Acousto-optic cooperation-based arcing monitoring method and device

Abstract

The invention provides an arcing monitoring method and device based on acousto-optic cooperation, and relates to the technical field of arcing monitoring, and the method comprises the steps: obtaining a sound signal and an optical signal of a pantograph based on a sensor; performing multi-stage denoising and signal reconstruction on the sound signal to obtain a target sound signal; determining first data based on the target sound signal, wherein the first data is the number of arcing and the time of each arcing; determining second data based on the optical signal, wherein the second data is the number of arcing and the time of each arcing; based on a Bayesian formula, the first data and the second data, an arcing occurrence probability value is obtained through calculation; and determining an arcing result of the pantograph based on the arcing occurrence probability value. According to the invention, an acousto-optic synergistic pantograph net arcing monitoring method is adopted, monitoring is carried out by combining a sound signal and an optical signal, the defects of a traditional image detection method, a contact detection method, an ultraviolet detection method and a single sound detection method are effectively overcome, and the arcing monitoring accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of arc monitoring, and in particular to an arc monitoring method and device based on acoustic and optical synergy. Background Art

[0002] At present, the pantograph-net arcing monitoring device still has many defects: (1) The monitoring device using image recognition technology needs to store and process a large amount of image data, which leads to high hardware requirements, high cost and long processing time; (2) The contact monitoring device needs to install the sensor on the pantograph, which will change the pantograph structure and affect the pantograph performance; (3) The commonly used ultraviolet detection method, although the device is simple, will be affected by other light sources, resulting in inaccurate monitoring results; (4) At present, the monitoring device needs to be connected to the power, communication and other systems of the electric locomotive, which is complicated to install. If arcing occurs, the overvoltage may damage the electric locomotive. Therefore, there is an urgent need for an arcing monitoring device that does not rely on the power supply of the electric locomotive and has a simple algorithm that is not easily interfered with. Summary of the invention

[0003] The purpose of the present invention is to provide an arc monitoring method and device based on acoustic and optical synergy to improve the above problems. In order to achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0004] In a first aspect, the present application provides an arc monitoring method based on acoustic-optical synergy, comprising:

[0005] Acquiring sound and light signals of the pantograph based on sensors;

[0006] Performing multi-level denoising and signal reconstruction on the sound signal to obtain a target sound signal;

[0007] Determine first data based on the target sound signal, the first data being the number of arcs and the time of each arc;

[0008] Determine second data based on the optical signal, the second data being the number of arcs and the time of each arc;

[0009] Based on the Bayesian formula, the first data and the second data, a probability value of arcing occurrence is calculated;

[0010] An arcing result of the pantograph is determined based on the arcing occurrence probability value, and the arcing result is used to indicate whether arcing occurs in the pantograph.

[0011] In a second aspect, the present application also provides an arc monitoring device based on acoustic and optical synergy, comprising:

[0012] A first acquisition unit, configured to acquire a sound signal and a light signal of the pantograph based on a sensor;

[0013] A denoising and reconstruction unit, used for performing multi-level denoising and signal reconstruction on the sound signal to obtain a target sound signal;

[0014] A first determining unit, configured to determine first data based on the target sound signal, wherein the first data is the number of arcs and the time of each arc;

[0015] A second determining unit, configured to determine second data based on the optical signal, wherein the second data is the number of arcs and the time of each arc;

[0016] A first calculation unit, configured to calculate an arcing occurrence probability value based on a Bayesian formula, the first data, and the second data;

[0017] The third determining unit is used to determine the arcing result of the pantograph based on the arcing occurrence probability value, and the arcing result is used to indicate whether the pantograph has arcing.

[0018] In a third aspect, the present application also provides an arc monitoring device based on acoustic and optical synergy, comprising:

[0019] Memory for storing computer programs;

[0020] A processor is used to implement the steps of the arc monitoring method based on sound and light coordination when executing the computer program.

[0021] In a fourth aspect, the present application further provides a readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the above-mentioned arc monitoring method based on sound and light coordination are implemented.

[0022] The beneficial effects of the present invention are:

[0023] The present invention adopts an acoustic-optical coordinated bow-cat arc monitoring method, which effectively overcomes the shortcomings of traditional image detection method, contact detection method, ultraviolet detection method and single sound detection method by combining sound signal and light signal for monitoring, and improves the accuracy of arc monitoring.

[0024] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by implementing the embodiments of the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0026] Figure 1 Schematic diagram of the flow of the arc monitoring method based on sound and light synergy described in an embodiment of the present invention;

[0027] Figure 2 Schematic diagram of the structure of an arc monitoring device based on sound and light synergy described in an embodiment of the present invention;

[0028] Figure 3 It is a schematic diagram of the structure of the arc monitoring device based on sound and light coordination described in an embodiment of the present invention.

