Non-stationary data enhancement method, device, electronic device and computer storage medium

By decomposing and reconstructing the non-steady state vibration signals of the dock equipment, predicting energy parameters, data enhancement under high noise is achieved, and the problem of insufficient significance of non-steady state data characteristics in the fault diagnosis of dock equipment is solved, and the problem of improved Diagnostic efficiency and accuracy.

CN118427599BActive Publication Date: 2024-11-22LEVI INTELLIGENT (SHENZHEN) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410889687.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-04
Publication Date
2024-11-22
Estimated Expiration
2044-07-04

AI Technical Summary

Technical Problem

In dock equipment scenarios, due to the alternation of equipment dynamics and statics and rapid changes in motor speed, fault impact characteristics do not comply with fixed cycles, traditional fault information enhancement algorithms cannot be effectively applied, and the diagnosis difficulty in high noise environments increases, resulting in insufficient enhancement of non-steady state data characteristics. Significant.

Method used

By obtaining the non-steady state original vibration signal of the target device, performing feature decomposition to obtain eigenvalues ​​and eigenvectors, characterizing the signal change direction and amplitude, reconstructing the main signal components, predicting energy parameters, and enhancing data based on these parameters to generate enhanced signals to generate enhanced signals data.

Benefits of technology

Accurately characterize the energy intensity of the instantaneous pulse of the signal in a high noise environment, improve the efficiency and accuracy of mechanical fault diagnosis, good noise resistance and low complexity, and is suitable for hardware terminal deployment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118427599B_ABST
    Figure CN118427599B_ABST
Patent Text Reader

Abstract

The present application relates to the field of signal processing technology, and discloses a non-steady-state data enhancement method, device, electronic device, and computer storage medium, including: obtaining multiple non-steady-state original vibration signals of a target device in a target time period; performing eigendecomposition on the vibration signal matrix corresponding to the original vibration signal to obtain eigenvalues ​​and eigenvectors corresponding to the eigenvalues; reconstructing the original vibration signal of the target time period according to the change direction corresponding to the eigenvector, the signal change amplitude corresponding to the eigenvalue, and the vibration signal matrix to obtain the main signal components of the target device in the target time period; predicting the energy of the main signal components in the target time period to obtain the target energy parameters; performing data enhancement on the original vibration signal based on the target energy parameters to obtain enhanced signal data. The present application realizes accurate data enhancement of non-steady-state original vibration signals under high noise, further improving the diagnostic efficiency and accuracy of mechanical faults.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of signal processing technology, and in particular to a non-steady-state data enhancement method, device, electronic device and computer-readable storage medium. Background Art

[0002] The port industry has always been an equipment-intensive industry, with large-scale port machinery such as quay cranes, yard cranes, rail cranes and gantry cranes being widely used. These devices are often complex in structure, with numerous power mechanisms and functional units and extremely heavy loads. In order to effectively ensure the normal and safe operation of the equipment, the predictive maintenance system of terminal equipment is increasingly valued. By installing acceleration sensors at key locations of the equipment and collecting vibration signal data of mechanical equipment at high frequency, the predictive maintenance system can issue warnings before serious damage to the equipment occurs, and guide users to perform maintenance operations, thereby reducing the risk of unplanned equipment failures and downtime, and reducing economic losses.

[0003] However, unlike other industrial scenarios where equipment operates stably for a long time, terminal equipment is generally operated by operators, and one operation can control multiple mechanisms. Acceleration, deceleration, and start and stop of power equipment are common, and load and no-load conditions appear alternately. The impact is that it is difficult for the system to collect data on the stable operation of the mechanism, so that the typical impact characteristics of equipment failure are submerged in system noise and non-stationary data, increasing the difficulty of diagnosis.

[0004] In the fault diagnosis of typical equipment, since general fault pulses are presented in a fixed periodic manner, the design idea of ​​most existing data enhancement algorithms is to amplify the periodic components in the signal as much as possible and suppress the non-periodic components. These methods are well suited for processing scenarios where vibration data is stable and does not change drastically. However, in the port equipment scenario, since the motor speed changes very quickly, the equipment often alternates between movement and stillness, and the spectrum components of the observed data change over time, the typical fault impact characteristics do not follow a fixed period, which leads to the fact that traditional fault information enhancement algorithms are not well applied in the terminal equipment diagnosis scenario. In addition, since the distance between multiple mechanisms of terminal equipment (such as lifting mechanism, pitch mechanism and trolley mechanism) is very close, the vibration data is seriously affected by the cross-operation between mechanisms, the observation noise is greater than the traditional diagnosis and analysis scenario, and the traditional method cannot work in this high-noise environment. Summary of the invention

[0005] The embodiments of the present application provide a non-stationary data enhancement method, device, electronic device and computer-readable storage medium, which can solve the technical problem that the current non-stationary data feature enhancement is not significant enough.

[0006] The present application provides a non-stationary data enhancement method, including:

[0007] Acquire multiple non-steady-state original vibration signals of the target device in a target period;

[0008] Performing eigendecomposition on a vibration signal matrix corresponding to the original vibration signal to obtain an eigenvalue and an eigenvector corresponding to the eigenvalue, wherein the eigenvector is used to characterize a change direction of the original vibration signal in the target time period, and the eigenvalue is used to characterize a signal change amplitude of the original vibration signal in the change direction in the target time period;

[0009] Reconstructing the original vibration signal of the target time period according to the change direction, the signal change amplitude and the vibration signal matrix to obtain the main signal components of the target device in the target time period;

[0010] Predicting the energy of the main signal component in the target time period to obtain a target energy parameter;

[0011] The original vibration signal is data enhanced based on the target energy parameter to obtain enhanced signal data.

