationship between the dominant noise source and the secondary noise source, generate a cooperative control logic chain to synchronously adjust the action timings of the active damper and the vibration isolator.

[0067] Furthermore, the dynamic adjustment module M500 is further configured to execute the following method: When it is detected that the deviation of the sound pressure distribution uniformity exceeds the deviation threshold, trigger the path parameter recalibration mechanism; under the path parameter recalibration mechanism, fuse geographical data and meteorological parameters, establish an optimized model for the noise propagation path, and determine the compensation value for the multi-path reflection superposition effect; based on the compensation value for the multi-path reflection superposition effect, perform dynamic optimization on the fluctuation of the corresponding stiffness dynamic threshold of the vibration isolator to generate a stiffness adjustment gradient matching the terrain undulation.

[0068] In summary, any step can be stored as computer instructions or programs in an unrestricted computer memory and can be called and recognized by an unrestricted computer processor, without further limitation here.

[0069] Furthermore, the above technical solution only reflects the preferred technical solution of the technical solution of the embodiments of the present application. Some changes that those skilled in the art may make to some parts thereof all reflect the principles of the novel embodiments of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application.

Claims

1. A vibration noise intelligent control method using spectrum characteristics, characterized in that: The method comprises: Collect vibration and noise signals of rail transit and extract mixed features in time and frequency domains; Performing multi-source separation on the vibration noise signal to generate a noise source contribution matrix, wherein the noise source contribution matrix includes a wheel-rail contact noise contribution value, an aerodynamic noise contribution value, and an environmental reflection superposition coefficient; According to the time-frequency domain hybrid characteristics and the noise source contribution matrix, an adaptive control strategy set including active damper frequency band adjustment parameters, sound absorption material layout optimization scheme, vibration isolator stiffness dynamic threshold and phase compensation strategy is configured; At the same time, the synchronous attenuation rate of noise and the change of vibration transfer function are monitored to generate dynamic feedback signals; Based on the adaptive control strategy set and in combination with the dynamic feedback signal, an active damper is dynamically adjusted, and the active damper is used to suppress vibration noise.

2. A vibration noise intelligent control method using spectrum characteristics as claimed in claim 1, characterized in that: The time-frequency domain mixed features include spectrum energy distribution, harmonic distortion rate, transient impact peak value and background noise baseline; Based on the time-frequency domain hybrid features, combined with direction of arrival estimation and coherence analysis, a noise source spatial distribution heat map is generated; Independent components are extracted from overlapping noise sources in the noise source spatial distribution heat map and correlated with vibration sensor data to verify the multi-source separation result.

3. A vibration noise intelligent control method using spectrum characteristics as claimed in claim 2, characterized in that: Dynamically adjusting an active damper, the method comprising: Using the spectrum energy distribution and harmonic distortion rate in the time-frequency domain hybrid features to predict the resonance frequency band, and determine the target suppression frequency band of the active damper; According to the target suppression frequency band, a frequency band adjustment parameter of the active damper is determined.

4. The vibration noise intelligent control method using spectrum characteristics as claimed in claim 3, characterized in that: Determine the porosity dynamic adaptation range of the sound absorbing material by using the transient impact peak value and the background noise baseline in the time-frequency domain hybrid feature; The layout optimization scheme of the sound absorbing material is adjusted according to the dynamic adaptation range of the porosity of the sound absorbing material.

5. The vibration noise intelligent control method using spectrum characteristics as claimed in claim 4, characterized in that: Setting noise control effect evaluation indicators, which include sound pressure level attenuation, vibration acceleration reduction rate and strategy execution energy consumption indicators; At the same time, the parameters of the adaptive control strategy set are optimized to balance the noise reduction effect and the equipment life loss, and the triggering conditions of the phase compensation strategy are dynamically adjusted according to the historical control records.

6. The vibration noise intelligent control method using spectrum characteristics as claimed in claim 3, characterized in that: The resonance frequency band is predicted using the spectrum energy distribution and harmonic distortion rate in the time-frequency domain hybrid feature, and the method further includes: Based on the non-stationary characteristics of the frequency spectrum of the vibration noise signal, a time-varying transfer function is set, and the time-varying transfer function is used to correct the phase delay parameter of the noise propagation path; Using wavelet packet decomposition to separate broadband random components and narrowband impact components in the vibration noise signal, and generate frequency band priority weight coefficients; Based on the time-varying transfer function and in combination with the frequency band priority weight coefficient, the center frequency tracking rate of the adaptive notch filter is adjusted to suppress the secondary reflection of the periodic noise.

7. The vibration noise intelligent control method using spectrum characteristics as claimed in claim 6, characterized in that: Performing cluster analysis on the time domain correlation of multi-source noise under the noise propagation path positioning to identify the coupling relationship between the dominant noise source and the secondary noise source; According to the coupling relationship between the dominant noise source and the secondary noise source, a cooperative control logic chain is generated to synchronously adjust the action timing of the active damper and the vibration isolator.

8. The vibration noise intelligent control method using spectrum characteristics as claimed in claim 7, characterized in that: When it is detected that the sound pressure distribution uniformity deviation exceeds the deviation threshold, the path parameter recalibration mechanism is triggered; Under the path parameter recalibration mechanism, geographic data and meteorological parameters are integrated to establish a noise propagation path optimization model and determine the multipath reflection superposition effect compensation value; Based on the multi-path reflection superposition effect compensation value, the stiffness dynamic threshold fluctuation corresponding to the vibration isolator is dynamically optimized to generate a stiffness adjustment gradient that matches the terrain undulation.

9. An intelligent vibration noise control system using spectrum characteristics, characterized in that: A method for intelligently controlling vibration noise using spectral characteristics according to any one of claims 1 to 8, the system comprising: A signal acquisition module, which is used to collect vibration and noise signals of rail transit and extract mixed features in time and frequency domains; A multi-source separation module, wherein the multi-source separation module is used to perform multi-source separation on the vibration noise signal to generate a noise source contribution matrix, wherein the noise source contribution matrix includes a wheel-rail contact noise contribution value, an aerodynamic noise contribution value, and an environmental reflection superposition coefficient; An adaptive configuration module, the adaptive configuration module is used to configure an adaptive control strategy set including active damper frequency band adjustment parameters, sound absorption material layout optimization scheme, vibration isolator stiffness dynamic threshold and phase compensation strategy according to the time-frequency domain hybrid characteristics and the noise source contribution matrix; A monitoring module, the monitoring module is used to simultaneously monitor the noise synchronous attenuation rate and the vibration transfer function change amount to generate a dynamic feedback signal; A dynamic adjustment module is used to dynamically adjust an active damper based on the adaptive control strategy set and in combination with the dynamic feedback signal, wherein the active damper is used to suppress vibration noise.

Images