[0029] Markings in the figure:

[0030] 10. First acquisition unit; 20. De-noising and reconstruction unit; 30. First determination unit; 40. Second determination unit; 50. First calculation unit; 60. Third determination unit; 800. Arc monitoring device based on acoustic and optical synergy; 801. Processor; 802. Memory; 803. Multimedia component; 804. I / O interface; 805. Communication component. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0032] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0033] Embodiment 1:

[0034] This embodiment provides an arc monitoring method based on sound and light coordination.

[0035] See also Figure 1 , the figure shows that the method includes step S10, step S20, step S30, step S40, step S50 and step S60.

[0036] Step S10. Acquire the sound signal and light signal of the pantograph based on the sensor;

[0037] Specifically, the sound sensor was used to collect the sound signal, and the collected sound signal was analyzed by combining the fast Fourier transform and the spectrogram. It was found that when the train was running, the sound signals were mostly concentrated below 5kHz, and the main frequencies in this band included the running sound of the traction motor, the vibration sound of the train, and the aerodynamic noise. The frequency band of the arcing sound signal was widely distributed, usually concentrated in the range of 5 to 17kHz.

[0038] The ultraviolet phototube is used to collect the ultraviolet light band emitted by the pantograph. The ultraviolet light band when arcing occurs is mainly concentrated in the entire band (10-400nm): 240-260nm, 300-330nm, 390-400nm. The ultraviolet phototube will only capture ultraviolet light with a wavelength of 185nm-260nm, because the natural ultraviolet light in this range will be absorbed by the ozone layer, thus avoiding external interference. After capturing ultraviolet light, the ultraviolet phototube will produce a photoelectric effect, generating irregular voltage pulse signals. After being processed, the voltage pulse signals become signals that can be recognized by the terminal and transmitted to the terminal.

[0039] Step S20. Perform multi-level denoising and signal reconstruction on the sound signal to obtain a target sound signal;

[0040] Specifically, step S20 specifically includes step S21, step S22, step S23, step S24, step S25 and step S26:

[0041] Step S21. Filter the sound signal based on the filter to obtain a first sound signal;

[0042] Specifically, the sound signal is processed using a bandpass filter, and the filtered sound signal is transmitted to the terminal. The bandpass filter allows signals within the frequency range of 5 to 17 kHz to pass through, while suppressing signals outside this range. The bandpass filter used in this application is designed based on a Butterworth filter. Different bandpass filters will be used for different filtering requirements, and there is no special restriction here.

[0043] Step S22. Decompose the first sound signal based on wavelet decomposition to obtain a plurality of sub-signals of different scales;

[0044] Specifically, wavelet decomposition is a technique used for signal analysis, which decomposes a signal into multiple sub-signals of different frequency scales, each sub-signal representing a component of the original signal within a specific frequency range.

[0045] Step S23. Analyze multiple sub-signals based on grey correlation analysis algorithm and wavelet decomposition to obtain approximate coefficients and multiple detail coefficients;

[0046] Specifically, the grey correlation analysis algorithm is a method for evaluating sequence similarity. The present application uses the algorithm to perform correlation analysis on sub-signals of different frequency scales, thereby avoiding the overall analysis being affected by abnormal signals on a single frequency scale.

[0047] Specifically, step S23 specifically includes step S231, step S232, step S233, step S234, step S235 and step S236:

[0048] Step S231. Extract features from the sub-signals at each scale and encode them to obtain multiple feature matrices, where the feature matrix includes a first matrix and multiple second matrices, the features corresponding to the first matrix are peak features, and the features corresponding to the second matrix are trough features;

[0049] Step S232. Calculate the correlation values ​​of the pairwise feature data in the feature matrix corresponding to the sub-signals of adjacent scales based on the grey correlation analysis algorithm, where the pairwise feature data are feature data located at the same time scale in different feature matrices;

[0050] Specifically, for each pair of sub-signals of adjacent scales, the feature data in their feature matrices at the same time point are compared to measure the similarity between these feature data.

[0051] Step S233. Consider each feature data in the feature matrix corresponding to the sub-signals of adjacent scales as a node, and regard the correlation value of each pair of feature data as the edge between two nodes;

[0052] Specifically, the relationship between feature data is constructed into a graph structure, where each node represents a feature data and the edge weights between nodes reflect their association.

[0053] Step S234. Perform graph convolution operation based on the structural feature extraction neural network algorithm, pass the feature information of each node to the adjacent nodes, and update the features in combination with the information of the adjacent nodes to obtain updated feature node information;

[0054] Specifically, the graph convolution operation combines the feature information of each node with the information of its adjacent nodes. Through multi-layer transmission and updating, the feature information of each node is made more representative and global, thereby optimizing the feature expression and making the useful information in the signal more prominent.

