Abnormal signal monitoring method and device of vibrating conveyor, electronic equipment and medium

Abstract

The invention provides an abnormal signal monitoring method and device for a vibrating conveyor, electronic equipment and a medium, and the method comprises the steps: carrying out the discrete wavelet transformation of a vibration signal, and obtaining a signal feature vector; constructing a spatial filtering matrix based on the normal vibration signal and the abnormal vibration signal; based on the plurality of signal feature vectors and the spatial filtering matrix, extracting a plurality of feature values of the multi-path vibration signals; and inputting the plurality of feature values into a classification model, and obtaining an abnormal vibration signal judgment result output by the classification model. Local analysis of non-stationary vibration signals is realized through discrete wavelet transform, and the accuracy of determining signal feature vectors is improved. Accurate feature extraction of the signal feature vector is realized through the spatial filtering matrix. By combining discrete wavelet transform and a common spatial mode algorithm, efficient feature extraction of the vibration signal is realized, and the efficiency and accuracy of monitoring the abnormal signal of the vibrating conveyor are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular, to a method, device, electronic device and medium for monitoring abnormal signals of a vibrating conveyor. Background Technique

[0002] The vibrating trough conveyor (vibrating conveyor) in the cigarette making workshop is mainly used to spread tobacco leaves and cut tobacco evenly so that they can enter the next process smoothly. Among them, the vibrating conveyor is mainly used for heating and humidifying tobacco leaves and cut tobacco before drying equipment and feeding equipment. Its stable operation is related to the moisture control, flavor formation and feeding effect of tobacco leaves and cut tobacco. Since the vibrating conveyor has been in operation for many years, its mechanical structure has aged, which may cause the trough body to crack or the physical properties to fail to meet the design requirements, resulting in the accumulation of cut tobacco and the inability to convey it forward. Moreover, the vibrating conveyor has a closed cover plate and cannot be observed, which may cause the cut tobacco to agglomerate, ultimately leading to unstable wet bulb and moisture, and there are quality hazards.

[0003] The existing monitoring of vibrating conveyors mainly relies on manual regular inspections, and the monitoring efficiency is low. Summary of the Invention

[0004] The present invention provides a method, device, electronic device and medium for monitoring abnormal signals of a vibrating conveyor, so as to solve the defect of low monitoring efficiency of vibrating conveyors in the prior art, and achieve the improvement of the monitoring efficiency of abnormal signals of vibrating conveyors.

[0005] In a first aspect, the present invention provides a method for monitoring abnormal signals of a vibrating conveyor, including: performing discrete wavelet transform on each path of vibration signals of the vibrating conveyor to obtain an approximation component and a plurality of detail components, and constructing a signal feature vector of each path of vibration signals based on the approximation component and the plurality of detail components; constructing a spatial filtering matrix of the vibration signals based on the normal vibration signals and abnormal vibration signals of the conveyor; extracting a plurality of eigenvalues of the multi-path vibration signals based on the plurality of signal feature vectors and the spatial filtering matrix; inputting the plurality of eigenvalues into a classification model to obtain a judgment result of abnormal vibration signals output by the classification model; wherein, the classification model is trained based on a preset model and the labels of sample feature vectors and sample abnormal vibration signal judgment results.

[0006] According to the abnormal signal monitoring method of a vibrating conveyor provided by the present invention, a spatial filtering matrix of vibration signals is constructed based on the normal vibration signals and abnormal vibration signals of the conveyor, including: obtaining a first covariance matrix of the normal vibration signals and a second covariance matrix of the abnormal vibration signals, and obtaining a third covariance matrix of the mixed space based on the sum of the first covariance matrix and the second covariance matrix; performing eigenvalue decomposition on the third covariance matrix to obtain an eigenvector matrix and an eigenvalue diagonal matrix; determining whitening processing parameters based on the eigenvector matrix and the eigenvalue diagonal matrix; performing whitening processing on the first covariance matrix based on the whitening processing parameters to obtain a first whitening processing result, and performing whitening processing on the second covariance matrix based on the whitening processing parameters to obtain a second whitening processing result; performing eigenvalue decomposition on the first whitening processing result and the second whitening processing result to obtain common eigenvectors, the sum of the first diagonal matrix corresponding to the common eigenvectors and the second diagonal matrix corresponding to the common eigenvectors is the identity matrix, the first diagonal matrix is obtained based on the eigenvalue decomposition result of the first whitening processing result, and the second diagonal matrix is obtained based on the eigenvalue decomposition result of the second whitening processing result; obtaining a spatial filtering matrix based on the common eigenvectors and the whitening processing parameters.

[0007] According to the abnormal signal monitoring method of a vibrating conveyor provided by the present invention, determining whitening processing parameters based on the eigenvector matrix and the eigenvalue diagonal matrix includes: sorting the rows and columns of the eigenvector matrix in descending order to obtain a target eigenvector matrix, and sorting the rows and columns of the eigenvalue diagonal matrix in descending order to obtain a target eigenvalue diagonal matrix; obtaining a target transpose matrix of the target eigenvector matrix, and obtaining a target square root of the target eigenvalue diagonal matrix; obtaining the whitening processing parameters based on the ratio of the target transpose matrix and the target square root.