[0012] Furthermore, before performing eigendecomposition on the vibration signal matrix corresponding to the original vibration signal to obtain eigenvalues ​​and eigenvectors corresponding to the eigenvalues, the method further includes:

[0013] Sorting the original vibration signals according to the collection time corresponding to each of the original vibration signals in the target time period to generate a vibration signal sequence;

[0014] Obtaining the moving step size and window length corresponding to the target time period;

[0015] An interference signal is determined according to the moving step length and the window length, and the vibration signal sequence is dimensionally transformed based on the interference signal and the original vibration signal to generate the vibration signal matrix.

[0016] Furthermore, the above-mentioned performing eigendecomposition on the vibration signal matrix corresponding to the original vibration signal to obtain eigenvalues ​​and eigenvectors corresponding to the eigenvalues ​​includes:

[0017] According to the positions corresponding to the original vibration signal and the interference signal in the vibration signal matrix, the similarity between the signals at any two positions is calculated, and a correlation matrix is ​​generated according to the similarity;

[0018] Perform eigenvalue decomposition on the correlation matrix to obtain the eigenvalues ​​and eigenvectors corresponding to the eigenvalues.

[0019] Furthermore, generating a correlation matrix according to the similarity includes:

[0020] Generate a correlation matrix to be solved according to the vibration signal matrix and a transposed matrix corresponding to the vibration signal matrix;

[0021] The parameters in the correlation matrix to be solved are solved according to the similarity to obtain the correlation matrix.

[0022] Furthermore, the above-mentioned reconstructing the original vibration signal of the target time period according to the change direction, the signal change amplitude and the vibration signal matrix to obtain the main signal components of the target device in the target time period includes:

[0023] According to the magnitude of the characteristic value, a target characteristic value is selected from the plurality of characteristic values;

[0024] Based on the signal change amplitude and change direction corresponding to the target eigenvalue, the vibration signal matrix is ​​mapped to obtain the main signal components of the target device in the target time period.

[0025] Furthermore, the above-mentioned predicting the energy of the main signal component in the target time period to obtain the target energy parameter includes:

[0026] Acquire a reference signal component of the target device in an adjacent time period, where the adjacent time period is a time period adjacent to the target time period, and the reference signal component is a signal component corresponding to the target device in the adjacent time period;

[0027] According to a preset energy prediction model, the reference signal component and the main signal component are fused to obtain the target energy parameter.

[0028] Furthermore, the data enhancement of the original vibration signal based on the target energy parameter to obtain enhanced signal data includes:

[0029] Performing noise filtering on the target energy parameter to obtain a filtered target energy parameter;

[0030] The original vibration signal is data enhanced according to the filtered target energy parameter to obtain the enhanced signal data.

[0031] Accordingly, an embodiment of the present application provides a non-steady-state data enhancement device, comprising:

[0032] An acquisition module, used to acquire a plurality of non-steady-state original vibration signals of a target device in a target period;

[0033] a decomposition module, configured to perform eigendecomposition on a vibration signal matrix corresponding to the original vibration signal to obtain an eigenvalue and an eigenvector corresponding to the eigenvalue, wherein the eigenvector is used to characterize a change direction of the original vibration signal in the target time period, and the eigenvalue is used to characterize a signal change amplitude of the original vibration signal in the change direction in the target time period;

[0034] A reconstruction module, used to reconstruct the original vibration signal of the target time period according to the change direction, the signal change amplitude and the vibration signal matrix, to obtain the main signal components of the target device in the target time period;

[0035] A prediction module, used for predicting the energy of the main signal component in the target time period to obtain a target energy parameter;

[0036] The enhancement module is used to perform data enhancement on the original vibration signal based on the target energy parameter to obtain enhanced signal data.

[0037] In addition, an embodiment of the present application also provides an electronic device, including a processor and a memory, wherein the memory stores a computer program, and the processor is used to run the computer program in the memory to implement the non-steady-state data enhancement method provided in an embodiment of the present application.

[0038] In addition, an embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. The computer program is suitable for loading by a processor to execute any non-steady-state data enhancement method provided in the embodiment of the present application.

[0039] In addition, an embodiment of the present application also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the non-steady-state data enhancement methods provided in the embodiments of the present application.

[0040] In an embodiment of the present application, multiple non-steady-state original vibration signals of a target device in a target time period are obtained; the vibration signal matrix corresponding to the original vibration signal is eigen-decomposed to obtain eigenvalues ​​and eigenvectors corresponding to the eigenvalues, the eigenvectors are used to characterize the change direction of the original vibration signal in the target time period, and the eigenvalues ​​are used to characterize the signal change amplitude of the original vibration signal in the change direction in the target time period, thereby realizing efficient decomposition of the vibration signal matrix; according to the change direction, signal change amplitude and vibration signal matrix, the original vibration signal of the target time period is reconstructed to obtain the main signal components of the target device in the target time period, thereby realizing accurate reconstruction of the original vibration signal; then, the energy of the main signal components in the target time period is predicted to obtain the target energy parameters, so that the energy intensity of the instantaneous pulse of the signal can be accurately characterized even under high noise interference; finally, the original vibration signal is data enhanced based on the target energy parameters to obtain enhanced signal data, thereby realizing accurate data enhancement of the non-steady-state original vibration signal under high noise, further improving the diagnostic efficiency and accuracy of mechanical faults, and having good anti-noise performance, low complexity, and being able to be flexibly deployed in hardware terminals. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0042] Figure 1 It is a flowchart of the non-steady-state data enhancement method provided by this application;

[0043] Figure 2 is a schematic diagram of an original vibration signal provided in an embodiment of the present application;

[0044] Figure 3 is a schematic diagram of a vibration signal after data enhancement provided in an embodiment of the present application;

[0045] Figure 4 is a schematic diagram of a third-order energy operator provided in an embodiment of the present application;

[0046] Figure 5 is a schematic diagram of the structure of a non-steady-state data enhancement device provided in an embodiment of the present application;

[0047] Figure 6 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0049] The embodiments of the present application provide a non-steady-state data enhancement method, device, electronic device and computer-readable storage medium. The non-steady-state data enhancement device can be integrated in an electronic device, which can be a server or a terminal.