[0055] Step S235. Perform integration and reconstruction processing based on the characteristic node information to obtain a second sound signal;

[0056] Specifically, a second sound signal is reconstructed according to the updated feature node information. The second sound signal is reconstructed based on richer and more accurate feature data, and part of the noise in the original signal is eliminated.

[0057] Step S236. Decompose the second sound signal based on wavelet decomposition to obtain an approximate coefficient and a plurality of detail coefficients;

[0058] Specifically, the second sound signal is subjected to wavelet decomposition to obtain approximate coefficients and detail coefficients, and the second sound signal is further denoised based on these coefficients.

[0059] Step S24. Determine the optimal threshold value for each scale based on the soft threshold function and multiple detail coefficients;

[0060] Specifically, the soft threshold function is used to perform denoising on the signal. By applying the soft threshold processing to the detail coefficients of each frequency scale, the noise can be removed and the useful signal can be retained.

[0061] Specifically, step S24 specifically includes step S241, step S242, step S243, step S244 and step S245:

[0062] Step S241. Obtain a preset threshold value at each scale to obtain multiple initial threshold values;

[0063] Specifically, the initial threshold is usually set according to the statistical characteristics of the signal, empirical data or algorithm, and is not particularly limited here.

[0064] Step S242. Calculate the absolute value of each detail coefficient to obtain an absolute detail coefficient;

[0065] Step S243. Calculate the difference between the absolute detail coefficient and the initial threshold value at the same scale to obtain a target difference;

[0066] Step S244. When the target difference is not greater than the preset second threshold, the optimal threshold is 0;

[0067] Step S245. When the target difference is greater than the second set threshold, the product of the sign function value of the detail coefficient and the target difference is calculated to obtain the optimal threshold;

[0068] Specifically, the soft threshold function calculation formula is:

[0069] x t =sign(x)*max(|x|-λ,0)

[0070] Among them, x t is the optimal threshold; sign() is the sign function; λ is the initial threshold; max() is the maximum value function; x is the detail coefficient.

[0071] Step S25. Perform soft threshold processing on the detail coefficient of each scale based on the optimal threshold to obtain a plurality of processed target detail coefficients;

[0072] Specifically, the optimal threshold determined above is applied to detail coefficients of each scale and soft threshold processing is performed thereon, so that noise components below the optimal threshold can be eliminated and only signal components above the threshold are retained.

[0073] Step S26. Reconstruct the signal based on the inverse wavelet transform, multiple target detail coefficients and approximate coefficients to obtain the target sound signal;

[0074] Specifically, the processed target detail coefficients and approximation coefficients are recombined into the target sound signal using inverse wavelet transform. This transformation process ensures that the main features and structures of the signal are preserved, while the noise components are effectively suppressed.

[0075] Step S30. Determine first data based on the target sound signal, the first data being the number of arcs and the time of each arc;

[0076] Specifically, there is a significant difference in the energy of the sound signal when arcing occurs and when not. Therefore, it is considered to describe the characteristic changes of the sound signal of the bow-catenary arc through short-time average energy analysis, so as to realize the identification of the bow-catenary arc.

[0077] Short-time energy is the energy of a signal in a short time window, which can be used to detect the instantaneous energy change of the signal. For a given window length N and step length H, the short-time energy calculation formula is as follows:

[0078]

[0079] Among them, n is the index of the current time window; H is the step size; m is the sample index in the current time window, ranging from 0 to N-1; N is the window length; x is the sampling sequence of the input signal.

[0080] The time, number and relative intensity of arcing can be determined from the sound signal based on the short-term energy.

[0081] Step S40. Determine second data based on the optical signal, the second data being the number of arcs and the time of each arc;

[0082] Specifically, a pulse monitoring device is provided in the terminal, which can record the input signal, so as to determine the time and number of arcing according to the optical signal.

[0083] When the pulse monitoring device captures the first voltage pulse, it enters the interrupt start timer. If the timer recording time does not exceed the set time when the second voltage pulse comes, it continues to run; if no new voltage pulse is captured within the set time after the previous voltage pulse, the timer records the time between the first voltage pulse and the last voltage pulse. This is an arc, and the time and number of single arcs can be counted. Usually, if the arc interval time exceeds 40ms, it is divided into two arcs, so the set time of this application is 40ms.

[0084] Step S50. Calculate the arcing probability value based on the Bayesian formula, the first data and the second data;

[0085] Specifically, arc monitoring based only on sound signals is easily affected by interference signals such as braking and whistling, which may lead to misjudgment; and in the process of arc monitoring based only on light signals, although a solar-blind ultraviolet photoelectric tube is used, it may still be interfered by artificial light sources. Therefore, the two characteristic signals, sound signal and light signal, are processed in a coordinated manner to effectively reduce their respective interference. These two signals are selected for analysis not only because they have significant characteristics, but also because the equipment corresponding to signal acquisition and analysis is low-cost and can significantly improve the accuracy of monitoring.

