Noise monitoring method for charging posts, noise monitoring system, and readable medium

Abstract

To provide a method and system for monitoring the noise of a charge post, and a readable medium.SOLUTION: A method for monitoring noise of a charge post including use, for a sound noise monitoring system, of a 3D ball array consisting of a plurality of microphones, comprises the steps of: obtaining respectively detected noise data including sound pressure signals measured by the plurality of microphones; obtaining an initial noise signal; obtaining a noise signal having a target analysis frequency; uniformly classifying a scanning area into some discrete scanning grid points; obtaining a vector constituted of beam formation energy estimation values at all scanning grid points; creating an energy transfer model; and calculating a high resolution sound source energy distribution of a charge post to be measured in the energy transfer model.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] This application relates to the technical field of noise measurement, and more particularly to a noise monitoring method, noise monitoring system, and readable medium for charging posts. [Background technology]

[0002] While the current charging stations vary in technological level, the demands of modern urban life dictate that people are striving to charge electric vehicles as quickly as possible. As a result, DC charging stations have a broad future for development, but this has also brought design issues for DC charging stations to the forefront.

[0003] Currently, many DC charging stations have noise control issues and, at the same time, their inappropriate placement in public spaces causes disturbances to residents. According to measurements from related experiments, the noise level in the vicinity of a DC fast charging station when it is operating is about 70 dB, and the noise level in residential areas is about 55 dB, sounding like a large vacuum cleaner used in supermarkets. In many situations, people are generally inside their cars or near the vehicles while electric vehicles are charging, so the noise generated during the operation of DC charging stations is likely to disturb those inside or near the vehicles. Therefore, measuring and positioning noise surround conditions is of important reference value for the design and placement of fast charging stations. The noise from DC fast charging stations mainly originates from the cooling fan system and transformer.

[0004] Currently, the most common measurement methods are continuous scanning and discrete-point measurement, but both methods have their drawbacks. Continuous scanning requires very strict control of the measurement device during measurement, and discrete-point measurement cannot guarantee the accuracy of results because the number of data samples collected is small. A common method is to use a sound level meter to perform a reasonable inspection of the equipment, but this method is only a rough inspection of high-frequency noise and does not allow for effective monitoring of low-frequency noise, making it difficult to effectively orient and position the noise formation. The majority of the noise generated by DC rapid charging posts is low-frequency, which is why there are few effective measurement methods for DC rapid charging posts. [Overview of the Initiative] [Problems that the invention aims to solve]

[0005] Based on this, the challenge is to provide a noise monitoring method, noise monitoring system, and readable medium for charging posts in order to solve the problem that conventional measurement methods can only perform rough inspections of high-frequency noise, and that effective monitoring of low-frequency noise is almost impossible, and that effective orientation and positioning of noise formation is not possible. [Means for solving the problem]

[0006] This invention includes a 3D ball array consisting of multiple microphones used in a noise monitoring system. The charging post under test Operating state (on state) In this process, noise detection data including sound pressure signals measured by multiple microphones at several measurement positions surrounding the circumferential direction of the charging post under test is acquired. The process involves obtaining an initial noise signal by applying noise reduction processing to the acquired noise detection data, and The process involves filtering the initial noise signal to obtain the noise signal at the target analysis frequency, and The measurement surface directly in front of the 3D ball array (the measurement surface facing the 3D ball array from the front) is defined as the sound source plane.The process involves selecting a scanning area from the sound source plane and uniformly dividing the scanning area into several discrete scanning grid points, The aforementioned objective analysis A step in which a vector is obtained consisting of estimated beamforming energy values ​​for all scanning grid points using a frequency noise signal, The steps include creating an energy transfer model using a vector and energy transfer matrix consisting of beamforming energy estimates for at least every scanning grid point, and The present invention provides a noise monitoring method for a charging post, which includes the step of calculating a high-resolution sound source energy distribution of the charging post under measurement in the energy transfer model using a Bayesian estimation method.

