Hoisting machinery fault diagnosis method based on multivariate adaptive singular spectrum analysis

Abstract

The invention relates to a hoisting machinery fault diagnosis method based on multivariate adaptive singular spectrum analysis. The method comprises the following steps: firstly, collecting multi-channel vibration data; then, performing multivariate adaptive singular spectrum decomposition on the multi-channel vibration data to obtain a multi-channel modal component and a residual signal; then, calculating a complexity index of the multi-channel modal component, and determining a feature matrix of multi-channel fault information; and finally, based on the feature matrix, a convolutional neural network is adopted to carry out fault mode identification, and the fault type of the hoisting machinery is determined. The method comprises the following steps: performing multivariate adaptive singular spectrum decomposition preprocessing on a multi-channel signal, extracting a modal component with a distinguishing feature, calculating a complexity index of the multi-channel modal component, and training the complexity index in combination with a convolutional neural network, so that representative fault features can be extracted; and a characteristic matrix containing multi-channel fault information is automatically obtained, automatic fault mode identification is realized, and the diagnosis efficiency and accuracy are greatly improved.

Description

Technical Field

[0001] The present application relates to the fields of artificial intelligence and mechanical engineering technology, and in particular to a method for diagnosing lifting machinery faults based on multivariate adaptive singular spectrum analysis. Background Art

[0002] With the continuous advancement of industrialization, lifting machinery plays a vital role in various types of engineering construction. Lifting machinery systems are widely used in equipment lifting, material handling, transportation and other fields, especially in the construction, mining, port and logistics industries. However, during long-term and high-load operation, the lifting machinery system is prone to various faults due to the complexity and harshness of its working environment, which seriously affects the normal operation and work efficiency of the equipment, and may even cause equipment damage or casualties, bringing huge safety hazards and economic losses to production. Therefore, in order to improve the safety and stability of equipment operation, it is of great necessity and significance to carry out fault diagnosis research on lifting machinery systems.

[0003] At present, many methods have been successfully applied to the fault diagnosis of hoisting systems, including wavelet transform, maximum correlation coefficient deconvolution, spectral coefficient, morphological filtering and singular spectrum analysis. However, each of these methods has certain limitations. For example, since wavelet transform is limited by the Heisenberg uncertainty principle, the resolution and aggregation of the time-frequency trajectory obtained by it are insufficient. In addition, when wavelet transform is used to analyze actual non-stationary vibration signals, it lacks adaptive ability. Maximum correlation coefficient deconvolution can reveal continuous pulses related to bearing or gear faults, but it requires manual selection of appropriate deconvolution period and filter length. Spectral coefficient can describe a series of transient features in the original vibration signal, but its two key parameters (i.e., passband center frequency and resonance bandwidth) have a great influence on its detection performance. Morphological filtering matches the fault information in the vibration signal through a structural element detector, but the type and size of the structural element are difficult to select effectively. The embedding dimension and time delay of the phase space reconstruction in singular spectrum analysis cannot be automatically determined.

[0004] In addition, some adaptive signal decomposition methods (such as empirical mode decomposition, local mean decomposition, intrinsic time scale decomposition, empirical wavelet transform and variational mode decomposition) have also been developed for fault diagnosis of mechanical equipment. However, the above signal decomposition methods mainly focus on single-channel data analysis, ignoring the fault information contained in other channels in the crane multi-sensor system. When the key parts of the lifting equipment are damaged, the inherent fault characteristics are usually propagated in the multi-channel signals obtained by multi-sensors. Therefore, in order to obtain more comprehensive fault information, some multi-channel data processing methods (such as multivariate empirical mode decomposition (MEMD), multivariate intrinsic time scale decomposition (MITD) and multivariate variational mode decomposition (MVMD)) have been gradually proposed to mine fault information in the multi-sensor data of mechanical equipment. However, the existing multi-channel signal decomposition methods still have the same shortcomings as the traditional single-channel signal decomposition methods. For example, multivariate empirical mode decomposition suffers from mode aliasing and endpoint effect problems; due to the use of linear transformation, multivariate intrinsic time scale decomposition will experience local data fluctuations and curve distortion when processing mechanical vibration data; the selection of two important parameters of multivariate variational mode decomposition (i.e., penalty factor and modulus) is highly dependent on professional knowledge and experience, and the selection of these parameters directly affects its decomposition performance.