[0086] Specifically, step S50 specifically includes step S51, step S52, step S53 and step S54:

[0087] Step S51. Determine a priori parameters and weight parameters based on historical arcing data, the a priori parameters include a priori probability of arcing occurring, a priori probability of arcing not occurring, a probability of sound signal response when arcing occurs, a probability of light signal response when arcing occurs, a probability of sound signal response when arcing does not occur, and a probability of light signal response when arcing does not occur, and the weight parameters include a weight of the sound signal, a weight of the light signal, and a weight of the a posteriori probability;

[0088] Step S52. Determine a marginal probability value based on the prior parameters, the first data and the second data;

[0089] Step S53. Determine the posterior probability of arcing based on the marginal probability value and the priori parameters;

[0090] Specifically, considering that the environmental factors of different lines may vary greatly, this application combines the Bayesian network with the weighted voting method. By adjusting the parameters of the Bayesian network, this solution can better adapt to different environments and further improve the robustness and stability of this application.

[0091] A Bayesian network was constructed, and the nodes were defined as: A: arcing occurrence (yes / no); S: sound signal detection (yes / no); U: light signal detection (yes / no).

[0092] The prior parameters are set based on previous experience. The actual values ​​of the prior parameters corresponding to different lines may be different, and no special restrictions are made here. The prior parameters include: P(A) the prior probability of arcing; P(S|A) the probability of sound signal response when arcing occurs; P(U|A) the probability of light signal response when arcing occurs; The probability of response of the sound signal when arcing does not occur; The response probability of the optical signal when arcing does not occur; The a priori probability that arcing does not occur.

[0093] Considering that in this application, the sound sensor corresponding to the sound signal S has two outputs, 1 and 0, respectively, and the light signal U is the same, that is, the corresponding prior parameters include: P(1|A): the response probability of the sound signal when arcing occurs; P(1|A): the response probability of the light signal when arcing occurs; The probability of response of the sound signal when arcing does not occur; The probability of response of the optical signal when arcing does not occur; P(0|A): The probability of non-response of the sound signal when arcing occurs; P(0|A): The probability of non-response of the optical signal when arcing occurs; The probability of no response to the sound signal when arcing does not occur; The probability of non-response of the optical signal when arcing does not occur.

[0094] The calculation formula for the posterior probability of arcing occurrence P(A|S,U) is:

[0095]

[0096] Among them, P(A|S,U) is the posterior probability of arcing; P(S|A) is the response probability of the sound signal when arcing occurs; P(U|A) is the response probability of the light signal when arcing occurs; P(S,U) is the marginal probability; P(A): the prior probability of arcing;

[0097] The calculation formula of marginal probability P(S,U) is:

[0098]

[0099] Among them, P(S|A) is the response probability of the sound signal when arcing occurs; P(U|A) is the response probability of the light signal when arcing occurs; is the response probability of the sound signal when arcing does not occur; is the response probability of the optical signal when arcing does not occur; P(A) is the prior probability of arcing occurring; is the prior probability that arcing does not occur.

[0100] Step S54. Calculate the response probability of the sound signal when arcing occurs, the response probability of the light signal when arcing occurs, and the posterior probability of arcing, and sum them with the product of the weight of the sound signal, the weight of the light signal, and the weight of the posterior probability, respectively, to obtain the probability value of arcing occurrence;

[0101] Specifically, the calculation formula for the arcing probability value is:

[0102] W A =w S *P(S|A)+w U *P(U|A)+w P *P(A|S,U)

[0103] Among them, W A is the probability value of arcing; w S is the weight of the sound signal; w U is the weight of the optical signal; w P is the weight of the posterior probability; P(A|S,U) is the posterior probability of arcing; P(S|A) is the response probability of the sound signal when arcing occurs; P(U|A) is the response probability of the light signal when arcing occurs; w S +w U +w P =1.

[0104] Among all the prior parameters, except for the prior probability of arcing, the other prior parameters will not change during the calculation process of the current scenario. Only the prior probability of arcing will be updated, that is, the posterior probability of arcing calculated each time will be brought into the formula as the prior probability of arcing calculated next time. The following is an example:

[0105] For example, within 100 seconds, the sound sensor and the light sensor responded three times and transmitted data three times. The first time was sound signal S=1 and light signal U=0; the second time was sound signal S=0 and light signal U=1; the third time was sound signal S=1 and light signal U=1. The above calculation formula in this application is used to determine whether arcing occurs each time.