[0007] This application further states that At the measurement location, a 3D ball array is used to collect noise data from the charging post under test, Coupled memory and processing unit (Processor, same applies below) The present invention provides a noise monitoring system that includes an information processing device that connects signals to the 3D ball array, wherein the memory is used for storing program data, and the processing device is used for executing the program data to realize the noise monitoring method for the charging post described above.

[0008] This application further provides a computer-readable medium on which a computer program is stored, and which is executed by a processing unit to realize the noise monitoring method for a charging post described above. [Effects of the Invention]

[0009] This application relates to a method for monitoring the noise of a charging post, a noise monitoring system, and a mobile medium. Among them, the method for monitoring the noise of a charging post uses a 3D ball array to perform asynchronous measurements and collect noise detection data at several measurement positions. At the same time, through noise removal processing and filtering processing on the noise detection data, a noise signal at the target analysis frequency that can represent the noise characteristics of the measured charging post is obtained. Furthermore, a vector composed of the beamforming energy estimate values of all scanning grid points is obtained. As a result, the inverse analysis of the visible sound source can be quickly realized, the spatial resolution of the sound source positioning can be improved, and its robustness and intuitiveness are good. In addition, by using the Bayesian estimation method to explain the variables in the sound field from the perspective of statistical optimization, the distribution of the sound source energy at the scanning grid points is calculated, and the fault position in the measured charging post is accurately positioned. As a result, the resolution of the imaging result is high, the anti-noise ability of the algorithm is strong, and the self-adaptive ability is strong. Therefore, the beamforming sound source positioning method based on the Bayesian estimation method can quickly position the fault of the DC fast charging post by quickly orienting the low-frequency noise generated from the measured charging post.

Brief Description of the Drawings

[0010] [Figure 1] It is a schematic flowchart of a method for monitoring the noise of a charging post provided by an embodiment of this application. [Figure 2] In the method for monitoring the noise of a charging post provided by an embodiment of this application, it is a schematic diagram showing the spatial distribution of the measurement positions for the measured charging post. [Figure 3] In the method for monitoring the noise of a charging post provided by an embodiment of this application, it is a schematic diagram showing the division of the scanning grid points. [Figure 4] It is a spatial distribution diagram of the collected noise signal and the average noise when the 3D ball array is located at the direct front measurement point of the measured charging post. [Figure 5]It is the spatial distribution diagram of the collected noise signal and the average noise when the 3D ball array is located at the measurement point 30° clockwise from the front of the measured charging post. [Figure 6] It is the spatial distribution diagram of the collected noise signal and the average noise when the 3D ball array is located at the measurement point 60° clockwise from the front of the measured charging post. [Figure 7] It is the spatial distribution diagram of the collected noise signal and the average noise when the 3D ball array is located at the measurement point 90° clockwise from the front of the measured charging post. [Figure 8] It is the spatial distribution diagram of the collected noise signal and the average noise when the 3D ball array is located at the measurement point 120° clockwise from the front of the measured charging post. [Figure 9] It is the spatial distribution diagram of the collected noise signal and the average noise when the 3D ball array is located at the measurement point 150° clockwise from the front of the measured charging post. [Figure 10] It is the spatial distribution diagram of the collected noise signal and the average noise when the 3D ball array is located at the measurement point 180° clockwise from the front of the measured charging post. [Figure 11] It is the spatial distribution diagram of the collected noise signal and the average noise when the 3D ball array is located at the measurement point 210° clockwise from the front of the measured charging post. [Figure 12] It is the spatial distribution diagram of the collected noise signal and the average noise when the 3D ball array is located at the measurement point 240° clockwise from the front of the measured charging post. [Figure 13] It is the spatial distribution diagram of the collected noise signal and the average noise when the 3D ball array is located at the measurement point 270° clockwise from the front of the measured charging post. [Figure 14] It is the spatial distribution diagram of the collected noise signal and the average noise when the 3D ball array is located at the measurement point 300° clockwise from the front of the measured charging post. [Figure 15]This is a spatial distribution map of the noise signal and average noise collected when the 3D ball array is positioned at a measurement point 330° clockwise directly in front of the charging post under test. [Figure 16] This is the spatial distribution of noise comprising any measurement position provided by one embodiment of this application. [Figure 17] This is a schematic diagram of noise grade classification of the spatial distribution of noise consisting of any measurement position provided by one embodiment of this application. [Figure 18] A noise monitoring method for a charging post provided by one embodiment of this application, the present invention provides a sound source energy distribution diagram based on beamforming. [Figure 19] A noise monitoring method for a charging post provided by one embodiment of this application, the present invention relates to a sound source energy distribution diagram based on Bayesian estimation and beamforming. [Modes for carrying out the invention]