[0005] Therefore, in the related technology, there is an urgent need for a method that can improve the accuracy and adaptability of various fault diagnosis of lifting machinery. Summary of the invention

[0006] Based on this, it is necessary to provide a crane fault diagnosis method based on multivariate adaptive singular spectrum analysis that can improve the accuracy and adaptability of various crane fault diagnosis in response to the above technical problems.

[0007] In a first aspect, the present application provides a method for diagnosing hoisting machinery faults based on multivariate adaptive singular spectrum analysis. The method comprises:

[0008] Collect multi-channel vibration data;

[0009] Performing multivariate adaptive singular spectrum decomposition on the multi-channel vibration data to obtain multi-channel modal components and residual signals;

[0010] Calculating the complexity index of the multi-channel modal components and determining the characteristic matrix of the multi-channel fault information;

[0011] Based on the characteristic matrix, a convolutional neural network is used to perform fault pattern recognition to determine the type of lifting machinery fault.

[0012] Optionally, in one embodiment of the present application, after collecting the multi-channel vibration data, the step further includes:

[0013] Initialize the decomposition parameters of multivariate adaptive singular spectrum analysis, including sampling frequency, decomposition stop threshold, and the maximum number of layers that the signal can be decomposed.

[0014] Optionally, in one embodiment of the present application, the objective function of performing multivariate adaptive singular spectrum decomposition on the multi-channel vibration data is:

[0015]

[0016] Among them, J is the objective function, x i (t) is the channel signal, y ik (t) is the modal component; ‖·‖ 2 represents the bi-norm, which is used to measure the error; λ is the regularization parameter, which is used to control the complexity of signal decomposition; ‖y ik (t)‖ p is the modal component y ik The p-norm of (t) is used to constrain the complexity of the modal components.

[0017] Optionally, in one embodiment of the present application, the calculation formula of the complexity index is:

[0018]

[0019] Among them, D(y ik (t)) is the improved Kolmogorov complexity, P i is the probability that the observation system is located at the i-th unit at the τth time, ε is the number of symbols, N represents the length of the symbol time series, y ik (t) is the modal component.

[0020] Optionally, in one embodiment of the present application, the using a convolutional neural network to perform fault mode recognition based on the feature matrix includes:

[0021] A cross entropy loss function is used to optimize the training process of the convolutional neural network.

[0022] Optionally, in one embodiment of the present application, the method further includes:

[0023] The precision, recall and F1 measurement indicators are used to verify the fault identification performance.

[0024] Optionally, in one embodiment of the present application, the method further includes:

[0025] The intra-class distance, inter-class distance and the ratio of inter-class to intra-class distance are used to evaluate the effectiveness of the multi-channel fault information extraction.

[0026] In a second aspect, the present application also provides a hoisting machinery fault diagnosis device based on multivariate adaptive singular spectrum analysis. The device comprises:

[0027] Data acquisition module, used to collect multi-channel vibration data;

[0028] A multivariate adaptive singular spectrum analysis module, used for performing multivariate adaptive singular spectrum decomposition on the multi-channel vibration data to obtain multi-channel modal components and residual signals;

[0029] A multi-channel fault information determination module, used to calculate the complexity index of the multi-channel modal component and determine the characteristic matrix of the multi-channel fault information;

[0030] The hoisting machinery fault diagnosis module is used to perform fault pattern recognition based on the feature matrix using a convolutional neural network to determine the type of hoisting machinery fault.

[0031] In a third aspect, the present application further provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor executes the steps of the methods described in the above embodiments.