[0106] Before the operation, set the prior parameters, where:

[0107] P(A) = 0.1 (prior probability of arcing); (prior probability that arcing does not occur); P(S=1|A)=0.8 (probability of sound signal response when arcing occurs); P(S=0|A)=0.2 (probability of sound signal non-response when arcing occurs); P(U=1|A)=0.9 (probability of light signal response when arcing occurs); P(U=0|A)=0.1 (probability of light signal non-response when arcing occurs); (Probability of response of sound signal when arcing does not occur); (Probability of no response to sound signal when arcing does not occur); (Probability of response of optical signal when arcing does not occur); (the probability of non-response of the optical signal when arcing does not occur); w P =0.5 (weight of posterior probability); w S =0.3 (weight of sound signal); w U =0.2 (weight of optical signal);

[0108] For the first transmission of data, calculate the marginal probability:

[0109]

[0110] Compute the posterior probability:

[0111]

[0112] Bring in the weighted formula:

[0113] W A =w S *P(S=1|A)+w U *P(U=0|A)+w P *P(A|S=1,U=0)

[0114] =0.3*0.8+0.2*0.1+0.5*0.0357=0.01785+0.24+0.02

[0115] =0.27785

[0116] If the judgment threshold is set to 0.3, it is determined that no arcing occurs this time; and the posterior probability of arcing obtained this time is brought into the next calculation as the prior probability of arcing in the second transmission data calculation process.

[0117] The calculation process is the same as that of the first transmission data. Only the prior probability of arcing is changed. The results are as follows:

[0118] The arcing probability value W corresponding to the second data transmission A = 0.2629 is smaller than the threshold value 0.3, and it is determined that arcing does not occur.

[0119] The arcing probability value W corresponding to the third data transmission A =0.60275 is greater than the threshold value 0.3, and it is determined that arcing occurs.

[0120] In actual operation, when both sensors do not respond, that is, both output 0, the above operation is not performed. Only when any one of the sensors responds will the operation be performed to determine whether arcing occurs.

[0121] Step S60: determining the arcing result of the pantograph based on the arcing probability value, where the arcing result is used to indicate whether the pantograph has arcing;

[0122] Specifically, based on the arcing probability value W A Determine whether arcing occurs, where:

[0123]

[0124] Among them, θ is the decision threshold; W A is the probability value of arcing occurrence.

[0125] By setting a suitable threshold value θ, it is possible to effectively determine whether arcing occurs. The selection of the threshold value θ needs to be comprehensively considered based on safety requirements, false alarm risks, and historical data and experimental results in actual applications.

[0126] After calculating the arcing time and arcing relative intensity, the terminal can determine whether to notify relevant personnel for processing according to the preset threshold. For example, if the single arcing time exceeds 100ms or the single arcing intensity exceeds 10, an alarm operation will be performed, that is, the relevant information of the arcing will be sent to the relevant personnel, thereby improving the efficiency of maintaining the bow-net system.

[0127] Embodiment 2:

[0128] like Figure 2 As shown, this embodiment provides an arc monitoring device based on sound and light coordination, and the device includes:

[0129] A first acquisition unit 10, configured to acquire a sound signal and a light signal of the pantograph based on a sensor;

[0130] The denoising and reconstruction unit 20 is used to perform multi-level denoising and signal reconstruction on the sound signal to obtain a target sound signal;

[0131] A first determination unit 30, configured to determine first data based on the target sound signal, the first data being the number of arcs and the time of each arc;

[0132] A second determination unit 40, configured to determine second data based on the optical signal, the second data being the number of arcs and the time of each arc;

[0133] A first calculation unit 50, configured to calculate an arcing occurrence probability value based on a Bayesian formula, the first data, and the second data;

[0134] The third determination unit 60 is used to determine the arcing result of the pantograph based on the arcing occurrence probability value, and the arcing result is used to indicate whether arcing occurs in the pantograph.

[0135] In a specific implementation disclosed in the present application, the denoising and reconstruction unit 20 includes:

[0136] A filtering unit, configured to filter the sound signal based on a filter to obtain a first sound signal;

[0137] A first decomposition unit, configured to decompose the first sound signal based on wavelet decomposition to obtain a plurality of sub-signals of different scales;

[0138] An analysis unit, used for analyzing multiple sub-signals based on a grey correlation analysis algorithm and wavelet decomposition to obtain approximate coefficients and multiple detail coefficients;

[0139] A fourth determination unit, configured to determine an optimal threshold value for each scale based on the soft threshold function and a plurality of detail coefficients;

[0140] A processing unit, used for performing soft threshold processing on the detail coefficient of each scale based on the optimal threshold to obtain a plurality of processed target detail coefficients;

[0141] The first reconstruction unit is used to perform signal reconstruction based on inverse wavelet transform, multiple target detail coefficients and approximate coefficients to obtain a target sound signal.