[0008] According to the abnormal signal monitoring method of a vibrating conveyor provided by the present invention, based on multiple signal eigenvectors and the spatial filtering matrix, extracting multiple eigenvalues of multiple vibration signals includes: obtaining a filtered signal feature matrix based on the product of the signal eigenvector matrix and the spatial filtering matrix, and the signal eigenvector matrix is a matrix composed of multiple signal eigenvectors; sorting each row of filtered signal feature vectors in the filtered signal feature matrix, and selecting the row of filtered signal feature vectors from front to back as the first target filtered signal feature vector according to the sorting result, and selecting the row of filtered signal feature vectors from back to front as the second target filtered signal feature vector, and the total number of rows of the filtered signal feature vectors in the filtered signal feature matrix is greater than ; determining the ratio of the target feature variance of each first target filtered signal feature vector or second target filtered signal feature vector to the total feature variance to obtain multiple eigenvalues, and the total feature variance is the sum of all target feature variances.

[0009] According to the abnormal signal monitoring method of the vibrating conveyor provided by the present invention, discrete wavelet transform is performed on each channel of vibration signal of the vibrating conveyor to obtain an approximation component and multiple detail components, including: performing discrete wavelet transform on each channel of vibration signal to obtain an approximation component and all initial detail components; sorting all the initial detail components, and selecting multiple initial detail components as the multiple detail components based on the sorting result.

[0010] According to the abnormal signal monitoring method of the vibrating conveyor provided by the present invention, the classification model is constructed based on the following steps: obtaining the sample feature vectors of multiple channels of sample vibration signals of the vibrating conveyor and the labels of the judgment results of sample abnormal vibration signals; obtaining training samples based on the sample feature vectors and the labels of the judgment results of sample abnormal vibration signals; training and validating a preset model based on the training samples until the error of the preset model is less than a set threshold to obtain a classification model.

[0011] According to the abnormal signal monitoring method of the vibrating conveyor provided by the present invention, obtaining the first covariance matrix of the normal vibration signal includes: performing discrete wavelet transform on the normal vibration signal to obtain the normal signal feature vectors of the normal vibration signal; obtaining the normal filter signal feature matrix based on the product of the transposed vector of the normal signal feature vectors and the normal signal feature vectors; determining the matrix trace of the normal filter signal feature matrix based on the sum of the diagonal elements of the normal filter signal feature matrix; obtaining the first covariance matrix based on the ratio of the normal filter signal feature matrix to the matrix trace.

[0012] In a second aspect, the present invention further provides an abnormal signal monitoring device for a vibrating conveyor, including: a discrete wavelet transform module for performing discrete wavelet transform on each channel of vibration signal of the vibrating conveyor to obtain an approximation component and multiple detail components, and constructing a signal feature vector of each channel of vibration signal based on the approximation component and the multiple detail components; a spatial filter matrix determination module for constructing a spatial filter matrix of the vibration signal based on the normal vibration signal and the abnormal vibration signal of the conveyor; an eigenvalue determination module for extracting multiple eigenvalues of multiple channels of vibration signals based on the multiple signal feature vectors and the spatial filter matrix; a judgment module for inputting the multiple eigenvalues into the classification model to obtain the judgment result of the abnormal vibration signal output by the classification model; wherein, the classification model is trained based on the sample feature vectors and the labels of the judgment results of sample abnormal vibration signals on the basis of a preset model.

[0013] In a third aspect, the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, it implements any one of the above-mentioned abnormal signal monitoring methods of the vibrating conveyor.

[0014] Fourthly, the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the abnormal signal monitoring method of any one of the above vibration conveyors.

[0015] The abnormal signal monitoring method, device, electronic device and medium of the vibration conveyor provided by the present invention perform discrete wavelet transform on each path of vibration signal of the vibration conveyor to obtain an approximation component and multiple detail components, and construct a signal feature vector of each path of vibration signal based on the approximation component and multiple detail components; construct a spatial filtering matrix of the vibration signal based on the normal vibration signal and abnormal vibration signal of the conveyor; extract multiple eigenvalues of the multi-path vibration signal based on the multiple signal feature vectors and the spatial filtering matrix; input the multiple eigenvalues into a classification model to obtain a judgment result of the abnormal vibration signal output by the classification model; wherein, the classification model is trained based on the sample feature vector and the label of the sample abnormal vibration signal judgment result on the basis of a preset model. The local analysis of the non-stationary vibration signal is realized through discrete wavelet transform, and the accuracy of determining the signal feature vector is improved. The accurate feature extraction of the signal feature vector is realized through the spatial filtering matrix. By combining discrete wavelet transform and common spatial pattern algorithm, the efficient feature extraction of the vibration signal is realized, which is beneficial to improving the efficiency and accuracy of the abnormal signal monitoring of the vibration conveyor. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0017] Figure 1 It is a schematic flowchart of the abnormal signal monitoring method of the vibration conveyor provided by the present invention.