[0011] To further clarify the purpose, technical means, and advantages of this application, the following drawings and examples are used to provide a more detailed explanation of this application. It should be understood that the specific examples described herein are used solely for illustrative purposes and are not intended to limit this application.

[0012] This application provides a noise monitoring method for a charging post, which includes a 3D ball array consisting of multiple microphones, used in a noise monitoring system. Since the 3D ball array is an existing technology, its specific details will not be described again.

[0013] As shown in Figure 1, in one embodiment of this application, the noise monitoring method for the charging post includes the following steps S100 to S700.

[0014] S100 is the charging post under test Operating state (on state) In this step, noise detection data is obtained from several measurement positions surrounding the circumferential direction of the charging post under test,

[0015] Specifically, the noise detection data is a sound pressure signal measured by a microphone in a 3D ball array.

[0016] S200 is the stage in which the acquired noise detection data undergoes noise reduction processing to obtain the initial noise signal. Specifically, depending on the location environment of the charging post under measurement, there may be interference from background noise (such as noise from other DC fast charging posts, pedestrians, vehicles, and shops) and environmental interference (temperature, air density). Therefore, in this application, noise reduction processing is applied to the acquired noise detection data to eliminate the influence of the background environment and improve the accuracy of subsequent noise positioning results.

[0017] S300 is the step in which the noise signal at the target analysis frequency is obtained after filtering the initial noise signal. Specifically, the target analysis frequency is a frequency segment, a particular frequency value, or a set of specific frequency values ​​that can represent the noise characteristics of the charging post under test.

[0018] More specifically, since the noise generated by DC rapid charging posts is mostly low-frequency noise, the target frequency can be set to a low frequency, allowing for effective detection of the noise from the charging post under test.

[0019] The specific filtering method involves using a Bessel circuit to eliminate high-frequency noise from the initial noise signal, thereby obtaining low-frequency noise.

[0020] S400 is a step in which a scanning area is selected from the sound source plane and the scanning area is uniformly divided into several discrete scanning grid points. Specifically, the sound source plane is the measurement surface directly in front of the 3D ball array. The scanning area is the scanning detection area of ​​the 3D ball array, centered on the charging post under test, including the outer surface of the charging post on the side directly in front of the 3D ball array, and also including the area that can be positioned at each measurement position along the circumference.

[0021] For example, in the embodiment described below, as shown in Figure 3, the black dots are discrete scanning grid points, uniformly distributed across the scanning area. Through the division of the scanning grid points, the noise output of the charging post under measurement is considered as the combined output of multiple discrete sound sources.

[0022] S500 is the step in which a vector consisting of estimated beamforming energies for all scanning grid points is obtained using a noise signal of the target frequency.

[0023] S600 is the step of creating an energy transfer model using a vector and energy transfer matrix consisting of beamforming energy estimates for at least every scanning grid point.

[0024] S700 is the step of calculating the high-resolution sound source energy distribution of the charging post under measurement in the energy transfer model using Bayesian estimation.