[0050] Among them, the server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, network acceleration services (Content Delivery Network, CDN), as well as big data and artificial intelligence platforms.

[0051] The terminal may be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, etc., but is not limited thereto. The terminal and the server may be directly or indirectly connected via wired or wireless communication, and this application does not limit this.

[0052] In addition, the term "plurality" in the embodiments of the present application refers to two or more than two. The terms "first" and "second" in the embodiments of the present application are used to distinguish descriptions and should not be understood as implying relative importance.

[0053] It should be noted that the description order of the following embodiments is not intended to limit the preferred order of the embodiments.

[0054] See also Figure 1 , Figure 1 : is a flow chart of a non-stationary data enhancement method provided by an embodiment of the present application. The non-stationary data enhancement method may include:

[0055] S101. Acquire a plurality of non-steady-state original vibration signals of a target device in a target period of time.

[0056] In this embodiment, the target device is a source mechanism of the vibration signal, such as a lifting mechanism in a quay crane device in a dock scenario. Multiple original vibration signals of the target device in a target period can be obtained through a vibration signal sensor. The multiple original vibration signals are non-steady-state vibration signals collected at continuous collection moments, and the continuous collection moments constitute the target period. For example, the original vibration signal of the target period is collected from a faulty reducer of a lifting mechanism in a quay crane device in the dock through a vibration signal sensor. Among them, due to equipment failure, a pulse signal corresponding to the fault will be generated in the original vibration signal. In non-steady-state scenarios such as docks, the generated fault pulses often do not have periodicity. Therefore, there are often non-periodically generated fault pulse signals in the original vibration signal, as well as noise signals generated by alternating acceleration, deceleration, and start-stop of the equipment.

[0057] like Figure 2 As shown, Figure 2 Schematic diagram of the original vibration signal provided in the embodiment of the present application. The original vibration signal is a vibration signal collected by the vibration sensor at a sampling frequency of 51200 Hz, a vibration data length of K=163840, and a duration of 3.12s. The horizontal axis represents the number of sampling points, and the vertical axis represents the acceleration. Figure 2 As shown, due to equipment failure, there are multiple abnormal impact signals in the schematic diagram of the original vibration signal.

[0058] S102, performing eigendecomposition on the vibration signal matrix corresponding to the original vibration signal to obtain eigenvalues ​​and eigenvectors corresponding to the eigenvalues, wherein the eigenvectors are used to characterize the change direction of the original vibration signal in the target time period, and the eigenvalues ​​are used to characterize the signal change amplitude of the original vibration signal in the change direction in the target time period.

[0059] In this embodiment, the vibration signal matrix is ​​a matrix generated according to the original vibration signal. By performing eigendecomposition on the vibration signal matrix, eigenvalues ​​and eigenvectors can be obtained. Specifically, the eigendecomposition of the signal can adopt singular value decomposition, wavelet transform, etc. By performing eigendecomposition on the vibration signal matrix, the eigenvector obtained is used to characterize the signal change direction of the original vibration signal in the target period; the eigenvalue is used to characterize the signal change amplitude of the original vibration signal in the target period in the change direction.

[0060] S103: reconstruct the original vibration signal of the target time period according to the change direction, the signal change amplitude and the vibration signal matrix to obtain the main signal components of the target device in the target time period.

[0061] In this embodiment, since the eigenvalue and the eigenvector represent the change amplitude of the signal and the change direction of the signal respectively, the original vibration signal of the target period can be reconstructed according to the change direction, the signal change amplitude and the vibration signal matrix, and then the main signal component of the target device in the target period can be obtained. The main signal component is the reconstructed signal feature, and the main signal component can be used to characterize the signal component with the largest change amplitude in the original vibration signal.

[0062] S104, predicting the energy of the main signal component in the target time period to obtain a target energy parameter;

[0063] In this embodiment, the target energy parameter is the instantaneous energy generated by the main signal component in the target time period. In some embodiments, there are multiple main signal components, so the target energy parameter can be the energy parameter corresponding to the total signal component of the sum of the multiple main signal components. The target energy parameter can be obtained by predicting the energy of the main signal component in the target time period through a pre-set energy prediction model, such as the energy prediction model corresponding to the Teager energy operator.

[0064] S105. Perform data enhancement on the original vibration signal based on the target energy parameter to obtain enhanced signal data.

[0065] In this embodiment, the target energy parameter can reflect the instantaneous energy of the signal. Therefore, by performing data enhancement on the original vibration signal through the target energy parameter, the instantaneous pulse caused by the fault in the original vibration signal can be accurately enhanced, thereby obtaining enhanced signal data with significant pulse characteristics.

[0066] like Figure 3 As shown, Figure 3 is a schematic diagram of a vibration signal after data enhancement provided by an embodiment of the present application. The horizontal axis represents the number of sampling points, and the vertical axis represents the normalized amplitude. Figure 3 It can be seen that, through the non-steady-state data enhancement method in the present application, the non-steady-state signal in the original vibration signal is significantly enhanced, and the fault pulse therein can be accurately identified.

