Method and system for visualizing noise components of a permanent magnet motor based on a microphone array

Abstract

The application relates to a microphone array-based permanent magnet motor noise component visualization method and system, which combines equi-angle resampling, focused beamforming, order analysis, principal component analysis and cluster analysis on the basis of beamforming acoustic imaging to give the noise distribution of each part of the permanent magnet motor, the essence of which is a point set of cluster analysis, each point containing sound pressure level information and frequency information, which is subjected to image fusion and area positioning with a motor side sectional view to give a sound pressure level diagram of each large-noise structure in the motor, and contribution degree analysis is carried out through the image, the method is also applicable to the constant speed condition.Compared with the prior art, the application directly shows the change trend of the noise components of each structure of the permanent magnet motor, realizes the correspondence of each structure of the motor to the order analysis result, is convenient for subsequent optimization, has high order grading accuracy and strong anti-interference capability.

Description

Technical Field

[0001] This invention relates to the field of permanent magnet motor vibration and noise testing technology, and in particular to a method and system for visualizing the noise components of a permanent magnet motor based on a microphone array. Background Technology

[0002] Due to the rapid development of new energy vehicles in recent years, the sound quality and noise source issues of automotive drive motors have received widespread attention. Beamforming technology is now a mature and commonly used acoustic method for noise source localization. By measuring the microphone array, beamforming the multi-channel signals can image the sound signals and visually identify the noise source. Permanent magnet motors often operate under variable speed conditions, such as acceleration and deceleration during startup, and fluctuations in actual speed during constant speed operation. Under variable speed conditions, non-steady-state vibrations are excited, and the radiated sound field will also be non-steady-state. The acoustic characteristics of permanent magnet motors change over time. Traditional frequency domain beamforming methods are only applicable to stationary sound sources and not to non-steady-state signals. Therefore, using time domain beamforming methods for non-stationary noise source localization in permanent magnet motors seems more appropriate. A non-stationary noise source localization method based on Volk-Kalman filters and time domain beamforming was proposed, and the spatial acoustic map of the permanent magnet motor was reconstructed using characteristic orders. The effectiveness of the proposed method was verified through motor noise source localization experiments. This method can obtain the localization results of non-stationary sound sources with order information. However, this method only locates the approximate position of the motor, resulting in low spatial resolution. It does not accurately locate the various internal structures of the motor, and it only locates a certain order, failing to reflect that different orders of noise from the motor correspond to different internal structures.

[0003] Chinese Patent Publication No. CN104266747A discloses a fault diagnosis method based on vibration signal order analysis, comprising the following steps: A. obtaining the time spectrum of the vibration signal of the wind turbine; B. resampling each order component in the time spectrum at equal angles; C. performing order spectrum analysis on the resampled signal to identify the fault location and issue an early warning.

[0004] The aforementioned application implements fault diagnosis based on vibration signals for wind turbines. However, due to size limitations, fault diagnosis based on vibration signals is not applicable to small permanent magnet motors. In addition, single-channel analysis suffers from drawbacks such as large interference and weak directivity.

[0005] At present, single-channel microphones are difficult to solve the problem of sound source localization, and multi-channel microphone arrays can only roughly locate the sound source. There is an urgent need for a technology that can locate the sound source of each structure inside the motor under variable speed conditions, and can visualize and analyze the noise components and order and calculate the noise contribution through imaging. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art by providing a method and system for visualizing the noise components of a permanent magnet motor based on a microphone array, so as to intuitively show the changing trends of the noise components of each structure of the permanent magnet motor.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] One aspect of the present invention provides a method for visualizing the noise components of a permanent magnet motor based on a microphone array, comprising the following steps:

[0009] S1: Acquire multi-channel noise data of permanent magnet motor collected by microphone array under variable speed operation, and perform equal-angle resampling on the data of each channel. Perform bandpass filtering on the data after equal-angle resampling and perform time-domain beamforming imaging to obtain multi-frame sound pressure level map.

[0010] S2, obtain the focal point coordinates based on the multi-frame sound pressure level map, perform focusing beamforming based on the current focal point coordinates to obtain the main order of the current focal point, remove the main order in the data after equal angle resampling by band-stop filtering and re-perform time-domain beamforming imaging to obtain a new multi-frame sound pressure level map;

[0011] S3. Repeat step S2 until there is a frame of sound pressure level map whose amplitude is less than the background noise amplitude, and obtain a scatter plot formed by superimposing the sound pressure level maps corresponding to each major order.