[0025] In this embodiment, noise detection data from several measurement locations is collected through asynchronous measurement using a 3D ball array. Simultaneously, noise reduction and filtering processes are performed on the noise detection data to obtain a noise signal at a target frequency that can represent the characteristics of the charging post under test. Furthermore, a vector composed of estimated beamforming energy values ​​at scanning grid points is obtained from this signal. This enables rapid inverse analysis of the visualized sound source, improving the spatial resolution of sound source positioning, and resulting in good robustness and intuitiveness. In addition, by explaining the variables in the sound field from the perspective of statistical optimization using Bayesian estimation, the distribution of sound source energy at scanning grid points is calculated, and the fault location in the charging post under test is accurately positioned. This results in high resolution imaging results, strong noise resistance of the algorithm, and strong self-adaptive capabilities. Therefore, the beamforming sound source positioning method based on Bayesian estimation can rapidly orient to low-frequency noise generated from the charging post under test, enabling rapid positioning of faults in DC rapid charging posts.

[0026] By using the above method to promptly detect and address DC charging post failures in their initial stages, the probability of various accidents caused by failures can be significantly reduced. Furthermore, when abnormal noises or sounds occurring at abnormal frequency intervals occur during the operation of DC rapid charging posts, the problem can be promptly fed back, playing a role in preventing accidents and allowing for directional adjustments to the structural design of the charging posts.

[0027] In one embodiment of this application, S100 includes S110 to S120.

[0028] S110 is a step in which multiple measurement positions are set using the center of the contour line of the charging post to be measured as the origin, a predetermined distance as the radius, and predetermined angular intervals. Specifically, the outer surface contour of the charging post equipment to be measured is drawn in a plane, and the center of the contour is set as the origin of the movement trajectory of the 3D ball array.

[0029] S120 is a step in which the measurement height is set in advance, and noise detection is performed at least once at each measurement position along the predetermined direction of movement, Specifically, as shown in Figure 2, while maintaining a constant distance from the origin, the system moves at a rotation angle of 30° clockwise (or counterclockwise), completing a full circuit around the charging post under test and collecting noise detection data at every measurement position. In Figure 2, the dark-colored ball array represents the current measurement position of the 3D ball array, while the lighter-colored ball array represents other measurement positions.

[0030] Of course, as another measurement method of this application, while the distance to the outer surface of the charging post under test remains constant, the device moves at a rotation angle of 30° clockwise, circling the charging post under test to acquire noise detection data at all measurement positions.

[0031] The aforementioned preset measurement height is either the height of the cooling fan system or transformer of the charging post under test, or half the height of the charging post under test.

[0032] In this embodiment, the noise generated by the cooling fan system and transformer of the charging post under test is considered the main noise source of the charging post, and therefore it is designated as the main monitoring point.

[0033] In one embodiment of this application, S200 includes S210 to S230.

[0034] S210 is the stage in which the noise detection data is subjected to a short-time Fourier transform to obtain the initial spectrum and initial phase.

[0035] S220 is a step in which the initial spectrum is sent to a noise prediction model, the noise prediction model is run, and the background noise spectrum output from the noise prediction model is obtained. Specifically, the noise prediction model is for the charging post being measured. Non-operating state (off state)The convergence training is performed using the following specific modeling method: the pure signal and noise signal are sampled at a constant frequency to form multiple new pure and noise signals of the same length. The pure and noise signals are randomly compressed and expanded and then mixed to obtain diverse mixed audio, ensuring diversity in the mixed audio. After applying a Fourier transform to the various mixed audio signals, the final input spectrum is obtained and used to train the noise prediction model.

[0036] S230 is the step in which the background noise spectrum is subtracted from the initial spectrum, and after a Fourier transform, the initial noise signal after noise reduction is obtained.

[0037] In this embodiment, noise reduction processing is performed on the acquired noise detection data to eliminate the influence of the background environment and improve the accuracy of the subsequent noise positioning results.