[0032] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method described in each of the above embodiments are implemented.

[0033] The above-mentioned method for diagnosing hoisting machinery faults based on multivariate adaptive singular spectrum analysis first collects multi-channel vibration data; then, performs multivariate adaptive singular spectrum decomposition on the multi-channel vibration data to obtain multi-channel modal components and residual signals; then, calculates the complexity index of the multi-channel modal components to determine the characteristic matrix of the multi-channel fault information; finally, based on the characteristic matrix, a convolutional neural network is used to perform fault mode recognition to determine the type of hoisting machinery fault. In other words, by pre-processing the multi-channel signal with multivariate adaptive singular spectrum decomposition, extracting modal components with distinguishing characteristics, calculating the complexity index of the multi-channel modal components, and training the complexity index in combination with a convolutional neural network (CNN), the noise component in the signal can be effectively separated, representative fault features can be extracted, and a characteristic matrix containing multi-channel fault information can be automatically obtained, realizing the automatic fault mode recognition of the hoisting machinery system, solving the problem of difficult feature selection, and significantly improving the robustness of the system, greatly improving the diagnostic efficiency and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is an application environment diagram of a crane fault diagnosis method based on multivariate adaptive singular spectrum analysis in one embodiment;

[0035] Figure 2 A schematic diagram of a flow chart of a method for diagnosing a hoisting machinery fault based on multivariate adaptive singular spectrum analysis in one embodiment;

[0036] Figure 3 A schematic diagram of a time domain waveform, a Fourier spectrum, and an envelope spectrum of a multi-channel fault signal collected by a core component in one embodiment;

[0037] Figure 4 A schematic diagram of the structure of a multivariate adaptive singular spectrum decomposition result in one embodiment;

[0038] Figure 5 It is a structural block diagram of a crane fault diagnosis device based on multivariate adaptive singular spectrum analysis in one embodiment;

[0039] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0041] The embodiment of the present application provides a method for diagnosing hoisting machinery faults based on multivariate adaptive singular spectrum analysis, which can be applied to Figure 1 In the application environment shown. Among them, the terminal communicates with the server through the network. The data storage system can store the data that the server needs to process. The data storage system can be integrated on the server, or it can be placed on the cloud or other network servers. Among them, the terminal can be but not limited to various personal computers, laptops, smart phones, tablets, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server can be implemented as an independent server or a server cluster consisting of multiple servers.

[0042] In one embodiment, Figure 2 As shown in the figure, a method for hoisting machinery fault diagnosis based on multivariate adaptive singular spectrum analysis is provided. Figure 1 The server in the example is used to illustrate the following steps:

[0043] S201: Collect multi-channel vibration data.

[0044] In the embodiment of the present application, first, a multi-sensor device is installed on the lifting mechanical system to collect multi-channel vibration data in different health states. The collected multi-channel vibration data is a multi-channel signal X(t)={x 1 (t),x 2 (t),…,x m (t)}, where x m (t)=[p(1),p(2),…,p(N)] T ,m=1,2,… represents the mth variable signal, p(N) is the signal x m (t) is the Nth data point. Figure 3 Shown are the time domain waveforms, Fourier spectra and envelope spectra of multi-channel fault signals collected from four core components (i.e., equalizing beam, A-frame, flying arm and lifting arm).

[0045] Afterwards, in one embodiment of the present application, the collecting of multi-channel vibration data further includes:

[0046] Initialize the decomposition parameters of multivariate adaptive singular spectrum analysis, including sampling frequency, decomposition stop threshold, and the maximum number of layers that the signal can be decomposed.

[0047] S203: performing multivariate adaptive singular spectrum decomposition on the multi-channel vibration data to obtain multi-channel modal components and residual signals.