[0018] Figure 2 It is a schematic structural diagram of the abnormal signal monitoring device of the vibration conveyor provided by the present invention.

[0019] Figure 3 It is a schematic structural diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts fall within the protection scope of the present invention.

[0021] Vibration signals are generally processed using the discrete Fourier transform (DFT), which is to decompose the collected vibration signals into a combination of multiple different sine waves, transform the time-domain signals into frequency-domain signals, and then analyze the spectral characteristics of the signals. As a global transformation, the Fourier transform has certain limitations. For example, the Fourier transform does not have the ability of localized analysis and cannot analyze non-stationary signals, etc.

[0022] The wavelet transform (WT) can perform multi-scale refined analysis on vibration signals by using operations such as translation and dilation of basis functions, perform time-frequency analysis on the signals to obtain the time-frequency spectrum, and overcome the problems that the Fourier transform does not have the ability of localized analysis and cannot analyze non-stationary signals, etc.

[0023] The Common Spatial Pattern (CSP) algorithm uses the theory of simultaneous diagonalization of matrices in algebra to find a set of spatial filters, such that under the action of this set of filters, the variance of one type of vibration signal reaches a maximum, and the variance of another type of vibration signal reaches a minimum, so as to achieve the purpose of extracting eigenvalues.

[0024] In view of the deficiencies of the Fourier transform, the present invention combines the discrete wavelet transform and the common spatial pattern to process vibration signals.

[0025] The following combines Figures 1 - 3 to describe the abnormal signal monitoring method, device, and electronic equipment of the vibrating conveyor provided by the embodiments of the present invention.

[0026] Figure 1 is a schematic flowchart of the abnormal signal monitoring method of the vibrating conveyor provided by the present invention. As Figure 1 shown, the abnormal signal monitoring method of the vibrating conveyor includes steps S100 to step S400, and the specific steps are as follows.

[0027] S100: Perform discrete wavelet transform on each path of vibration signals of the vibrating conveyor to obtain an approximation component and multiple detail components, and construct a signal feature vector of each path of vibration signals based on the approximation component and the multiple detail components.

[0028] It should be noted that the execution subject of the embodiments of this application can be a server or a computer device, such as a mobile phone, a tablet computer, a notebook computer, a handheld computer, a vehicle-mounted electronic device, a wearable device, an ultra-mobile personal computer (UMPC), a netbook, or a personal digital assistant (PDA), etc.

[0029] Perform discrete wavelet transform on each path of vibration signal of the vibrating conveyor to obtain an approximation component and multiple detail components. Specifically, perform discrete wavelet transform on each path of vibration signal to obtain an approximation component and all initial detail components; sort all the initial detail components, and select multiple initial detail components as multiple detail components based on the sorting result.

[0030] Collect n paths of vibration signals of the vibrating conveyor. The expression formula of the vibration signal is as follows.

[0031] ; Where is the vibration signal set, is the th path of vibration signal, is the number of sampling points of each path of vibration signal.

[0032] Select the same wavelet basis, and perform discrete wavelet transform on each path of vibration signal to decompose it into an approximation component and an initial detail component. The expression formulas of the approximation component and the initial detail component are as follows.

[0033] ; Where is the initial signal feature vector corresponding to the th path of vibration signal, is the approximation component of the th path of vibration signal, is the th initial detail component of the th path of vibration signal, is the number of initial detail components.

[0034] Sort all the initial detail components, and select the top initial detail components with the top ranking according to the sorting result as multiple detail components. Construct a signal feature vector based on the approximation component and multiple detail components. The expression formula of the signal feature vector is as follows.

[0035] ; Where is the signal feature vector corresponding to the -th road vibration signal, is the approximate component of the -th road vibration signal, is the -th detail component of the , where is the number of detail components, .

[0036] S200: Construct a spatial filtering matrix of the vibration signal based on the normal vibration signal and the abnormal vibration signal of the conveyor.

[0037] Construct a spatial filtering matrix of the vibration signal based on the normal vibration signal and the abnormal vibration signal of the conveyor. Specifically, obtain the first covariance matrix of the normal vibration signal and the second covariance matrix of the abnormal vibration signal, and obtain the third covariance matrix of the mixed space based on the sum of the first covariance matrix and the second covariance matrix; perform eigenvalue decomposition on the third covariance matrix to obtain the eigenvector matrix and the eigenvalue diagonal matrix; determine the whitening processing parameter based on the eigenvector matrix and the eigenvalue diagonal matrix; perform whitening processing on the first covariance matrix based on the whitening processing parameter to obtain the first whitening processing result, and perform whitening processing on the second covariance matrix based on the whitening processing parameter to obtain the second whitening processing result; perform eigenvalue decomposition on the first whitening processing result and the second whitening processing result to obtain the common eigenvector, the sum of the first diagonal matrix corresponding to the common eigenvector and the second diagonal matrix corresponding to the common eigenvector is the identity matrix, the first diagonal matrix is obtained based on the eigenvalue decomposition result of the first whitening processing result, and the second diagonal matrix is obtained based on the eigenvalue decomposition result of the second whitening processing result; obtain the spatial filtering matrix based on the common eigenvector and the whitening processing parameter.