[0038] In one embodiment of this application, the formula for calculating the vector consisting of the estimated beamforming energy values ​​of all scanning grid points in S500 is:

[0039]

number

[0040]

number

[0041]

number

[0042]

number

[0043]

number

[0044] In actual calculations, the effects of noise can be eliminated by setting the main diagonal elements of the cross-power spectrum matrix C to zero.

[0045] In this embodiment, by using a 3D ball array, the inverse analysis of a visualized sound source can be quickly realized by a beamforming method measured asynchronously, which has good robustness and good intuitiveness. At each measurement position, it is possible to obtain the inspection grid points related to the maximum noise direction with respect to the measured charging post. Furthermore, the maximum noise positions at all measurement positions around the measured charging post can be obtained, and by means of the asynchronous measurement method, the purpose of complementation can be achieved, and the vertical resolution can be further improved.

[0046] Although the above beamforming method can quickly realize the inverse analysis of the visualized sound source, as shown in FIG. 18, its imaging result is not clear, especially in the case of low frequencies, it is more blurred.

[0047] In an embodiment of the present application, the S600 includes the following S610 to S620. S610 is the step of creating an energy transfer matrix shown in the calculation formula 6 (Equation 6 ), where

[0048]

Equation

[0049]

Equation

[0050] In one embodiment of this application, S700 includes S710 to S730.

[0051] S710 is a variational prior distribution q1(x|γ x ),q2(γ x ) and q3(γ ε Through ), the binding distribution p(x,γ x γ ε This is the stage where we approach |y)∝p(θ,y), Of these, q1(·), q2(·), q3(·) and p(·) are all probability density functions, and ∝ indicates a positive correlation.

[0052] The S720 uses mean-field theory to determine the variational prior distribution parameter x (i.e., the high-resolution sound source energy distribution), γ x and gamma ε At the stage of integrating into the variable θ, q(θ) = q1(x)q2(γ x )q3(γ ε ) satisfies the following: γ x is the covariance matrix of x, and γ ε This is the covariance matrix of ε, and it can be kept constant. The minimum KL (Kullback-Leibler) variance rule is given by formula 8.

[0053]

number

[0054]

number

[0055]

number

[0056]

number

[0057]

number

[0058] In this embodiment, Bayesian estimation can be used to explain the variables in the sound field from the perspective of statistical optimization. By obtaining a solution for the high-resolution sound source energy distribution x, beamforming can be combined with Bayesian estimation to perform sound source positioning. Furthermore, rapid orientation can be performed for low-frequency noise generated by DC rapid charging posts. As shown in Figure 19, the imaging result has high resolution, strong noise resistance of the algorithm, and strong self-adaptive ability.

[0059] In one embodiment of this application, the noise monitoring method for the charging post, from S700 onward, further includes S810 to S820.

[0060] S810 is the step of calculating the average noise level at the measurement location of the charging post under test based on the noise detection data.

[0061] S820 is the stage in which a noise radiation diagram is created consisting of all measurement locations, based on the azimuth angle of each measurement location and the noise value corresponding to each individual measurement location.

[0062] For example, in one embodiment described below, Figures 3 to 15 show the spatial distribution of noise signals and average noise collected from individual measurement locations, and the step involves correlating the average noise value with the corresponding measurement location in the spatial distribution diagram of the highest temperature.

[0063] Figure 16 shows the spatial distribution of noise across all measurement locations, including the average noise level at each location.

[0064] Figure 17 shows the noise level classifications for the spatial distribution of noise at various measurement locations. During the experiment, only the left charging head of the charging post under test was operating. As can be seen from Figure 17, the loudest noise was at the exhaust vents on both sides of the charging post, and the noise in the direction of the left exhaust vent already exceeded the Class 4 noise standard, thus constituting noise pollution.

[0065] In this embodiment, creating a noise radiation diagram allows for an intuitive representation of the noise surround situation near the charging post under test, thus providing important reference value for manufacturers' designs and the placement of DC charging posts.