[0048] In the embodiment of the present application, multivariate adaptive singular spectrum decomposition is performed on the collected multi-channel vibration data, i.e., the multi-channel signal X(t), to obtain the multi-channel modal components containing the discriminant information. Specifically, for each channel signal x i (t), the goal of its multivariate adaptive singular spectrum decomposition is to decompose it into multiple modal components y ik (t), and remove noise and unnecessary information as much as possible. The decomposition process is optimized by minimizing the relevant objective function.

[0049] Specifically, in one embodiment of the present application, the objective function of performing multivariate adaptive singular spectrum decomposition on the multi-channel vibration data is:

[0050]

[0051] Among them, J is the objective function, x i (t) is the channel signal, y ik (t) is the modal component; ‖·‖ 2 represents the bi-norm, which is used to measure the error; λ is the regularization parameter, which is used to control the complexity of signal decomposition; ‖y ik (t)‖ p is the modal component y ikThe p-norm of (t) is used to constrain the complexity of the modal components.

[0052] Whether the decomposition result meets expectations is usually determined by whether the stop decomposition criterion is met, that is, when the residual signal V (k) When the normalized mean square error between (t) and the multi-channel signal X(t) is less than the preset decomposition stop threshold θ, it proves that the decomposition result meets the expectations and the signal decomposition process stops. At this point, K multi-channel modal components can be obtained and K residual signals If the stop decomposition criterion is not met, the extracted component is subtracted from the current residual signal to obtain a new residual signal, and then the new residual signal is used as the input signal to continue decomposing until the stop decomposition criterion is met. Finally, the multivariate adaptive singular spectrum analysis result of the multichannel signal is expressed as a set of K multichannel modal components and the corresponding residual signal, which is expressed by the following formula.

[0053]

[0054] Among them, y ik (t) can be intuitively understood as a space vector with three directions (i.e., X-axis, Y-axis, and Z-axis). The X-axis direction represents the number of multi-channel modal components, the Y-axis direction represents the length of the multi-channel modal components, and the Z-axis direction represents the number of channels of the input multi-channel signal. Figure 4 Shown are examples of decomposition results, (a) equalizing beam, (b) A-frame, (c) flying arm, and (d) lifting arm.

[0055] S205: Calculate the complexity index of the multi-channel modal components and determine the characteristic matrix of the multi-channel fault information.

[0056] In the embodiment of the present application, after obtaining the multi-channel modal components, the modal components y are calculated. ik (t) to capture the dynamic characteristics of the signal, and based on this, determine the feature matrix containing multi-channel fault information. Optionally, by calculating the improved Kolmogorov complexity D(y ik (t)) is used to measure the rate of change of the modal component.

[0057] Specifically, in one embodiment of the present application, the calculation formula of the complexity index is:

[0058]

[0059] Among them, D(y ik (t)) is the improved Kolmogorov complexity, P i is the probability that the observation system is located at the i-th unit at the τth time, ε is the number of symbols, N represents the length of the symbol time series, yik (t) is the modal component.

[0060] By calculating D(y ik (t)) value, the feature matrix F containing multi-channel fault information can be obtained as shown below.

[0061]

[0062] Where l is the number of channels, R is the number of modal components, and D ik is the complexity index of the kth modal component of the ith channel. By calculating the characteristic matrix F, rich multi-channel fault information can be provided for subsequent fault mode identification.

[0063] S207: Using a convolutional neural network to perform fault pattern recognition based on the feature matrix to determine the type of hoisting machinery fault.

[0064] In an embodiment of the present application, the extracted features containing multi-channel fault information are randomly divided into training samples and test samples in a ratio of 1:1, wherein the training samples are used to train a convolutional neural network (CNN) model, and the test samples are input into the trained CNN model to realize automatic recognition of different fault modes of the lifting machinery system and determine the type of lifting machinery fault.

[0065] Specifically, in one embodiment of the present application, the fault mode recognition using a convolutional neural network based on the feature matrix includes:

[0066] A cross entropy loss function is used to optimize the training process of the convolutional neural network.