[0067] The non-steady-state data enhancement method proposed in the present application obtains multiple non-steady-state original vibration signals of the target device in the target time period; performs feature decomposition on the vibration signal matrix corresponding to the original vibration signal to obtain eigenvalues ​​and eigenvectors corresponding to the eigenvalues, wherein the eigenvectors are used to characterize the change direction of the original vibration signal in the target time period, and the eigenvalues ​​are used to characterize the signal change amplitude of the original vibration signal in the change direction in the target time period, thereby realizing efficient decomposition of the vibration signal matrix; reconstructs the original vibration signal of the target time period according to the change direction, signal change amplitude and vibration signal matrix, obtains the main signal components of the target device in the target time period, thereby realizing accurate reconstruction of the original vibration signal; then, predicts the energy of the main signal components in the target time period, obtains the target energy parameters, so that the energy intensity of the instantaneous pulse of the signal can be accurately characterized even under high noise interference; finally, performs data enhancement on the original vibration signal based on the target energy parameters to obtain enhanced signal data, thereby realizing accurate data enhancement of the non-steady-state original vibration signal under high noise, further improving the diagnostic efficiency and accuracy of mechanical faults, and having good anti-noise performance, low complexity, and being able to be flexibly deployed in hardware terminals.

[0068] In some embodiments of the present application, before performing eigendecomposition on the vibration signal matrix corresponding to the original vibration signal to obtain eigenvalues ​​and eigenvectors corresponding to the eigenvalues, the method further includes:

[0069] Sorting the original vibration signals according to the collection time corresponding to each of the original vibration signals in the target time period to generate a vibration signal sequence;

[0070] Obtaining the moving step size and window length corresponding to the target time period;

[0071] An interference signal is determined according to the moving step length and the window length, and the vibration signal sequence is dimensionally transformed based on the interference signal and the original vibration signal to generate the vibration signal matrix.

[0072] In this embodiment, the original vibration signals in the target period are vibration signals collected at consecutive collection times, and the original vibration signals are sorted according to the collection times corresponding to the original vibration signals to obtain a vibration signal sequence. The vibration signal sequence can be expressed as , where 1 to N represent the corresponding acquisition time, to Represents the original vibration signal corresponding to each acquisition moment. The vibration signal sequence can be dimensionally adjusted through a preset window to generate a vibration signal matrix. Therefore, the moving step refers to the distance between the first elements of each row in the vibration signal matrix. The number of rows of the generated vibration signal matrix can be determined by the moving step; the window length refers to the number of columns of the generated vibration signal matrix. The moving step and window length can be determined by the number of original vibration signals corresponding to the target time period and the computing power of the computer. Among them, the larger the moving step, the faster the calculation speed.

[0073] For example, the original vibration signal sequence is , then the vibration signal matrix can be expressed as:

[0074]

[0075] Among them, K is the window length and L is the moving step size.

[0076] In some embodiments, if , the window length is 3, the moving step is 1, then the vibration signal matrix can be:

[0077]

[0078] The interference signal is a signal used to fill the vibration signal matrix. According to the moving step size and the window length, the interference signal can be determined. Specifically, when the vibration signal matrix is ​​generated, the original vibration signal in the vibration signal sequence is adjusted by the moving step size and the window length, and there will be repeated original vibration signals, and the repeated original vibration signals are interference signals. According to the interference signal and the original vibration signal, the vibration signal sequence is dimensionally converted to obtain a vibration signal matrix. For example, in the above example, the repeated 2, 3, and 4 are interference signals.

[0079] This embodiment determines the interference signal through the moving step and window length corresponding to the target time period, and performs dimension conversion on the vibration signal sequence based on the interference signal and the original vibration signal to obtain a vibration signal matrix, thereby achieving accurate and rapid generation of the vibration signal matrix, so that the characteristics of the original vibration signal can be quickly acquired through the generated vibration signal matrix.

[0080] In some embodiments of the present application, the above-mentioned performing eigendecomposition on the vibration signal matrix corresponding to the original vibration signal to obtain eigenvalues ​​and eigenvectors corresponding to the eigenvalues ​​includes:

[0081] According to the positions corresponding to the original vibration signal and the interference signal in the vibration signal matrix, the similarity between the signals at any two positions is calculated, and a correlation matrix is ​​generated according to the similarity;

[0082] Perform eigenvalue decomposition on the correlation matrix to obtain the eigenvalues ​​and eigenvectors corresponding to the eigenvalues.

[0083] In the present embodiment, in order to capture the relationship between the features of the original vibration signal, it is necessary to determine the correlation between the original vibration signals, and the correlation can be represented by the similarity between the signals. And in order to realize the eigendecomposition of the vibration signal matrix corresponding to the original vibration signal, it is necessary to generate the correlation matrix corresponding to the original vibration signal, so that the vibration signal matrix can be efficiently characterized by the correlation matrix. Therefore, according to the positions corresponding to the original vibration signal and the interference signal in the vibration signal matrix, the similarity between the signals at any two positions is calculated. Since the interference signal is included in the vibration signal matrix, the similarity between the signals at any two positions includes the similarity between any two original vibration signals and the similarity between the original vibration signal and the interference signal. The similarity can be obtained by variance calculation.

[0084] Generate a correlation matrix based on the similarity, and perform eigendecomposition on the correlation matrix to obtain eigenvalues ​​and eigenvectors corresponding to the eigenvalues. There are multiple eigenvalues ​​and eigenvectors, and one eigenvalue corresponds to one eigenvector. The eigenvalue can be obtained by Indicates that the eigenvector can be obtained by express.