[0038] Obtain the first covariance matrix of the normal vibration signal, including: performing discrete wavelet transform on the normal vibration signal to obtain the normal signal feature vector of the normal vibration signal; obtaining the normal filtering signal feature matrix based on the product of the transposed vector of the normal signal feature vector and the normal signal feature vector; determining the matrix trace of the normal filtering signal feature matrix based on the sum of the diagonal elements of the normal filtering signal feature matrix; obtaining the first covariance matrix based on the ratio of the normal filtering signal feature matrix to the matrix trace.

[0039] Further, obtaining the second covariance matrix of the abnormal vibration signal includes: performing discrete wavelet transform on the abnormal vibration signal to obtain the abnormal signal feature vector of the abnormal vibration signal; obtaining the abnormal filtering signal feature matrix based on the product of the transposed vector of the abnormal signal feature vector and the abnormal signal feature vector; determining the matrix trace of the abnormal filtering signal feature matrix based on the sum of the diagonal elements of the abnormal filtering signal feature matrix; and obtaining the second covariance matrix based on the ratio of the abnormal filtering signal feature matrix to the matrix trace.

[0040] The normal vibration signal is the vibration signal extracted when the vibrating conveyor is in a normal state. The abnormal vibration signal is the vibration signal extracted when the vibrating conveyor is in an abnormal state. Update the abnormal state of the vibrating conveyor, and then update the second covariance.

[0041] The expression formula of the first covariance matrix is as follows.

[0042] ; Where, is the first covariance matrix, is the normal signal feature vector, is the transposed vector of the normal signal feature vector, is the normal filtering signal feature matrix, is the matrix trace of the normal filtering signal feature matrix.

[0043] Further, the expression formula of the second covariance matrix is as follows.

[0044] ; Where, is the second covariance matrix, is the abnormal signal feature vector, is the transposed vector of the abnormal signal feature vector, is the abnormal filtering signal feature matrix, is the matrix trace of the abnormal filtering signal feature matrix.

[0045] Add the first covariance matrix and the second covariance matrix to obtain the third covariance matrix of the mixed space. The calculation formula of the third covariance matrix is as follows.

[0046] ; Where, is the third covariance matrix, is the first covariance matrix, is the second covariance matrix.

[0047] Perform eigenvalue decomposition on the third covariance matrix to obtain the eigenvector matrix and the eigenvalue diagonal matrix. The expression formula of the eigenvalue decomposition of the third covariance matrix is as follows.

[0048] ; Among them, is the third covariance matrix, is the eigenvector matrix, is the eigenvalue diagonal matrix, is the transpose matrix of the eigenvector matrix.

[0049] Determine the whitening processing parameters based on the eigenvector matrix and the eigenvalue diagonal matrix. Specifically, sort the rows and columns of the eigenvector matrix in descending order to obtain the target eigenvector matrix, and sort the rows and columns of the eigenvalue diagonal matrix in descending order to obtain the target eigenvalue diagonal matrix; obtain the target transpose matrix of the target eigenvector matrix, and obtain the target square root of the target eigenvalue diagonal matrix; obtain the whitening processing parameters based on the ratio of the target transpose matrix and the target square root.

[0050] The calculation formula of the whitening processing parameters is as follows.

[0051] ; Among them, is the whitening processing parameter, is the target transpose matrix, is the target eigenvector matrix after sorting the rows and columns in descending order, is the target eigenvalue diagonal matrix after sorting the rows and columns in descending order, is the target square root.

[0052] The expression formula of the whitening processing is as follows.

[0053] ; Among them, is the first covariance matrix, is the second covariance matrix, is the first whitening processing result, is the second whitening processing result, is the whitening processing parameter.

[0054] Perform the eigenvalue decomposition on the first whitening processing result and the second whitening processing result to obtain the common eigenvector. The expression formula of the common eigenvector is as follows.

[0055] ; Among them, is the common eigenvector, is the first whitening processing result, is the second whitening processing result, is the first diagonal matrix, is the second diagonal matrix, is the identity matrix.

[0056] A spatial filtering matrix is obtained based on the common eigenvector and the whitening processing parameter. The calculation formula of the spatial filtering matrix is as follows.

[0057] ; where is the spatial filtering matrix, is the transposed vector of the common eigenvector, is the whitening processing parameter.

[0058] S300: Based on multiple signal eigenvectors and the spatial filtering matrix, extract multiple eigenvalues of multiple vibration signals.