[0012] S4. For the scatter plot, cluster analysis is used to obtain the group center and group weight of each structure inside the permanent magnet motor, and the contribution of each structure to the total noise is calculated.

[0013] S5: Obtain a side image of the motor after calibration and with the permanent magnet motor structure division layer superimposed. Based on the contribution of each structure and the side image of the motor, output a visualized signal of permanent magnet motor speed-structure contribution under variable speed conditions.

[0014] As a preferred technical solution, in step S2, the calculation of the focal point coordinates includes:

[0015] S21, after summing and averaging the multiple frames of sound pressure level images after beamforming, the coordinates of the maximum sound pressure level point on the image are used as the coordinates of the focal point.

[0016] As a preferred technical solution, the calculation of the main order in S2 includes the following:

[0017] S22, perform order analysis on the acoustic signal at the focal point to obtain the order information of the focal point noise.

[0018] S23, taking the order of the structure where the focal point is located inside the permanent magnet motor as the principal component and the orders of other structures as minor components, the principal order of the focal point is obtained by performing principal component analysis on the order information.

[0019] As a preferred technical solution, the calibration process in S5 includes:

[0020] S51 calibrates the side image of the motor using markers pre-set on the side shell of the permanent magnet motor.

[0021] As a preferred technical solution, the cluster analysis process in S4 includes:

[0022] S41, using the sound pressure level of the focal point as a clustering variable, calculates the similarity between each focal point;

[0023] S42, grouping based on preset clustering algorithm and clustering parameters to obtain multiple groups, and calculating group centers and group weights.

[0024] As a preferred technical solution, the contribution is the ratio of the sound pressure level of each group within the structural area to the total sound pressure level of the permanent magnet motor calculated based on the group center and the number of people, or the ratio of the sound pressure level of each group within the structural area to the sum of the sound pressure levels of each group.

[0025] As a preferred technical solution, S5 includes:

[0026] S52 transforms the scale from the angular domain to the time domain through inverse transformation. Based on the time-speed information of the permanent magnet motor, it obtains the speed-structure contribution data of the permanent magnet motor under variable speed conditions, that is, a scatter plot of the changes in the contribution of each structure based on the speed change.

[0027] S53 uses nonlinear regression to fit the scatter plot in S52 into a curve, obtaining a visualized signal reflecting the changes in various structural noise components under variable speed conditions.

[0028] As a preferred technical solution, time-domain beamforming is achieved using any one of the algorithms of energy correction, mean square error optimization, and phase offset compensation, and focused beamforming is achieved using any one of the algorithms of GSC-LMS, GSC-RLS, and GSC-ICA.

[0029] As a preferred technical solution, in step S2, the number of frames in the sound pressure level map formed by the time-domain beamforming is determined by the following formula:

[0030] S Frames =l / f s ×n Frames

[0031] Among them, S FramesLet n be the total number of frames, l be the total number of signal points, f be the sampling frequency, and n be the total number of frames. Frames The number of frames generated per second.

[0032] Another aspect of the present invention provides a visualization system for the noise components of a permanent magnet motor based on a microphone array, comprising:

[0033] The permanent magnet motor under test;

[0034] A microphone array is disposed on the side of the permanent magnet motor under test to acquire multi-channel noise data of the permanent magnet motor under variable speed conditions. A camera for capturing images of the side of the motor is disposed at the center of the microphone array, and the camera is at the same height as the permanent magnet motor.

[0035] A speed sensor is used to measure the speed of a permanent magnet motor.

[0036] The data processing unit includes a memory and an actuator, wherein the memory stores instructions for executing the above-described method for visualizing the noise components of a permanent magnet motor based on a microphone array.

[0037] A visualization device for displaying the visualization signal output by the data processing unit.

[0038] Compared with the prior art, the present invention has the following advantages:

[0039] (1) Visually displaying the changing trends of noise components in various structures of a permanent magnet motor: This invention first resamples and beamforms multi-channel noise data of a permanent magnet motor collected by a microphone array under variable speed conditions. After order analysis, the main order of the noise source is obtained. Then, the main order is divided into various structural regions of the permanent magnet motor, and the noise contribution is calculated using images. Finally, a side image of the motor with a permanent magnet motor structure division layer is superimposed to form a visual signal that shows the changing trends of noise components in various structures of the permanent magnet motor. Since focused beamforming is used to focus on the point with the maximum average sound pressure level during the order search process, the main order information is enhanced, and the interference caused by sidelobe sound signals is effectively suppressed. Compared with single-channel order analysis, it is more directional and can correspond one-to-one with the internal structure of the motor, improving the intuitiveness of the noise and helping to detect faults and monitor operating conditions.