[0085] This embodiment obtains eigenvalues ​​and eigenvectors by performing eigenvalue decomposition on the correlation matrix, thereby achieving efficient acquisition of eigenvalues ​​and eigenvectors, so that the original vibration signal can be accurately reconstructed through the acquired eigenvalues ​​and eigenvectors.

[0086] In some embodiments of the present application, the above-mentioned generating a correlation matrix according to the similarity includes:

[0087] Generate a correlation matrix to be solved according to the vibration signal matrix and a transposed matrix corresponding to the vibration signal matrix;

[0088] The parameters in the correlation matrix to be solved are solved according to the similarity to obtain the correlation matrix.

[0089] In this embodiment, the vibration signal matrix is ​​multiplied by the transposed matrix corresponding to the vibration signal matrix to obtain the correlation matrix to be solved. The correlation matrix to be solved is a symmetric matrix, such as a covariance matrix. The correlation matrix to be solved can be obtained by It indicates that, is the vibration signal matrix, is the transposed matrix of the vibration signal matrix. The elements in the correlation matrix to be solved are composed of the similarities of the signals at any two positions in the vibration signal matrix. Therefore, the parameters in the correlation matrix to be solved can be solved through the similarities of the signals at any two positions in the vibration signal matrix, that is, the correlation matrix can be obtained.

[0090] This embodiment generates a correlation matrix to be solved by using a vibration signal matrix and its corresponding transposed matrix, and then solves the correlation matrix to be solved according to the similarity, that is, obtains the correlation matrix corresponding to the original vibration signal, so that the characteristics of the original vibration signal can be efficiently decomposed through the correlation matrix, thereby improving the decomposition efficiency.

[0091] In some embodiments of the present application, the original vibration signal of the target time period is reconstructed according to the change direction, the signal change amplitude and the vibration signal matrix to obtain the main signal components of the target device in the target time period, including:

[0092] According to the magnitude of the characteristic value, a target characteristic value is selected from the plurality of characteristic values;

[0093] Based on the signal change amplitude and change direction corresponding to the target eigenvalue, the vibration signal matrix is ​​mapped to obtain the main signal components of the target device in the target time period.

[0094] In this embodiment, the eigenvalues ​​may be sorted in descending order according to their magnitudes to obtain a descending sequence corresponding to the eigenvalues, and the first n eigenvalues ​​are selected from the descending sequence as target eigenvalues.

[0095] In some implementations, all eigenvalues ​​may be summed to obtain a total eigenvalue; a preset amplitude ratio is multiplied by the total eigenvalue to obtain a reference eigenvalue. For example, if the preset amplitude ratio is 5%, the total eigenvalue is multiplied by the amplitude ratio of 5% to obtain a reference eigenvalue. All eigenvalues ​​are preliminarily screened according to the reference eigenvalue, and eigenvalues ​​greater than the reference value are screened out as the first eigenvalue. Then, all first eigenvalues ​​are sorted in descending order, and n eigenvalues ​​are selected from large to small as target eigenvalues.

[0096] Since each eigenvalue has its corresponding eigenvector, that is, the direction of change. Therefore, by obtaining the eigenvector corresponding to each target eigenvalue, the change direction corresponding to each target eigenvalue can be obtained. According to the signal change amplitude represented by the target eigenvalue and the change direction corresponding to the target eigenvalue, the vibration signal matrix is ​​mapped to obtain the main signal component of the target device in the target time period. The main signal component is used to characterize the signal component of the original vibration signal in the target time period in the main change direction of the original space, and the original vibration signal can be reconstructed and represented by the main signal component. Specifically, the main signal component can be calculated using the following formula:

[0097]

[0098] in, is the i-th main signal component, is the eigenvector corresponding to the i-th target eigenvalue, is the vibration signal matrix.

[0099] This embodiment selects the target eigenvalue from multiple eigenvalues, maps the vibration signal matrix according to the signal change amplitude and change direction corresponding to the target eigenvalue, and can reconstruct the original vibration signal to obtain the main signal components of the target device in the target time period, thereby achieving accurate expression of the original vibration signal.

[0100] In some embodiments of the present application, predicting the energy of the main signal component in the target time period to obtain the target energy parameter includes:

[0101] Acquire a reference signal component of the target device in an adjacent time period, where the adjacent time period is a time period adjacent to the target time period, and the reference signal component is a signal component corresponding to the target device in the adjacent time period;

[0102] According to a preset energy prediction model, the reference signal component and the main signal component are fused to obtain the target energy parameter.

[0103] In this embodiment, the reference signal component is a signal component generated by the vibration signal of the target device in the adjacent time period, such as a main signal component generated according to the vibration signal of the target device in the adjacent time period. The adjacent time period is a time period adjacent to the target time period, and there may be multiple adjacent time periods. For example, the current target time period is represented as t, then the adjacent time period may include the time period of t+1, the time period of t-1, the time period of t+2, and the time period of t-2. The method for obtaining the reference signal component is the same as the method for obtaining the main signal component of the target device in the target time period in the above embodiment, and will not be repeated here.

[0104] The energy prediction model is a pre-set prediction model, such as an energy prediction model based on the Teager energy operator, or a mathematical model based on integral calculation. Among them, the Teager energy operator is a nonlinear operator that can effectively extract signal energy. According to the preset energy prediction model, the reference signal component and the main signal component are fused to obtain the energy of the main signal component in the target time period, that is, the target energy parameter. Specifically, there may be multiple main signal components obtained through data reconstruction. Therefore, if there are multiple main signal components, all the main signal components are summed to obtain the total signal component of the target device in the target time period; correspondingly, the multiple main signal components of the target device in adjacent time periods are summed to obtain the total signal component of the target device in adjacent time periods. According to the total signal component of the target device in the target time period and the total signal component of the adjacent time periods, the target energy parameter of the target device in the target time period is obtained. Among them, the calculation formula of the total signal component is as follows:

[0105]

[0106] in, is the total signal component, It is the nth main signal component in the target period or adjacent period.