[0059] Based on multiple signal eigenvectors and the spatial filtering matrix, extract multiple eigenvalues of multiple vibration signals. Specifically, based on the product of the signal eigenvector matrix and the spatial filtering matrix, a filtered signal eigenmatrix is obtained. The signal eigenvector matrix is a matrix composed of multiple signal eigenvectors; sort each row of the filtered signal eigenvectors in the filtered signal eigenmatrix, and select rows of the filtered signal eigenvectors from the front to the back as the first target filtered signal eigenvectors, and select rows of the filtered signal eigenvectors from the back to the front as the second target filtered signal eigenvectors. The total number of rows of the filtered signal eigenvectors in the filtered signal eigenmatrix is greater than ; Determine the ratio of the target eigenvariance to the total eigenvariance of each first target filtered signal eigenvector or second target filtered signal eigenvector to obtain multiple eigenvalues. The total eigenvariance is the sum of all target eigenvariances.

[0060] The expression formula of the filtered signal eigenmatrix is as follows.

[0061] ; where is the filtered signal eigenmatrix, is the spatial filtering matrix, is the th signal eigenvector corresponding to the th vibration signal, is the total number of rows of the filtered signal eigenmatrix, is the th signal eigenvector corresponding to the th vibration signal,

[0062] Based on the rows of the filtered signal eigenvectors with the highest ranks in the filtered signal eigenmatrix and the The row filtering signal feature vectors are used to obtain a set of signal feature vectors. The expression formula of the set of feature vectors is as follows.

[0063] ; Where, is the set of target filtering signal feature vectors, is the th first target filtering signal feature vector or the th second target filtering signal feature vector, is the total number of columns of the filtering signal feature matrix, is the number of first target filtering signal feature vectors and second target filtering signal feature vectors, is the total number of rows of the filtering signal feature vectors in the filtering signal feature matrix.

[0064] The calculation formula of the eigenvalue is as follows.

[0065] ; Where, is the th first target filtering signal feature vector or the th second target filtering signal feature vector, is the number of first target filtering signal feature vectors and second target filtering signal feature vectors, is the th first target filtering signal feature vector or the th second target filtering signal feature vector's variance, is the sum of all target feature variances.

[0066] Multiple eigenvalues form an eigenvalue set. The expression formula of the eigenvalue set is as follows.

[0067] ; Where, is the eigenvalue set, is the th eigenvalue, is the number of eigenvalues.

[0068] S400: Input multiple eigenvalues into the classification model to obtain the judgment result of the abnormal vibration signal output by the classification model.

[0069] Among them, the classification model is trained based on the sample feature vectors and the labels of the sample abnormal vibration signal judgment results on the basis of a preset model.

[0070] The classification model is constructed based on the following steps: obtaining the sample feature vectors of the multi-channel sample vibration signals of the vibrating conveyor and the labels of the judgment results of the sample abnormal vibration signals; obtaining the training samples based on the sample feature vectors and the labels of the judgment results of the sample abnormal vibration signals; training and validating a preset model based on the training samples until the error of the preset model is less than a set threshold to obtain the classification model.

[0071] Obtain the sample feature vectors of the multi-channel sample vibration signals of the vibrating conveyor and the labels of the judgment results of the sample abnormal vibration signals. Mark the sample feature vectors according to the labels of the judgment results of the sample abnormal vibration signals to obtain the sample feature vectors carrying labels. Train and validate a preset model according to the sample feature vectors carrying labels to obtain the classification model. The preset model includes classification models such as support vector machines, decision trees, and logistic regression.

[0072] The abnormal signal monitoring method for the vibrating conveyor provided by the embodiment of the present invention obtains an approximation component and multiple detail components by performing discrete wavelet transform on each channel of vibration signal of the vibrating conveyor, and constructs a signal feature vector for each channel of vibration signal based on the approximation component and the multiple detail components; constructs a spatial filtering matrix of the vibration signals based on the normal vibration signals and abnormal vibration signals of the conveyor; extracts multiple eigenvalues of the multi-channel vibration signals based on the multiple signal feature vectors and the spatial filtering matrix; inputs the multiple eigenvalues into the classification model to obtain the judgment result of the abnormal vibration signal output by the classification model; wherein, the classification model is trained based on the sample feature vectors and the labels of the judgment results of the sample abnormal vibration signals on the basis of the preset model. The local analysis of the non-stationary vibration signals is realized through discrete wavelet transform, which improves the accuracy of determining the signal feature vectors. The accurate feature extraction of the signal feature vectors is realized through the spatial filtering matrix. By combining discrete wavelet transform and common spatial pattern algorithm, the efficient feature extraction of the vibration signals is realized, which is beneficial to improving the efficiency and accuracy of the abnormal signal monitoring of the vibrating conveyor.