[0066] In one embodiment of this application, S100 can collect multiple sets of noise detection data at each individual measurement location. The aforementioned S810 includes the following S811 to S815. S811 is the stage where a measurement location is selected. S812 is the step of determining whether the difference between the maximum and minimum values ​​in the multiple sets of noise detection data at the measurement location exceeds 5 dB. S813 is a step in which, when the difference between the maximum and minimum values ​​among multiple sets of noise detection data at the measurement location exceeds 5 dB, the average value is calculated for the multiple sets of noise detection data using the energy averaging method and is set as the average noise value at the measurement location.

[0067] Specifically, when calculating the average value using the energy averaging method, 10 sets of noise detection data are collected at the same time and time interval. The specific calculation method can be found in the ISO proximity field test method, and will not be explained again in detail.

[0068] S814 is a step in which, when the difference between the maximum and minimum values ​​among the multiple sets of noise detection data at the measurement location is less than 5 dB, an arithmetic mean is calculated for the multiple sets of noise detection data to be used as the average noise value at the measurement location.

[0069] S815 is the step of returning to the step of selecting one of the measurement positions until all possible measurement positions have been selected.

[0070] This application further provides a noise monitoring system.

[0071] In one embodiment of this application, the noise monitoring system includes a 3D ball array and information processing equipment.

[0072] Specifically, the 3D ball array used for collecting sound pressure signals includes multiple microphones. More specifically, the 3D ball array uses a high-performance 64-bit ball array to maximize the accuracy of noise detection, thereby highly simulating the noise disturbances that people experience in real environments.

[0073] The information processing equipment is connected to the 3D ball array, and further connected to the memory and the processing unit, with the memory being coupled to the processing unit. The memory is used to store program data, and the processing unit is used to execute the program data and implement the noise monitoring method for the charging post described above.

[0074] This application also provides a computer-readable medium on which a computer program is stored, and when the computer program is executed by a processing unit, the noise monitoring method for a charging post described above is realized.

[0075] The technical features described in the above embodiments can be combined in any way. The steps of each method do not limit the procedure of implementation, and for the sake of brevity, not every possible combination of the technical features in the above embodiments has been described in detail. However, as long as these combinations of technical features do not conflict with each other, they should be understood to fall within the scope described herein.

[0076] Although the aforementioned examples describe only a few embodiments of this application, and the descriptions are relatively specific and detailed, this should not be understood as limiting the scope of the claims of this application. A person of ordinary skill in the art could make many variations and modifications without deviating from the idea of ​​this application, and it goes without saying that all of these would fall within the scope of protection of this application.

Claims

1. A noise monitoring method for a charging post used in a noise monitoring system, comprising a 3D ball array consisting of multiple microphones, The process involves acquiring noise detection data, including sound pressure signals measured by the plurality of microphones, from several measurement positions surrounding the circumferential direction of the charging post under measurement, while the charging post under measurement is in operation. The process involves obtaining an initial noise signal after noise reduction processing on the acquired noise detection data, The process involves obtaining a noise signal at the target analysis frequency after filtering the initial noise signal, The steps include selecting a scanning area from the sound source plane and uniformly dividing the scanning area into several discrete scanning grid points, The steps include obtaining a vector consisting of estimated beamforming energy values ​​for all scanning grid points using a noise signal at the target analysis frequency, A step of creating an energy transfer model using a vector and energy transfer matrix consisting of the estimated beamforming energy values ​​for at least all of the scanning grid points, and The process includes the step of calculating the high-resolution sound source energy distribution of the measured charging post in the energy transfer model using Bayesian estimation. A method for monitoring noise at a charging station, characterized by the following features.

2. In the operating state of the charging post to be measured, the step of acquiring noise detection data from several measurement positions surrounding the circumferential direction of the charging post to be measured is: The steps include setting multiple measurement positions using the center of the contour line of the charging post to be measured as the origin, a predetermined distance as the radius, and predetermined angular intervals, and The process includes a step of pre-setting the measurement height and performing noise detection at least once at each measurement position along the pre-set direction of movement, The noise monitoring method for a charging post as described in feature 1.