[0107] like Figure 4 As shown, Figure 4 Schematic diagram of the third-order energy operator provided in the embodiment of the present application. The horizontal axis represents the number of sampling points, and the vertical axis represents the normalized amplitude. Figure 4 It can be seen that the irregular hidden impact signal in the original vibration signal can be effectively identified through the third-order energy operator.

[0108] In some implementations, taking the energy prediction model as a prediction model of a third-order energy operator as an example, the calculation formula of the third-order energy operator is as follows:

[0109]

[0110] in, is the target energy parameter, is the total signal component of the target period, is the total signal component of adjacent time periods.

[0111] This embodiment fuses the reference signal component and the main signal component through a preset energy prediction model, which can achieve accurate characterization of the instantaneous pulse energy intensity of the target device in the target time period, thereby improving the significance of the fault pulse in the vibration signal.

[0112] In some embodiments of the present application, the data enhancement of the original vibration signal based on the target energy parameter to obtain enhanced signal data includes:

[0113] Performing noise filtering on the target energy parameter to obtain a filtered target energy parameter;

[0114] The original vibration signal is data enhanced according to the filtered target energy parameter to obtain the enhanced signal data.

[0115] In the present embodiment, the enhanced signal data is data after the energy of the fault pulse in the original vibration signal is enhanced. By performing data enhancement on the original vibration signal, the fault pulse in the original vibration signal can be made more significant, that is, the fault pulse therein can be determined more accurately by enhancing the signal data, thereby achieving accurate judgment of the mechanical fault. Specifically, after the target energy parameter is obtained, there may still be residual noise in the target energy parameter. Therefore, the obtained target energy parameter is subjected to noise filtering to obtain the filtered target energy parameter. In some embodiments, the target energy parameter can be subjected to noise filtering by a filter such as a mean filter or a Gaussian filter to obtain the filtered energy parameter. Then, the original vibration signal is subjected to data enhancement based on the filtered target energy parameter to obtain enhanced signal data.

[0116] In some implementations, data enhancement processing may be performed on the original vibration signal using the following calculation formula to obtain enhanced signal data:

[0117]

[0118] Among them, K is the enhanced signal data, is the target energy parameter after filtering, and X is the original vibration signal.

[0119] This embodiment performs noise filtering on the target energy parameters to obtain filtered target energy parameters; then, data enhancement is performed on the original vibration signal according to the filtered target energy parameters to obtain enhanced signal data, thereby eliminating residual noise and making the enhanced signal data pulses after data enhancement more significant.

[0120] In order to facilitate better implementation of the non-steady-state data enhancement method provided in the embodiment of the present application, the embodiment of the present application also provides a device based on the above non-steady-state data enhancement method. The meanings of the terms are the same as those in the above non-steady-state data enhancement method, and the specific implementation details can refer to the description in the method embodiment.

[0121] For example, Figure 5As shown, the non-stationary data enhancement device may include: an acquisition module 501, a decomposition module 502, a reconstruction module 503, a prediction module 504 and an enhancement module 505.

[0122] An acquisition module 501 is used to acquire a plurality of non-steady-state original vibration signals of a target device in a target period;

[0123] A decomposition module 502 is used to perform eigendecomposition on the vibration signal matrix corresponding to the original vibration signal to obtain an eigenvalue and an eigenvector corresponding to the eigenvalue, wherein the eigenvector is used to characterize the change direction of the original vibration signal in the target time period, and the eigenvalue is used to characterize the signal change amplitude of the original vibration signal in the change direction in the target time period;

[0124] A reconstruction module 503, configured to reconstruct the original vibration signal of the target time period according to the change direction, the signal change amplitude and the vibration signal matrix, to obtain the main signal components of the target device in the target time period;

[0125] A prediction module 504, configured to predict the energy of the main signal component in the target time period to obtain a target energy parameter;

[0126] The enhancement module 505 is used to perform data enhancement on the original vibration signal based on the target energy parameter to obtain enhanced signal data.

[0127] In one embodiment of the present application, the non-steady-state data enhancement device further includes:

[0128] A sorting unit, used to sort the original vibration signals according to the collection time corresponding to each of the original vibration signals in the target time period to generate a vibration signal sequence;

[0129] An acquisition unit, used to acquire a moving step length and a window length corresponding to the target time period;

[0130] The conversion unit is used to determine the interference signal according to the moving step size and the window length, and perform dimension conversion on the vibration signal sequence based on the interference signal and the original vibration signal to generate the vibration signal matrix.

[0131] In one embodiment of the present application, the decomposition module 502 includes:

[0132] A generating unit, configured to calculate the similarity between signals at any two positions according to the positions corresponding to the original vibration signal and the interference signal in the vibration signal matrix, and generate a correlation matrix according to the similarity;

[0133] The decomposition unit is used to perform eigenvalue decomposition on the correlation matrix to obtain the eigenvalue and the eigenvector corresponding to the eigenvalue.

[0134] In one embodiment of the present application, the above-mentioned generation unit is specifically used to: generate a correlation matrix to be solved based on the vibration signal matrix and the transposed matrix corresponding to the vibration signal matrix; solve the parameters in the correlation matrix to be solved according to the similarity to obtain the correlation matrix.