[0040] (2) To make the motor structure more consistent with the order analysis results, which is convenient for subsequent optimization: The main orders of the noise sources of each part of the permanent magnet motor were imaged and cluster analysis was used to divide the main orders into the structural regions of the permanent magnet motor. The noise contribution was calculated by image, thus avoiding the situation that the theoretical calculation might not correspond to the order. While seeing the internal structural noise, the order was also retained, which provides a clearer idea for optimizing the internal structure of the permanent magnet motor for the purpose of vibration reduction and noise reduction, such as optimizing the parameters of the corresponding structure according to the level of contribution and which orders should be avoided.

[0041] (3) High accuracy of order classification: This invention uses equal-angle resampling technology in the data processing process, thus eliminating the influence of variable speed operation on the noise of permanent magnet motor. This method can be used in both constant speed and variable speed operation. By using cluster analysis, the original order points are integrated and estimated into groups. The attenuation of sound energy in each structural area inside the motor due to calculation is reduced, thus improving the accuracy of contribution analysis.

[0042] (4) Strong anti-interference capability: The array of this application is at the same height as the center point of the permanent magnet motor. It uses beamforming technology to enhance the target signal and reduce interference, so that the beam with the maximum amplitude (i.e. the main lobe) can be directly facing the target signal position of the rail vehicle. Therefore, after finding the maximum value point of the average sound pressure level, the array can focus the sound signal at that point and effectively suppress the interference caused by the side lobe sound signal. Attached Figure Description

[0043] Figure 1 This is a flowchart of a method for visualizing the noise components of a permanent magnet motor based on a microphone array, as described in the embodiment.

[0044] Figure 2 This is a schematic diagram showing the overlay of cluster analysis results and structural partitioning diagram;

[0045] Figure 3 This is a schematic diagram of a permanent magnet motor noise component visualization system based on a microphone array, as described in the embodiment.

[0046] Among them, 1-casing, 2-stator, 3-stator winding, 4-rotor, 5-bearing, 6-shaft, 7-camera, 8-microphone array. Detailed Implementation

[0047] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0048] Example 1

[0049] See Figure 1 To address the aforementioned problems, this embodiment provides a method for visualizing the noise components of a permanent magnet motor based on a microphone array, comprising the following steps:

[0050] Step S1: Use a microphone array to collect the noise emitted by the permanent magnet motor under different operating conditions. The data acquisition and storage system sends the sound signals collected by the microphone array to the host computer for data display and analysis through communication equipment, and then stores the data and sends it to the analysis program.

[0051] Step S2 involves resampling the data of each channel based on equal-angle resampling to eliminate the influence of variable speed operating conditions. The time domain scale mentioned later refers to the transformed angle domain.

[0052] Step S3: Bandpass filtering is performed on the channel data after equal-angle resampling. Assuming that the noise sources of each structure inside the permanent magnet motor are point sources and that each source is incoherent, time-domain beamforming acoustic imaging algorithms such as energy correction, mean square error optimization, and phase offset compensation can be used to beamform the channel data after equal-angle resampling onto the focusing grid.

[0053] Step S4: Since the principle of time-domain beamforming is to divide the signal into segments, delay and sum them, each segment becomes an image, which is a frame. After summing each frame image after beamforming, divide by the total number of frames to calculate the average, and find the coordinates of the maximum sound pressure level point on the image through a function, which is used as the coordinates of the focal point.

[0054] Specifically, the total number of frames is determined by the set sampling frequency, total signal length, and frames per second:

[0055] S Frames =l / f s ×n Frames

[0056] Among them, S Frames Let n be the total number of frames, l be the total number of signal points, f be the sampling frequency, and n be the total number of frames. Frames The number of frames generated per second.

[0057] Step S5: Based on focusing beamforming algorithms such as GSC-LMS, GSC-RLS, and GSC-ICA, the array signal is focused onto the focal point coordinates to form a focused beam, thereby obtaining the sound pressure signal in the time domain of the focal point, enhancing the target signal, and effectively suppressing the interference caused by the sidelobe sound signal.