[0067] In one embodiment of the present application, the training process of the convolutional neural network (CNN) model is optimized by minimizing the loss function, and the loss function adopts the cross entropy loss function, which is expressed as follows:

[0068]

[0069] Among them, y i is the actual label (fault type), is the predicted value output by the CNN model. The fault mode in the feature matrix is ​​learned through the back propagation algorithm of stochastic gradient descent to complete the identification of the fault mode.

[0070] In one embodiment of the present application, the method further includes:

[0071] The precision, recall and F1 measurement indicators are used to verify the fault identification performance.

[0072] In one embodiment of the present application, by calculating the predicted value With the actual label yi The precision p, recall rate Re and F1 measurement values ​​between are used to verify the fault recognition performance. The specific calculation formula is as follows:

[0073]

[0074] Among them, T p represents the number of true positives, B p represents the number of false positives, B N It represents the number of false negatives, β is a weighting parameter, and the higher the F1 measurement value, the better the recognition result, and its value usually ranges between 0 and 1.

[0075] In one embodiment of the present application, the method further includes:

[0076] The intra-class distance, inter-class distance and the ratio of inter-class to intra-class distance are used to evaluate the effectiveness of the multi-channel fault information extraction.

[0077] In one embodiment of the present application, by calculating the intra-class distance W d , class spacing C d , the ratio of the distance between classes and within classes C d / W d , in order to quantitatively evaluate the effectiveness of multi-channel fault information extraction. The specific calculation method is as follows:

[0078]

[0079] Among them, u() represents the Euclidean distance between two vectors, and denote the kth and lth samples of the ith category respectively, represents the lth sample of the jth category, N i and N j They represent the number of samples in the i-th category and the j-th category respectively.

[0080] In the above-mentioned method for diagnosing hoisting machinery faults based on multivariate adaptive singular spectrum analysis, first, multi-channel vibration data is collected; then, the multi-channel vibration data is subjected to multi-variate adaptive singular spectrum decomposition to obtain multi-channel modal components and residual signals; then, the complexity index of the multi-channel modal components is calculated to determine the characteristic matrix of the multi-channel fault information; finally, the fault mode recognition is performed based on the characteristic matrix using a convolutional neural network to determine the type of hoisting machinery fault. In other words, by pre-processing the multi-channel signal with multi-variate adaptive singular spectrum decomposition, extracting the modal components with distinguishing characteristics, calculating the complexity index of the multi-channel modal components, and training the complexity index in combination with a convolutional neural network (CNN), the noise component in the signal can be effectively separated, representative fault features can be extracted, and a characteristic matrix containing multi-channel fault information can be automatically obtained, thereby realizing the automatic fault mode recognition of the hoisting machinery system, solving the problem of difficult feature selection, and significantly improving the robustness of the system, and greatly improving the diagnostic efficiency and accuracy.

[0081] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0082] Based on the same inventive concept, the embodiment of the present application also provides a device for diagnosing a crane fault based on multivariate adaptive singular spectrum analysis for implementing the above-mentioned method for diagnosing a crane fault based on multivariate adaptive singular spectrum analysis. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in one or more embodiments of the device for diagnosing a crane fault based on multivariate adaptive singular spectrum analysis provided below can be referred to the limitations of the method for diagnosing a crane fault based on multivariate adaptive singular spectrum analysis above, and will not be repeated here.

[0083] In one embodiment, Figure 5As shown, a hoisting machinery fault diagnosis device 500 based on multivariate adaptive singular spectrum analysis is provided, comprising: a data acquisition module 501, a multivariate adaptive singular spectrum analysis module 503, a multi-channel fault information determination module 505 and a hoisting machinery fault diagnosis module 507, wherein:

[0084] The data acquisition module 501 is used to acquire multi-channel vibration data.

[0085] The multivariate adaptive singular spectrum analysis module 503 is used to perform multivariate adaptive singular spectrum decomposition on the multi-channel vibration data to obtain multi-channel modal components and residual signals.