[0135] In an embodiment of the present application, the reconstruction module 503 is specifically used for:

[0136] According to the magnitude of the characteristic value, a target characteristic value is selected from the plurality of characteristic values;

[0137] Based on the signal change amplitude and change direction corresponding to the target eigenvalue, the vibration signal matrix is ​​mapped to obtain the main signal components of the target device in the target time period.

[0138] In an embodiment of the present application, the prediction module 504 is specifically used for:

[0139] Acquire a reference signal component of the target device in an adjacent time period, where the adjacent time period is a time period adjacent to the target time period, and the reference signal component is a signal component corresponding to the target device in the adjacent time period;

[0140] According to a preset energy prediction model, the reference signal component and the main signal component are fused to obtain the target energy parameter.

[0141] In an embodiment of the present application, the enhancement module 505 is specifically used for:

[0142] Performing noise filtering on the target energy parameter to obtain a filtered target energy parameter;

[0143] The original vibration signal is data enhanced according to the filtered target energy parameter to obtain the enhanced signal data.

[0144] The non-steady-state data enhancement device proposed in this application realizes accurate data enhancement of non-steady-state original vibration signals under high noise, further improving the diagnostic efficiency and accuracy of mechanical faults. It also has good anti-noise performance, low complexity, and can be flexibly deployed in hardware terminals.

[0145] In specific implementation, the above modules can be implemented as independent entities, or can be arbitrarily combined and implemented as the same or several entities. The specific implementation methods and corresponding beneficial effects of the above modules can be found in the previous method embodiments, which will not be repeated here.

[0146] The present application also provides an electronic device, which may be a server or a terminal, etc. Figure 6 As shown, it shows a schematic diagram of the structure of the electronic device involved in the embodiment of the present application, specifically:

[0147] The electronic device may include components such as a processor 601 with one or more processing cores, a memory 602 with one or more computer-readable storage media, a power supply 603, and an input unit 604. Those skilled in the art will appreciate that Figure 6 The electronic device structure shown in the figure does not constitute a limitation on the electronic device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0148] The processor 601 is the control center of the electronic device, which uses various interfaces and lines to connect various parts of the entire electronic device, and executes various functions of the electronic device and processes data by running or executing computer programs and / or modules stored in the memory 602, and calling data stored in the memory 602. Optionally, the processor 601 may include one or more processing cores; preferably, the processor 601 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 601.

[0149] The memory 602 can be used to store computer programs and modules. The processor 601 executes various functional applications and data processing by running the computer programs and modules stored in the memory 602. The memory 602 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, a computer program required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 602 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 602 may also include a memory controller to provide the processor 601 with access to the memory 602.

[0150] The electronic device also includes a power supply 603 for supplying power to each component. Preferably, the power supply 603 can be logically connected to the processor 601 through a power management system, so as to manage charging, discharging, power consumption and other functions through the power management system. The power supply 603 can also include one or more DC or AC power supplies, recharging systems, power failure detection circuits, power converters or inverters, power status indicators and other arbitrary components.

[0151] The electronic device may further include an input unit 604, which may be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0152] Although not shown, the electronic device may further include a display unit, etc., which will not be described in detail herein. Specifically in this embodiment, the processor 601 in the electronic device will load the executable files corresponding to the processes of one or more computer programs into the memory 602 according to the following instructions, and the processor 601 will run the computer programs stored in the memory 602, thereby realizing various functions, such as:

[0153] Acquire multiple non-steady-state original vibration signals of the target device in a target period;

[0154] Performing eigendecomposition on a vibration signal matrix corresponding to the original vibration signal to obtain an eigenvalue and an eigenvector corresponding to the eigenvalue, wherein the eigenvector is used to characterize a change direction of the original vibration signal in the target time period, and the eigenvalue is used to characterize a signal change amplitude of the original vibration signal in the change direction in the target time period;

[0155] Reconstructing the original vibration signal of the target time period according to the change direction, the signal change amplitude and the vibration signal matrix to obtain the main signal components of the target device in the target time period;

[0156] Predicting the energy of the main signal component in the target time period to obtain a target energy parameter;

[0157] The original vibration signal is data enhanced based on the target energy parameter to obtain enhanced signal data.

[0158] The specific implementation methods and corresponding beneficial effects of the above operations can be found in the detailed description of the non-stationary data enhancement method above, which will not be elaborated here.

[0159] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by a computer program, or by controlling related hardware through a computer program. The computer program may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0160] To this end, an embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored, and the computer program can be loaded by a processor to execute the steps in any non-stationary data enhancement method provided in the embodiment of the present application. For example, the computer program can execute the following steps:

[0161] Acquire multiple non-steady-state original vibration signals of the target device in a target period;

[0162] Performing eigendecomposition on a vibration signal matrix corresponding to the original vibration signal to obtain an eigenvalue and an eigenvector corresponding to the eigenvalue, wherein the eigenvector is used to characterize a change direction of the original vibration signal in the target time period, and the eigenvalue is used to characterize a signal change amplitude of the original vibration signal in the change direction in the target time period;

[0163] Reconstructing the original vibration signal of the target time period according to the change direction, the signal change amplitude and the vibration signal matrix to obtain the main signal components of the target device in the target time period;

[0164] Predicting the energy of the main signal component in the target time period to obtain a target energy parameter;

[0165] The original vibration signal is data enhanced based on the target energy parameter to obtain enhanced signal data.