3. The step of obtaining an initial noise signal after noise reduction processing on the acquired noise detection data is as follows: The process involves obtaining the initial spectrum and initial phase from noise detection data through a short-time Fourier transform, and The process involves sending an initial spectrum to a noise prediction model, running the noise prediction model to obtain the background noise spectrum output from the noise prediction model, and the noise prediction model undergoing intensive training while the charging post under measurement is inactive. This includes the step of subtracting the background noise spectrum from the initial spectrum, performing a Fourier transform, and obtaining the initial noise signal after noise reduction. The noise monitoring method for a charging post as described in feature 1.

4. In the step of obtaining a vector consisting of beamforming estimates for all scanning grid points using the noise signal at the target analysis frequency, the vector consisting of beamforming energy estimates for all scanning grid points is: 【Number 1】 Eventually, y ｎ Here, the nth scanning grid point is the beamforming result at the target analysis frequency f, N represents the total number of scanning grid points, and T represents the transpose of the vector. y in the above number 1 ｎ Formula for (f): [Math 2] eventually, a ｎ is the steering vector from any microphone in the 3D ball array to the nth scanning grid point, * represents the conjugate transpose, and n represents the order of the scanning grid points, ||・|| ２ is L ２ The norm of is where C is the cross-power spectrum matrix, a in the above number 2 ｎ Formula for (f): [Math 3] In this case, M is the total number of microphones in the 3D ball array, m is the order of the microphones in the 3D ball array, and T represents the transpose of the vector. a in the above number 3 ｍ，ｎ The formula for calculation: [Math 4] Among them, j is a unit, and r ｍ，ｎ is the distance between the mth microphone and the nth scanning grid point, c is the speed of sound in air, and f is the target analysis frequency. The formula for calculating C(f) in the above equation 2: [Math 5] Among them, p ｍ (f) represents the f-frequency component obtained after noise removal processing and filtering processing of the sound pressure signal measured by the m-th microphone in the 3D ball array, E[·] represents the expected value, and T represents the transpose of the vector. The noise monitoring method for a charging post as described in feature 1.

5. The step of creating an energy transfer model using a vector and energy transfer matrix consisting of beamforming energy estimates for at least every scanning grid point is as follows: The energy transfer matrix shown in Equation 6: [Math 6] Eventually, h ｎ１，ｎ２ is an element in H, The creation of an energy transfer model is [Number 7] Of these, y is a vector composed of the estimated beamforming energy values ​​for all scanning grid points, where ε = (ε １ , ε ２ , ..., ε ｎ ) Ｔ is the model error, T is the vector transpose, and h ｎ１，ｎ２ n is an element in the energy transfer matrix H, and n １ = 1 ,...,N、n ２ =1 , . . . , N, a ｎ１ is the nth １ This is the steering vector from the individual scanning grid points to the 3D ball array, a ｎ２ is the nth ２ This is the steering vector from the individual scan grid points to the 3D ball array, where * indicates the conjugate transpose. The noise monitoring method for a charging post according to feature 4.

6. The step of obtaining a solution for the high-resolution energy distribution of the measured charging post in the energy transfer model based on the Bayesian estimation method is: Variational prior distribution q １ (x|γ ｘ ), q ２ (γ ｘ ) and q ３ (γ ε Through ), the binding distribution p(x,γ ｘ γ ε The stage where |y)∝p(θ,y) approaches, Eventually, q １ (・), q ２ (・), q ３ Both (•) and p(•) are probability density functions, and ∝ indicates a positive correlation. According to mean-field theory, variational prior distribution parameters x and γ ｘ and gamma ε Integrating into the variable θ, we get q(θ) = q １ (x)q ２ (γ ｘ )q ３ (γ ε ) the stage that satisfies, and of those, γ ｘ is the covariance matrix of x, and γ ε This is the covariance matrix of ε, The rule for minimizing KL variance is given by Equation 8. [Number 8] Based on prior distribution knowledge, the variation of x and the prior distribution are shown as the mean μ, the variance is a multidimensional Gaussian distribution of Σ, and based on the rarefaction and Gaussian mixing properties of the Student t distribution, γ ｘ and gamma ε The stage where the variational prior distribution is shown as the synergistic product of the inverse gamma distribution, [Number 9] Of these, N(•) follows a normal distribution. According to the KL variance minimization rule, the iterative process for x is as follows: [Number 10] Of these, l is the l-th repetition, γ ｘ and gamma ε The iterative process is as follows: [Math 11] The recurring deadline conditions are [Math 12] Therefore, ρ is a constant, indicating that it accepts the error. The noise monitoring method for a charging post according to claim 5, characterized by including the three steps described above.