[0142] In a specific embodiment disclosed in the present application, the analysis unit includes:

[0143] An encoding unit is used to extract features from the sub-signals at each scale and encode them to obtain multiple feature matrices, where the feature matrix includes a first matrix and multiple second matrices, the features corresponding to the first matrix are peak features, and the features corresponding to the second matrix are trough features;

[0144] A correlation unit, used for calculating the correlation value of each pair of feature data in a feature matrix corresponding to sub-signals of adjacent scales based on a grey correlation analysis algorithm, wherein each pair of feature data is feature data located at the same time scale in different feature matrices;

[0145] As a unit, it is used to regard each feature data in the feature matrix corresponding to the sub-signals of adjacent scales as a node, and regard the correlation value of each pair of feature data as the edge between two nodes;

[0146] The convolution unit is used to perform graph convolution operations based on the structural feature extraction neural network algorithm, pass the feature information of each node to the adjacent nodes, and update the features in combination with the information of the adjacent nodes to obtain the updated feature node information;

[0147] A second reconstruction unit is used to perform integration and reconstruction processing based on the characteristic node information to obtain a second sound signal;

[0148] The second decomposition unit is used to decompose the second sound signal based on wavelet decomposition to obtain an approximate coefficient and a plurality of detail coefficients.

[0149] In a specific implementation disclosed in the present application, the first calculation unit 50 includes:

[0150] a fifth determination unit, which determines a priori parameters and weight parameters based on historical arcing data, wherein the a priori parameters include a priori probability of arcing occurring, a priori probability of arcing not occurring, a response probability of a sound signal when arcing occurs, a response probability of a light signal when arcing occurs, a response probability of a sound signal when arcing does not occur, and a response probability of a light signal when arcing does not occur, and the weight parameters include a weight of the sound signal, a weight of the light signal, and a weight of the posterior probability;

[0151] a sixth determining unit, configured to determine a marginal probability value based on the priori parameter, the first data, and the second data;

[0152] a seventh determination unit, configured to determine a posterior probability of arcing occurrence based on the marginal probability value and the priori parameter;

[0153] The second calculation unit is used to calculate the response probability of the sound signal when the arcing occurs, the response probability of the light signal when the arcing occurs and the posterior probability of the arcing, and the sum of the products of the weight of the sound signal, the weight of the light signal and the weight of the posterior probability, respectively, to obtain the probability value of arcing occurrence.

[0154] In a specific implementation manner disclosed in the present application, the fourth determining unit includes:

[0155] A second acquisition unit is used to acquire a preset threshold value at each scale to obtain multiple initial threshold values;

[0156] A third calculation unit, used for calculating the absolute value of each detail coefficient to obtain an absolute detail coefficient;

[0157] A fourth calculation unit is used to calculate the difference between the absolute detail coefficient and the initial threshold value at the same scale to obtain a target difference;

[0158] A value taking unit, used for setting the optimal threshold value to 0 when the target difference value is not greater than a preset second setting threshold value;

[0159] The obtaining unit is used to calculate the product of the sign function value of the detail coefficient and the target difference value when the target difference value is greater than the second set threshold value, so as to obtain the optimal threshold value.

[0160] It should be noted that, regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment of the method, and will not be elaborated here.

[0161] Embodiment 3:

[0162] Corresponding to the above method embodiment, this embodiment also provides an arc monitoring device based on sound and light synergy. The arc monitoring device based on sound and light synergy described below and the arc monitoring method based on sound and light synergy described above can refer to each other.

[0163] Figure 3 FIG. 8 is a block diagram of an arc monitoring device 800 based on acoustic-optical coordination according to an exemplary embodiment. Figure 3 As shown, the arc monitoring device 800 based on acoustic-optical coordination may include: a processor 801 and a memory 802. The arc monitoring device 800 based on acoustic-optical coordination may also include one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0164] The processor 801 is used to control the overall operation of the arc monitoring device 800 based on sound and light coordination to complete all or part of the steps in the above-mentioned arc monitoring method based on sound and light coordination. The memory 802 is used to store various types of data to support the operation of the arc monitoring device 800 based on sound and light coordination, and these data may include, for example, instructions for any application or method operating on the arc monitoring device 800 based on sound and light coordination, as well as application-related data, such as contact data, sent and received messages, pictures, audio, video, etc. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, disk or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone, which is used to receive external audio signals. The received audio signal may be further stored in the memory 802 or sent via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules, and the above-mentioned other interface modules can be keyboards, mice, buttons, etc. These buttons can be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the arc monitoring device 800 based on sound and light coordination and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G or 4G, or a combination of one or more of them, so the corresponding communication component 805 can include: Wi-Fi module, Bluetooth module, NFC module.