[0073] The embodiment of the present invention also provides an abnormal signal monitoring device for a vibrating conveyor, as Figure 2 shown Figure 2 is a schematic structural diagram of the abnormal signal monitoring device for the vibrating conveyor provided by the present invention. It should be noted that the abnormal signal monitoring device for the vibrating conveyor provided by the embodiment of the present invention can execute the abnormal signal monitoring method for the vibrating conveyor described in any of the above embodiments during specific operation, and this embodiment will not be elaborated here.

[0074] Refer to Figure 3, an embodiment of the present invention provides an abnormal signal monitoring device for a vibrating conveyor, including: a discrete wavelet transform module 201, configured to perform discrete wavelet transform on each vibration signal of the vibrating conveyor to obtain an approximation component and a plurality of detail components, and construct a signal feature vector of each vibration signal based on the approximation component and the plurality of detail components.

[0075] A spatial filtering matrix determination module 202, configured to construct a spatial filtering matrix of the vibration signal based on the normal vibration signal and the abnormal vibration signal of the conveyor.

[0076] An eigenvalue determination module 203, configured to extract a plurality of eigenvalues of the multi-channel vibration signals based on the plurality of signal feature vectors and the spatial filtering matrix.

[0077] A judgment module 204, configured to input the plurality of eigenvalues into a classification model to obtain a judgment result of the abnormal vibration signal output by the classification model; wherein, the classification model is trained based on the sample feature vector and the label of the sample abnormal vibration signal judgment result on the basis of a preset model.

[0078] The abnormal signal monitoring device for a vibrating conveyor provided by the embodiment of the present invention performs discrete wavelet transform on each vibration signal of the vibrating conveyor to obtain an approximation component and a plurality of detail components, constructs a signal feature vector of each vibration signal based on the approximation component and the plurality of detail components; constructs a spatial filtering matrix of the vibration signal based on the normal vibration signal and the abnormal vibration signal of the conveyor; extracts a plurality of eigenvalues of the multi-channel vibration signals based on the plurality of signal feature vectors and the spatial filtering matrix; inputs the plurality of eigenvalues into a classification model to obtain a judgment result of the abnormal vibration signal output by the classification model; wherein, the classification model is trained based on the sample feature vector and the label of the sample abnormal vibration signal judgment result on the basis of a preset model. Through discrete wavelet transform, local analysis of non-stationary vibration signals is realized, and the accuracy of determining signal feature vectors is improved. Through the spatial filtering matrix, accurate feature extraction of signal feature vectors is realized. By combining discrete wavelet transform and common spatial pattern algorithm, efficient feature extraction of vibration signals is realized, which is beneficial to improving the efficiency and accuracy of abnormal signal monitoring of vibrating conveyors.

[0079] In one embodiment, the spatial filtering matrix determination module 202 is configured to: obtain a first covariance matrix of normal vibration signals and a second covariance matrix of abnormal vibration signals, and obtain a third covariance matrix of the mixed space based on the sum of the first covariance matrix and the second covariance matrix; perform eigenvalue decomposition on the third covariance matrix to obtain an eigenvector matrix and an eigenvalue diagonal matrix; determine whitening processing parameters based on the eigenvector matrix and the eigenvalue diagonal matrix; perform whitening processing on the first covariance matrix based on the whitening processing parameters to obtain a first whitening processing result, and perform whitening processing on the second covariance matrix based on the whitening processing parameters to obtain a second whitening processing result; perform eigenvalue decomposition on the first whitening processing result and the second whitening processing result to obtain common eigenvectors, the sum of the first diagonal matrix corresponding to the common eigenvectors and the second diagonal matrix corresponding to the common eigenvectors is the identity matrix, the first diagonal matrix is obtained based on the eigenvalue decomposition result of the first whitening processing result, and the second diagonal matrix is obtained based on the eigenvalue decomposition result of the second whitening processing result; obtain a spatial filtering matrix based on the common eigenvectors and the whitening processing parameters.

[0080] In one embodiment, the spatial filtering matrix determination module 202 is configured to: perform descending sorting on the rows and columns of the eigenvector matrix to obtain a target eigenvector matrix, and perform descending sorting on the rows and columns of the eigenvalue diagonal matrix to obtain a target eigenvalue diagonal matrix; obtain a target transpose matrix of the target eigenvector matrix, and obtain a target square root of the target eigenvalue diagonal matrix; obtain whitening processing parameters based on the ratio of the target transpose matrix to the target square root.

[0081] In one embodiment, the eigenvalue determination module 203 is configured to: obtain a filtered signal feature matrix based on the product of a signal feature vector matrix and a spatial filtering matrix, where the signal feature vector matrix is a matrix composed of multiple signal feature vectors; sort each row of filtered signal feature vectors in the filtered signal feature matrix, and select the row of filtered signal feature vectors from front to back as the first target filtered signal feature vector according to the sorting result, and select the row of filtered signal feature vectors from back to front as the second target filtered signal feature vector, and the total number of rows of the filtered signal feature vectors in the filtered signal feature matrix is greater than ; determine the ratio of the target feature variance of each first target filtered signal feature vector or second target filtered signal feature vector to the total feature variance to obtain multiple eigenvalues, where the total feature variance is the sum of all target feature variances.