7. After obtaining a solution for the high-resolution sound source distribution of the charging post under measurement in the energy transfer model based on Bayesian estimation, the noise monitoring method for the charging post under measurement further includes: The step of calculating the noise level at the measurement location of the charging post to be measured using the noise detection data, This includes the step of creating a noise radiation diagram composed of all measurement locations, based on the azimuth angle of each measurement location and the noise level corresponding to each individual measurement location. The noise monitoring method for a charging post as described in feature 1.

8. In the operating state of the charging post under measurement, from noise detection data obtained from several measurement positions surrounding the circumference of the charging post under measurement, multiple sets of noise detection data are collected from each individual measurement position, The step of measuring the noise level at the measurement location of the charging post to be measured using the noise detection data is as follows: The stage of selecting one measurement location, The step of determining whether the difference between the maximum and minimum values ​​in multiple sets of noise detection data from the aforementioned measurement location exceeds 5 dB, When the difference between the maximum and minimum values ​​in multiple sets of noise detection data at the aforementioned measurement location exceeds 5 dB, the average value is calculated for the multiple sets of noise detection data using the energy averaging method and is taken as the average noise value at the measurement location. In the step of calculating an arithmetic mean of the noise detection data for multiple sets of noise detection data at the measurement location when the difference between the maximum and minimum values ​​is less than 5 dB, and using this as the average noise value for the measurement location, The step includes returning to the step of selecting one of the measurement positions until all measurement positions have been selected, The noise monitoring method for a charging post according to feature 3.

9. A noise monitoring system used in the noise monitoring method for a charging post according to any one of claims 1 to 8, The 3D ball array used to collect the sound pressure signal generated in the operating state of the charging post under measurement, An information processing device connected by a signal to the 3D ball array, comprising a memory and a processor, wherein the memory is coupled to the processor, the memory is used to store program data, and the processor is used to (A) process and analyze noise detection data collected from the 3D ball array, (B) execute algorithms such as beamforming, (C) filtering, and (D) Bayesian estimation, to carry out a noise monitoring method for the charging post. The aforementioned processor, a. (A) involves controlling the 3D ball array to acquire noise detection data including sound pressure signals measured by multiple microphones at multiple measurement positions, performing noise reduction processing on the acquired noise detection data to obtain an initial noise signal, and further, (C) involves performing filtering processing on the initial noise signal to obtain a noise signal at the target analysis frequency. b. Selecting a scanning area on the sound source plane and dividing the scanning area equally into a plurality of discrete scanning grid points, c. The above (B) involves performing beamforming on the noise signal of the target analysis frequency and obtaining estimated beamforming energy values ​​at all scanning grid points, d. Constructing an energy transfer model based on a vector consisting of a pre-set energy transfer matrix and the estimated beamforming energy, e. The above (D) involves solving the energy propagation model using Bayesian estimation and obtaining a high-resolution sound source energy distribution map of the charging post under measurement, f. A noise monitoring system characterized by being configured to output or save a sound source distribution map for noise analysis and evaluation.

10. A computer-readable medium that stores a computer program, and from which the computer program is executed by a processor to realize the noise monitoring method for a charging post according to any one of claims 1 to 8.

Images