[0165] In an exemplary embodiment, the arc monitoring device 800 based on sound and light coordination can be implemented by one or more application specific integrated circuits (Application Specific Integrated Circuit, referred to as ASIC), digital signal processors (Digital Signal Processor, referred to as DSP), digital signal processing devices (Digital Signal Processing Device, referred to as DSPD), programmable logic devices (Programmable Logic Device, referred to as PLD), field programmable gate arrays (Field Programmable Gate Array, referred to as FPGA), controllers, microcontrollers, microprocessors or other electronic components to execute the above-mentioned arc monitoring method based on sound and light coordination.

[0166] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, and when the program instructions are executed by a processor, the steps of the above-mentioned arcing monitoring method based on sound and light synergy are implemented. For example, the computer-readable storage medium can be the above-mentioned memory 802 including program instructions, and the above-mentioned program instructions can be executed by the processor 801 of the arcing monitoring device based on sound and light synergy 800 to complete the above-mentioned arcing monitoring method based on sound and light synergy.

[0167] Embodiment 4:

[0168] Corresponding to the above method embodiment, a readable storage medium is also provided in this embodiment. The readable storage medium described below and the arc monitoring method based on sound and light synergy described above can refer to each other.

[0169] A readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the arc monitoring method based on sound and light coordination of the above method embodiment are implemented.

[0170] The readable storage medium may specifically be a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or other readable storage medium that can store program codes.

[0171] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0172] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An arc monitoring method based on acoustic and optical synergy, characterized in that: include: Acquiring sound and light signals of the pantograph based on sensors; Performing multi-level denoising and signal reconstruction on the sound signal to obtain a target sound signal; Determine first data based on the target sound signal, the first data being the number of arcs and the time of each arc; Determine second data based on the optical signal, the second data being the number of arcs and the time of each arc; Based on the Bayesian formula, the first data and the second data, a probability value of arcing occurrence is calculated; An arcing result of the pantograph is determined based on the arcing occurrence probability value, and the arcing result is used to indicate whether arcing occurs in the pantograph.

2. The arc monitoring method based on sound and light synergy according to claim 1 is characterized in that , performing multi-level denoising and signal reconstruction on the sound signal to obtain a target sound signal, including: Performing filtering processing on the sound signal based on a filter to obtain a first sound signal; Decomposing the first sound signal based on wavelet decomposition to obtain a plurality of sub-signals of different scales; Based on the grey correlation analysis algorithm and wavelet decomposition, multiple sub-signals are analyzed to obtain approximate coefficients and multiple detail coefficients; Determine the optimal threshold for each scale based on the soft threshold function and multiple detail coefficients; Performing soft threshold processing on the detail coefficient of each scale based on the optimal threshold to obtain a plurality of processed target detail coefficients; Signal reconstruction is performed based on inverse wavelet transform, multiple target detail coefficients and the approximate coefficients to obtain the target sound signal.

3. The arc monitoring method based on sound and light synergy according to claim 2 is characterized in that ,Based on the grey correlation analysis algorithm and wavelet decomposition, multiple sub-signals are analyzed to obtain approximate coefficients and multiple detail coefficients, including: Extract features from the sub-signals at each scale and encode them to obtain multiple feature matrices, wherein the feature matrix includes a first matrix and multiple second matrices, the features corresponding to the first matrix are peak features, and the features corresponding to the second matrix are trough features; Calculating the correlation values ​​of the pairwise feature data in the feature matrix corresponding to the sub-signals of adjacent scales based on the grey correlation analysis algorithm, wherein the pairwise feature data are feature data located at the same time scale in different feature matrices; Each feature data in the feature matrix corresponding to the sub-signals of adjacent scales is regarded as a node, and the correlation value of each pair of feature data is regarded as an edge between two nodes; Based on the structural feature extraction neural network algorithm, graph convolution operation is performed to pass the feature information of each node to the adjacent nodes, and the feature is updated in combination with the information of the adjacent nodes to obtain the updated feature node information; Performing integration and reconstruction processing based on the characteristic node information to obtain a second sound signal; The second sound signal is decomposed based on the wavelet decomposition to obtain an approximate coefficient and a plurality of detail coefficients.

4. The arc monitoring method based on sound and light synergy according to claim 1 is characterized in that , based on the Bayesian formula, the first data and the second data, the arcing probability value is calculated, including: Determine a priori parameters and weight parameters based on historical arcing data, the a priori parameters include a priori probability of arcing occurring, a priori probability of arcing not occurring, a response probability of a sound signal when arcing occurs, a response probability of a light signal when arcing occurs, a response probability of a sound signal when arcing does not occur, and a response probability of a light signal when arcing does not occur, and the weight parameters include a weight of the sound signal, a weight of the light signal, and a weight of the a posteriori probability; Determining a marginal probability value based on the priori parameter, the first data, and the second data; Determining a posterior probability of arcing occurrence based on the marginal probability value and the priori parameter; The response probability of the sound signal when the arcing occurs, the response probability of the light signal when the arcing occurs and the posterior probability of the arcing occurrence are calculated, and the sum of the products of the weight of the sound signal, the weight of the light signal and the weight of the posterior probability are respectively multiplied to obtain the probability value of arcing occurrence.