[0082] In one embodiment, the discrete wavelet transform module 201 is configured to: perform discrete wavelet transform on each vibration signal to obtain an approximation component and all initial detail components; sort all the initial detail components, and select multiple initial detail components as multiple detail components based on the sorting result.

[0083] In one embodiment, the judgment module 204 is configured to: obtain the sample feature vector of the multi-channel sample vibration signals of the vibrating conveyor and the label of the sample abnormal vibration signal judgment result; obtain training samples based on the sample feature vector and the label of the sample abnormal vibration signal judgment result; train and verify a preset model based on the training samples until the error of the preset model is less than a set threshold, so as to obtain a classification model.

[0084] In one embodiment, the spatial filtering matrix determination module 202 is configured to: perform discrete wavelet transform on the normal vibration signal to obtain the normal signal feature vector of the normal vibration signal; obtain the normal filtering signal feature matrix based on the product of the transposed vector of the normal signal feature vector and the normal signal feature vector; determine the matrix trace of the normal filtering signal feature matrix based on the sum of the diagonal elements of the normal filtering signal feature matrix; obtain the first covariance matrix based on the ratio of the normal filtering signal feature matrix to the matrix trace.

[0085] Figure 3 is a schematic structural diagram of the electronic device provided by the present invention. As Figure 3 shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330, and a communication bus 340. Among them, the processor 310, the communication interface 320, and the memory 330 communicate with each other through the communication bus 340. The processor 310 can call the logical instructions in the memory 330 to execute the abnormal signal monitoring method of the vibrating conveyor, and the method includes: performing discrete wavelet transform on each vibration signal of the vibrating conveyor to obtain an approximation component and multiple detail components, and constructing a signal feature vector of each vibration signal based on the approximation component and the multiple detail components; constructing a spatial filtering matrix of the vibration signals based on the normal vibration signals and the abnormal vibration signals of the conveyor; extracting multiple eigenvalues of the multi-channel vibration signals based on the multiple signal feature vectors and the spatial filtering matrix; inputting the multiple eigenvalues into the classification model to obtain the abnormal vibration signal judgment result output by the classification model; wherein, the classification model is obtained by training based on the sample feature vector and the label of the sample abnormal vibration signal judgment result on the basis of a preset model.

[0086] In addition, when the logical instructions in the above-mentioned memory 330 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0087] In another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is configured to execute the abnormal signal monitoring method of the vibrating conveyor provided in the above-mentioned embodiments. The method includes: performing discrete wavelet transform on each path of vibration signals of the vibrating conveyor to obtain an approximation component and multiple detail components, and constructing a signal feature vector of each path of vibration signals based on the approximation component and the multiple detail components; constructing a spatial filtering matrix of the vibration signals based on the normal vibration signals and abnormal vibration signals of the conveyor; extracting multiple eigenvalues of multiple paths of vibration signals based on the multiple signal feature vectors and the spatial filtering matrix; inputting the multiple eigenvalues into a classification model to obtain a judgment result of abnormal vibration signals output by the classification model; wherein the classification model is trained based on the sample feature vectors and the labels of the sample abnormal vibration signal judgment results on the basis of a preset model.

[0088] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative labor.

[0089] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring abnormal signals of a vibrating conveyor, characterized in that: include: Performing discrete wavelet transform on each vibration signal of the vibration conveyor to obtain an approximate component and a plurality of detail components, and constructing a signal feature vector of each vibration signal based on the approximate component and the plurality of detail components; Constructing a spatial filter matrix of the vibration signal based on the normal vibration signal and the abnormal vibration signal of the conveyor; Extracting multiple eigenvalues ​​of multiple vibration signals based on the multiple signal eigenvectors and the spatial filter matrix; Inputting the multiple characteristic values ​​into a classification model to obtain an abnormal vibration signal judgment result output by the classification model; The classification model is obtained by training based on a preset model and based on the sample feature vector and the label of the sample abnormal vibration signal judgment result.

2. The abnormal signal monitoring method of a vibrating conveyor according to claim 1, characterized in that: The step of constructing a spatial filter matrix of the vibration signal based on the normal vibration signal and the abnormal vibration signal of the conveyor comprises: Acquire a first covariance matrix of the normal vibration signal and a second covariance matrix of the abnormal vibration signal, and obtain a third covariance matrix of the mixed space based on the sum of the first covariance matrix and the second covariance matrix; Performing eigenvalue decomposition on the third covariance matrix to obtain an eigenvector matrix and an eigenvalue diagonal matrix; Determine whitening processing parameters based on the eigenvector matrix and the eigenvalue diagonal matrix; Performing whitening processing on the first covariance matrix based on the whitening processing parameters to obtain a first whitening processing result, and performing whitening processing on the second covariance matrix based on the whitening processing parameters to obtain a second whitening processing result; Performing the eigenvalue decomposition on the first whitening processing result and the second whitening processing result to obtain a common eigenvector, the sum of a first diagonal matrix corresponding to the common eigenvector and a second diagonal matrix corresponding to the common eigenvector is a unit matrix, the first diagonal matrix is ​​obtained based on the eigenvalue decomposition result of the first whitening processing result, and the second diagonal matrix is ​​obtained based on the eigenvalue decomposition result of the second whitening processing result; The spatial filter matrix is ​​obtained based on the common eigenvector and the whitening processing parameter.