[0166] The specific implementation methods and corresponding beneficial effects of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0167] The computer-readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0168] Since the computer program stored in the computer-readable storage medium can execute the steps in any one of the non-steady-state data enhancement methods provided in the embodiments of the present application, the beneficial effects that can be achieved by any one of the non-steady-state data enhancement methods provided in the embodiments of the present application can be achieved. Please refer to the previous embodiments for details and will not be repeated here.

[0169] According to one aspect of the present application, a computer program product or a computer program is provided, the computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the above-mentioned non-stationary data enhancement method.

[0170] The above is a detailed introduction to a non-steady-state data enhancement method, device, electronic device and computer-readable storage medium provided in the embodiments of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, according to the ideas of the present application, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A non-stationary data enhancement method, characterized in that: include: Acquire multiple non-steady-state original vibration signals of the target device in a target period; Performing eigendecomposition on a vibration signal matrix corresponding to the original vibration signal to obtain an eigenvalue and an eigenvector corresponding to the eigenvalue, wherein the eigenvector is used to characterize a change direction of the original vibration signal in the target time period, and the eigenvalue is used to characterize a signal change amplitude of the original vibration signal in the change direction in the target time period; Reconstructing the original vibration signal of the target time period according to the change direction, the signal change amplitude and the vibration signal matrix to obtain the main signal component of the target device in the target time period, wherein the main signal component is used to characterize the signal component of the original vibration signal in the main change direction of the original space; Predicting the energy of the main signal component in the target time period to obtain a target energy parameter, wherein the target energy parameter is an energy parameter corresponding to a total signal component of the sum of multiple main signal components; The original vibration signal is data enhanced based on the target energy parameter to obtain enhanced signal data.

2. The non-stationary data enhancement method according to claim 1, characterized in that: Before performing eigendecomposition on the vibration signal matrix corresponding to the original vibration signal to obtain eigenvalues ​​and eigenvectors corresponding to the eigenvalues, the method further includes: Sorting the original vibration signals according to the collection time corresponding to each of the original vibration signals in the target time period to generate a vibration signal sequence; Obtaining the moving step size and window length corresponding to the target time period; An interference signal is determined according to the moving step length and the window length, and the vibration signal sequence is dimensionally transformed based on the interference signal and the original vibration signal to generate the vibration signal matrix.

3. The non-stationary data enhancement method according to claim 2, characterized in that: The step of performing eigendecomposition on the vibration signal matrix corresponding to the original vibration signal to obtain eigenvalues ​​and eigenvectors corresponding to the eigenvalues ​​includes: According to the positions corresponding to the original vibration signal and the interference signal in the vibration signal matrix, the similarity between the signals at any two positions is calculated, and a correlation matrix is ​​generated according to the similarity; Perform eigenvalue decomposition on the correlation matrix to obtain the eigenvalues ​​and eigenvectors corresponding to the eigenvalues.

4. The non-stationary data enhancement method according to claim 3, characterized in that: The generating a correlation matrix according to the similarity comprises: Generate a correlation matrix to be solved according to the vibration signal matrix and a transposed matrix corresponding to the vibration signal matrix; The parameters in the correlation matrix to be solved are solved according to the similarity to obtain the correlation matrix.

5. The non-stationary data enhancement method according to claim 1, characterized in that: The reconstructing the original vibration signal of the target time period according to the change direction, the signal change amplitude and the vibration signal matrix to obtain the main signal components of the target device in the target time period includes: According to the magnitude of the characteristic value, a target characteristic value is selected from the plurality of characteristic values; Based on the signal change amplitude and change direction corresponding to the target eigenvalue, the vibration signal matrix is ​​mapped to obtain the main signal components of the target device in the target time period.

6. The non-stationary data enhancement method according to claim 1, characterized in that: The predicting the energy of the main signal component in the target time period to obtain a target energy parameter comprises: Acquire a reference signal component of the target device in an adjacent time period, where the adjacent time period is a time period adjacent to the target time period, and the reference signal component is a signal component corresponding to the target device in the adjacent time period; According to a preset energy prediction model, the reference signal component and the main signal component are fused to obtain the target energy parameter.

7. The non-stationary data enhancement method according to claim 1, characterized in that: The step of performing data enhancement on the original vibration signal based on the target energy parameter to obtain enhanced signal data includes: Performing noise filtering on the target energy parameter to obtain a filtered target energy parameter; The original vibration signal is data enhanced according to the filtered target energy parameter to obtain the enhanced signal data.

8. A non-steady-state data enhancement device, characterized in that: include: An acquisition module, used to acquire a plurality of non-steady-state original vibration signals of a target device in a target period; a decomposition module, configured to perform eigendecomposition on a vibration signal matrix corresponding to the original vibration signal to obtain an eigenvalue and an eigenvector corresponding to the eigenvalue, wherein the eigenvector is used to characterize a change direction of the original vibration signal in the target time period, and the eigenvalue is used to characterize a signal change amplitude of the original vibration signal in the change direction in the target time period; A reconstruction module, used to reconstruct the original vibration signal of the target time period according to the change direction, the signal change amplitude and the vibration signal matrix, to obtain the main signal component of the target device in the target time period, wherein the main signal component is used to characterize the signal component of the original vibration signal in the main change direction of the original space; A prediction module, used to predict the energy of the main signal component in the target time period to obtain a target energy parameter, where the target energy parameter is an energy parameter corresponding to a total signal component of the sum of multiple main signal components; The enhancement module is used to perform data enhancement on the original vibration signal based on the target energy parameter to obtain enhanced signal data.

9. An electronic device, characterized in that: It comprises a processor and a memory, wherein the memory stores a computer program, and the processor is used to run the computer program in the memory to execute the non-stationary data enhancement method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the non-stationary data enhancement method according to any one of claims 1 to 7.