[0058] Step S6: Focusing beamforming changes the signal amplitude but not the signal order information. Here, only the order information of the focal point is extracted. The order analysis of the acoustic signal at the focal point is performed to obtain the order information of the noise at that point. Inside the permanent magnet motor, the order of the structure where the focal point is located is the main component, and the order of other structures is the secondary component. The main order of the focal point can be obtained by principal component analysis of the order information.

[0059] Step S7: Bandpass filtering is performed on the signals after equal-angle resampling of each channel to filter out the required main order, and beamforming imaging is performed again.

[0060] Step S8: Band-stop filtering is performed on the signals after equal-angle resampling of each channel. After filtering out the main order obtained in step S7, steps S4-S7 are repeated. If the amplitude of the beamforming sound pressure level map in a certain frame is less than the background noise amplitude, the process jumps to step S9. The maximum amplitude in all frames of the sound pressure level map is taken as the comparison with the background noise amplitude. The background noise amplitude is pre-designed.

[0061] Step S9: Adjust the dynamic range of the sound pressure level map so that the maximum value of the sound pressure level map at each order approaches a point, which represents the location of the sound source. Superimpose the maps of each major order in each frame after the time-domain beamforming imaging to obtain a scatter plot of each major order in the focused area of ​​each frame, with amplitude and order information attached.

[0062] Step S10: Overlay the side sectional view of the permanent magnet motor design with the image captured by the central camera of the microphone array to divide the image into various structural regions inside the permanent magnet motor.

[0063] Specifically, the calibration of the permanent magnet motor design side section drawing is achieved through the following methods:

[0064] Make several calibration marks on the side shell of the permanent magnet motor, align the side view of the motor taken by the array center camera with the side view of the permanent magnet motor during design to ensure overlap, and then replace the side view with the side sectional view. Based on the side sectional view, manually divide the various structural areas inside the motor.

[0065] Step S11: Perform clustering algorithm analysis on the scatter plot of each frame composed of each major order to obtain the group center and group weight of each structure inside the motor in each frame, and display them on the structure partitioning map. The clustering analysis result of each frame is represented as the noise source and noise component result of each segment in the angular domain.

[0066] Specifically, cluster analysis is implemented in the following ways:

[0067] (1) Determine the clustering variables and distance measurement methods. Select the sound pressure level of the focal point as the clustering variable, and use algorithms such as Euclidean distance, Manhattan distance or correlation coefficient to calculate the similarity between each focal point.

[0068] (2) Select an appropriate clustering algorithm based on the actual number of focal points. Choose an appropriate clustering algorithm, such as K-Means or hierarchical clustering, based on the size and complexity of the dataset and the application objectives.

[0069] (3) Set clustering parameters, including the number of clusters, stopping criteria, and initial conditions. The number of clusters refers to the number of categories defined, the stopping criteria refers to the convergence condition for stopping the clustering iteration, and the initial conditions are the starting values ​​set when the clustering algorithm starts running.

[0070] (4) Run the clustering algorithm. Group all samples in the dataset according to the clustering algorithm.

[0071] (5) Evaluate and interpret the clustering results. By analyzing the size, characteristics, and trends of each cluster, interpret the clustering results to further conduct data mining and business decision-making.

[0072] Step S12: Based on the center of each group within the region, and using the group weight as a weighting coefficient, analyze the contribution of each structure to the total noise. The contribution of each structure is expressed as the ratio of the sound pressure level of each group within each region to the total sound pressure level of the permanent magnet motor calculated according to the group center and the group size. Alternatively, the proportion of noise source components can be calculated and expressed as the ratio of the sound pressure level of each group within each region to the sum of the sound pressure levels of each group.

[0073] Step S13: After calculating the structural contribution of each image obtained by superimposing each order in each frame, calculate the contribution according to the angle range ω of that frame. i ~ω i+1 ,Pick Using the horizontal axis as the scale transformation of equal-angle resampling in step S3, the angle domain is reversed back to the time domain. Then, based on the rotational speed information obtained from the rotational speed sensor, each moment corresponds to a rotational speed. The horizontal axis of the points of each structural contribution is changed from time to rotational speed, thus obtaining a scatter plot of the changes in the contribution of each structure based on the rotational speed.