5. An arc monitoring device based on sound and light synergy, characterized in that: include: A first acquisition unit, configured to acquire a sound signal and a light signal of the pantograph based on a sensor; A denoising and reconstruction unit, used for performing multi-level denoising and signal reconstruction on the sound signal to obtain a target sound signal; A first determining unit, configured to determine first data based on the target sound signal, wherein the first data is the number of arcs and the time of each arc; A second determining unit, configured to determine second data based on the optical signal, wherein the second data is the number of arcs and the time of each arc; A first calculation unit, configured to calculate an arcing occurrence probability value based on a Bayesian formula, the first data, and the second data; The third determining unit is used to determine the arcing result of the pantograph based on the arcing occurrence probability value, and the arcing result is used to indicate whether the pantograph has arcing.

6. The arc monitoring device based on sound and light coordination according to claim 5 is characterized in that: The denoising and reconstruction unit comprises: A filtering unit, configured to filter the sound signal based on a filter to obtain a first sound signal; A first decomposition unit, configured to decompose the first sound signal based on wavelet decomposition to obtain a plurality of sub-signals of different scales; An analysis unit, used for analyzing a plurality of sub-signals based on a grey correlation analysis algorithm and wavelet decomposition to obtain an approximate coefficient and a plurality of detail coefficients; A fourth determination unit, configured to determine an optimal threshold value for each scale based on the soft threshold function and a plurality of detail coefficients; A processing unit, configured to perform soft threshold processing on the detail coefficient of each scale based on the optimal threshold to obtain a plurality of processed target detail coefficients; The first reconstruction unit is used to perform signal reconstruction based on inverse wavelet transform, multiple target detail coefficients and the approximate coefficients to obtain the target sound signal.

7. The arc monitoring device based on sound and light synergy according to claim 6 is characterized in that: The analysis unit comprises: An encoding unit is used to extract features from the sub-signals at each scale and encode them to obtain multiple feature matrices, wherein the feature matrix includes a first matrix and multiple second matrices, the features corresponding to the first matrix are peak features, and the features corresponding to the second matrix are trough features; A correlation unit, used for calculating the correlation value of each pair of feature data in the feature matrix corresponding to the sub-signals of adjacent scales based on the grey correlation analysis algorithm, wherein the each pair of feature data is feature data located at the same time scale in different feature matrices; As a unit, it is used to regard each feature data in the feature matrix corresponding to the sub-signals of adjacent scales as a node, and regard the correlation value of each pair of feature data as an edge between two nodes; The convolution unit is used to perform graph convolution operations based on the structural feature extraction neural network algorithm, pass the feature information of each node to the adjacent nodes, and update the features in combination with the information of the adjacent nodes to obtain the updated feature node information; A second reconstruction unit, configured to perform integration and reconstruction processing based on the characteristic node information to obtain a second sound signal; The second decomposition unit is used to decompose the second sound signal based on the wavelet decomposition to obtain an approximate coefficient and a plurality of detail coefficients.

8. The arc monitoring device based on sound and light coordination according to claim 1 is characterized in that: The first computing unit comprises: a fifth determination unit, for determining a priori parameters and weight parameters based on historical arcing data, wherein the a priori parameters include a priori probability of arcing occurring, a priori probability of arcing not occurring, a response probability of a sound signal when arcing occurs, a response probability of a light signal when arcing occurs, a response probability of a sound signal when arcing does not occur, and a response probability of a light signal when arcing does not occur, and the weight parameters include a weight of a sound signal, a weight of a light signal, and a weight of a posterior probability; a sixth determining unit, configured to determine a marginal probability value based on the priori parameter, the first data, and the second data; a seventh determination unit, configured to determine a posterior probability of arcing occurrence based on the marginal probability value and the priori parameter; The second calculation unit is used to calculate the response probability of the sound signal when the arcing occurs, the response probability of the light signal when the arcing occurs and the posterior probability of the arcing, and the sum of the products of the weight of the sound signal, the weight of the light signal and the weight of the posterior probability, respectively, to obtain the probability value of arcing occurrence.

9. An arc monitoring device based on sound and light synergy, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the arc monitoring method based on sound and light synergy as claimed in any one of claims 1 to 4 when executing the computer program.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the arc monitoring method based on sound and light synergy as claimed in any one of claims 1 to 4 are implemented.

Images