3. The abnormal signal monitoring method of a vibrating conveyor according to claim 2, characterized in that: The determining of whitening processing parameters based on the eigenvector matrix and the eigenvalue diagonal matrix comprises: Sorting the rows and columns in the eigenvector matrix in descending order to obtain a target eigenvector matrix, and sorting the rows and columns in the eigenvalue diagonal matrix in descending order to obtain a target eigenvalue diagonal matrix; Obtain a target transposed matrix of the target eigenvector matrix, and obtain a target square root of the target eigenvalue diagonal matrix; The whitening processing parameter is obtained based on a ratio of the target transposed matrix and the target square root.

4. The abnormal signal monitoring method of a vibrating conveyor according to claim 1, characterized in that: The extracting multiple eigenvalues ​​of the multiple vibration signals based on the multiple signal eigenvectors and the spatial filter matrix comprises: Based on the product of the signal feature vector matrix and the spatial filter matrix, a filtering signal feature matrix is ​​obtained, wherein the signal feature vector matrix is ​​a matrix composed of a plurality of the signal feature vectors; Sort each row of the filter signal feature vectors in the filter signal feature matrix, and select from the front to the back according to the sorting result. The filtered signal feature vector is used as the first target filtered signal feature vector, and the filter signal feature vector is selected from the back to the front according to the sorting result. The filtered signal feature vector is used as the second target filtered signal feature vector, and the total number of rows of the filtered signal feature vectors of the filtered signal feature matrix is ​​greater than ; Determine the ratio of the target feature variance to the total feature variance of each of the first target filtered signal feature vectors or the second target filtered signal feature vectors to obtain a plurality of the feature values, wherein the total feature variance is the sum of all the target feature variances.

5. The abnormal signal monitoring method of a vibrating conveyor according to claim 1, characterized in that: The discrete wavelet transform is performed on each vibration signal of the vibrating conveyor to obtain an approximate component and multiple detail components, including: Performing discrete wavelet transform on each vibration signal to obtain the approximate component and all initial detail components; All the initial detail components are sorted, and a plurality of the initial detail components are selected as the plurality of detail components based on the sorting result.

6. The abnormal signal monitoring method of a vibration conveyor according to claim 1, characterized in that: The classification model is constructed based on the following steps: Obtaining sample feature vectors of multi-channel sample vibration signals of the vibrating conveyor and labels of sample abnormal vibration signal judgment results; Obtaining a training sample based on the sample feature vector and the label of the sample abnormal vibration signal judgment result; The preset model is trained and verified based on the training samples until the error of the preset model is less than a set threshold, thereby obtaining the classification model.

7. The abnormal signal monitoring method of a vibrating conveyor according to claim 2, characterized in that: The obtaining of a first covariance matrix of the normal vibration signal comprises: Performing the discrete wavelet transform on the normal vibration signal to obtain a normal signal feature vector of the normal vibration signal; Obtaining a normal filtering signal feature matrix based on the product of the transposed vector of the normal signal feature vector and the normal signal feature vector; determining a matrix trace of the normal filtered signal feature matrix based on a sum of diagonal elements of the normal filtered signal feature matrix; The first covariance matrix is ​​obtained based on a ratio of the normal filtered signal feature matrix and the matrix trace.

8. An abnormal signal monitoring device for a vibrating conveyor, characterized in that: include: A discrete wavelet transform module is used to perform discrete wavelet transform on each vibration signal of the vibration conveyor to obtain an approximate component and a plurality of detail components, and construct a signal feature vector of each vibration signal based on the approximate component and the plurality of detail components; A spatial filter matrix determination module, used for constructing a spatial filter matrix of the vibration signal based on the normal vibration signal and the abnormal vibration signal of the conveyor; An eigenvalue determination module, used for extracting multiple eigenvalues ​​of multiple vibration signals based on multiple signal eigenvectors and the spatial filter matrix; A judgment module is used to input the multiple feature values ​​into a classification model to obtain an abnormal vibration signal judgment result output by the classification model; wherein the classification model is obtained by training based on a preset model, based on a sample feature vector and a label of the sample abnormal vibration signal judgment result.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the abnormal signal monitoring method for a vibrating conveyor according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the abnormal signal monitoring method for a vibrating conveyor according to any one of claims 1 to 7 is implemented.

Images