[0074] In step S14, the scatter plot can be fitted into a curve using nonlinear regression to represent the changing trend of each structural noise component under variable speed conditions.

[0075] See Figure 2This diagram illustrates the overlay of cluster analysis results and structural partitioning diagrams. It represents the desired outcome of the method, using image-based contribution calculations for a more intuitive understanding and eliminating the need for tedious traditional order calculations, thus visualizing order information. The structural regions in the diagram are defined based on the side sectional view of the motor design. Before overlaying the structural regions with the cluster analysis results, the side view of the motor captured by the array's central camera must be aligned with the side sectional view of the motor design to ensure accurate positioning.

[0076] This invention performs cluster analysis on each frame of image obtained by time-domain beamforming to obtain the contribution at each moment, and then fits the contribution into a curve to reflect the change of noise components under variable speed conditions. Therefore, it not only retains the identification results of noise sources and noise components inside the permanent magnet motor at each moment, but also reflects the trend of noise contribution of each structure inside the permanent magnet motor changing with rotational speed.

[0077] Example 2

[0078] See Figure 3 This embodiment provides a permanent magnet motor noise component visualization system based on a microphone array, including a permanent magnet motor under test, a microphone array 8, a camera 7, a speed sensor, a data processing unit, and a visualization device.

[0079] The microphone array includes a housing 1, a stator 2, a stator winding 3, a rotor 4, a bearing 5, and a shaft 6.

[0080] The microphone array 8 is set on the side of the permanent magnet motor under test to acquire multi-channel noise data of the permanent magnet motor under variable speed conditions. The center of the microphone array 8 is set with a camera 7 for taking pictures of the side of the motor. The camera 7 is at the same height as the permanent magnet motor.

[0081] A speed sensor is used to measure the speed of a permanent magnet motor.

[0082] The data processing unit includes a memory and an actuator. The memory stores instructions for executing the method for visualizing the noise components of a permanent magnet motor based on a microphone array as described in Example 1.

[0083] A visualization device used to display visual signals output by a data processing unit.

[0084] See Figure 3Beamforming is a method that uses a microphone array to obtain highly directional beam characteristics. It involves arranging a group of microphones at different locations in space to form a microphone array to receive sound signals. After appropriate delay and summation, information such as the sound source location and intensity is extracted. The array's directivity allows for maximum gain of the signal in a specific direction; however, the array's focusing direction is almost independent of its azimuth. Therefore, different delay methods can focus in different directions, and all of these directions may contain useful signals.

[0085] This application designs an array at the same height as the center point of the permanent magnet motor. It utilizes GSC beamforming technology to enhance the target signal and reduce interference, ensuring that the beam with the maximum amplitude (i.e., the main lobe) is directly aligned with the target signal position of the rail vehicle. Therefore, after finding the point of maximum average sound pressure level, the array can capture the acoustic signal at that focal point and effectively suppress interference from sidelobe acoustic signals. Simultaneously, the array is also capable of adjusting and suppressing signals in non-monitoring directions, improving the signal-to-noise ratio and overcoming strong interference sources. The side view of the permanent magnet motor acquired by camera 7 at the center of the array is overlaid with the motor's internal structure diagram and cluster analysis results to solve the positioning problem.

[0086] This invention, based on beamforming acoustic imaging, combines equal-angle resampling, focused beamforming, order analysis, principal component analysis, and cluster analysis to provide noise distribution information for various parts of a permanent magnet motor. Essentially, it is a point set for cluster analysis, where each point contains sound pressure level and frequency information. By fusing these points with a side profile of the motor and locating the regions, it provides sound pressure level maps of various noisy structures inside the motor. Contribution analysis is then performed on the images. This method is also applicable to constant speed operating conditions.

[0087] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for visualizing the noise components of a permanent magnet motor based on a microphone array, characterized in that, Includes the following steps: S1: Acquire multi-channel noise data of permanent magnet motor collected by microphone array under variable speed operation, and perform equal-angle resampling on the data of each channel. Perform bandpass filtering on the data after equal-angle resampling and perform time-domain beamforming imaging to obtain multi-frame sound pressure level map. S2, obtain the focal point coordinates based on the multi-frame sound pressure level map, perform focusing beamforming based on the current focal point coordinates to obtain the main order of the current focal point, remove the main order in the data after equal angle resampling by band-stop filtering and re-perform time-domain beamforming imaging to obtain a new multi-frame sound pressure level map; S3. Repeat step S2 until there is a frame of sound pressure level map whose amplitude is less than the background noise amplitude, and obtain a scatter plot formed by superimposing the sound pressure level maps corresponding to each major order. S4. For the scatter plot, cluster analysis is used to obtain the group center and group weight of each structure inside the permanent magnet motor, and the contribution of each structure to the total noise is calculated. S5: Obtain a side image of the motor after calibration and with the permanent magnet motor structure division layer superimposed. Based on the contribution of each structure and the side image of the motor, output a visualized signal of permanent magnet motor speed-structure contribution under variable speed conditions.

2. The method for visualizing the noise components of a permanent magnet motor based on a microphone array according to claim 1, characterized in that, In S2, the calculation of the focal point coordinates includes: S21, after summing and averaging the multiple frames of sound pressure level images after beamforming, the coordinates of the maximum sound pressure level point on the image are used as the coordinates of the focal point.

3. The method for visualizing the noise components of a permanent magnet motor based on a microphone array according to claim 1, characterized in that, In S2, the calculation of the main order includes the following: S22, perform order analysis on the acoustic signal at the focal point to obtain the order information of the focal point noise. S23, taking the order of the structure where the focal point is located inside the permanent magnet motor as the principal component and the orders of other structures as minor components, the principal order of the focal point is obtained by performing principal component analysis on the order information.

4. The method for visualizing the noise components of a permanent magnet motor based on a microphone array according to claim 1, characterized in that, In S5, the calibration process includes: S51 calibrates the side image of the motor using markers pre-set on the side shell of the permanent magnet motor.

5. The method for visualizing the noise components of a permanent magnet motor based on a microphone array according to claim 1, characterized in that, In S4, the cluster analysis process includes: S41, using the sound pressure level of the focal point as a clustering variable, calculates the similarity between each focal point; S42, grouping based on preset clustering algorithm and clustering parameters to obtain multiple groups, and calculating group centers and group weights.

6. The method for visualizing the noise components of a permanent magnet motor based on a microphone array according to claim 5, characterized in that, The contribution is the ratio of the sound pressure level of each group within the structural region to the total sound pressure level of the permanent magnet motor calculated based on the group center and group weight, or the ratio of the sound pressure level of each group within the structural region to the sum of the sound pressure levels of each group.

7. The method for visualizing the noise components of a permanent magnet motor based on a microphone array according to claim 1, characterized in that, The S5 includes: S52 transforms the scale from the angular domain to the time domain through inverse transformation. Based on the time-speed information of the permanent magnet motor, it obtains the speed-structure contribution data of the permanent magnet motor under variable speed conditions, that is, a scatter plot of the changes in the contribution of each structure based on the speed change. S53 uses nonlinear regression to fit the scatter plot in S52 into a curve, obtaining a visualized signal reflecting the changes in various structural noise components under variable speed conditions.

8. The method for visualizing the noise components of a permanent magnet motor based on a microphone array according to claim 1, characterized in that, In S2, time-domain beamforming is achieved using any one of the following algorithms: energy correction, mean square error optimization, and phase offset compensation. Focused beamforming is achieved using any one of the following algorithms: GSC-LMS, GSC-RLS, and GSC-ICA.

9. The method for visualizing the noise components of a permanent magnet motor based on a microphone array according to claim 1, characterized in that, In S2, the number of frames in the sound pressure level map formed by the time-domain beamforming is determined by the following formula: S Frames =l / f s ×n Frames Among them, S Frames Let n be the total number of frames, l be the total number of signal points, f be the sampling frequency, and n be the total number of frames. Frames The number of frames generated per second.

10. A system for visualizing the noise components of a permanent magnet motor based on a microphone array, characterized in that, include: The permanent magnet motor under test; A microphone array is disposed on the side of the permanent magnet motor under test to acquire multi-channel noise data of the permanent magnet motor under variable speed conditions. A camera for capturing images of the side of the motor is disposed at the center of the microphone array, and the camera is at the same height as the permanent magnet motor. A speed sensor is used to measure the speed of a permanent magnet motor. The data processing unit includes a memory and an actuator, wherein the memory stores instructions for executing the method for visualizing the noise components of a permanent magnet motor based on a microphone array as described in any one of claims 1-9; A visualization device for displaying the visualization signal output by the data processing unit.

Images