[0086] The multi-channel fault information determination module 505 is used to calculate the complexity index of the multi-channel modal components and determine the characteristic matrix of the multi-channel fault information.

[0087] The hoisting machinery fault diagnosis module 507 is used to perform fault pattern recognition based on the feature matrix using a convolutional neural network to determine the type of hoisting machinery fault.

[0088] In one embodiment of the present application, the multivariate adaptive singular spectrum analysis module is further used for:

[0089] Initialize the decomposition parameters of multivariate adaptive singular spectrum analysis, including sampling frequency, decomposition stop threshold, and the maximum number of layers that the signal can be decomposed.

[0090] In one embodiment of the present application, the objective function of performing multivariate adaptive singular spectrum decomposition on the multi-channel vibration data is:

[0091]

[0092] Among them, J is the objective function, x i (t) is the channel signal, y ik (t) is the modal component; ‖·‖ 2 represents the bi-norm, which is used to measure the error; λ is the regularization parameter, which is used to control the complexity of signal decomposition; ‖y ik (t)‖ p is the modal component y ik The p-norm of (t) is used to constrain the complexity of the modal components.

[0093] In one embodiment of the present application, the calculation formula of the complexity index is:

[0094]

[0095] Among them, D(y ik (t)) is the improved Kolmogorov complexity, P iis the probability that the observation system is located at the i-th unit at the τth time, ε is the number of symbols, N represents the length of the symbol time series, y ik (t) is the modal component.

[0096] In one embodiment of the present application, the lifting machinery fault diagnosis module is also used for:

[0097] A cross entropy loss function is used to optimize the training process of the convolutional neural network.

[0098] In one embodiment of the present application, the method further includes:

[0099] The precision, recall and F1 measurement indicators are used to verify the fault identification performance.

[0100] In one embodiment of the present application, the method further includes:

[0101] The intra-class distance, inter-class distance and the ratio of inter-class to intra-class distance are used to evaluate the effectiveness of the multi-channel fault information extraction.

[0102] Each module in the above-mentioned hoisting machinery fault diagnosis device based on multivariate adaptive singular spectrum analysis can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each of the above modules.

[0103] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 6 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for diagnosing hoisting machinery faults based on multivariate adaptive singular spectrum analysis is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse, etc.

[0104] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0105] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0106] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0107] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0108] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0109] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0110] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0111] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for diagnosing hoisting machinery faults based on multivariate adaptive singular spectrum analysis, characterized in that: The method comprises: Collect multi-channel vibration data; Performing multivariate adaptive singular spectrum decomposition on the multi-channel vibration data to obtain multi-channel modal components and residual signals; Calculating the complexity index of the multi-channel modal components and determining the characteristic matrix of the multi-channel fault information; Based on the characteristic matrix, a convolutional neural network is used to perform fault pattern recognition to determine the type of lifting machinery fault.

2. The method for diagnosing hoisting machinery faults based on multivariate adaptive singular spectrum analysis according to claim 1 is characterized in that: After collecting the multi-channel vibration data, the method further includes: Initialize the decomposition parameters of multivariate adaptive singular spectrum analysis, including sampling frequency, decomposition stop threshold, and the maximum number of layers that the signal can be decomposed.

3. The method for diagnosing hoisting machinery faults based on multivariate adaptive singular spectrum analysis according to claim 1 is characterized in that: The objective function of performing multivariate adaptive singular spectrum decomposition on the multi-channel vibration data is: Among them, J is the objective function, x i (t) is the channel signal, y ik (t) is the modal component; ‖·‖2 represents the bi-norm, which is used to measure the error; λ is the regularization parameter, which is used to control the complexity of signal decomposition; ‖y ik (t)‖ p is the modal component y ik The p-norm of (t) is used to constrain the complexity of the modal components.

4. The method for diagnosing hoisting machinery faults based on multivariate adaptive singular spectrum analysis according to claim 1 is characterized in that: The calculation formula of the complexity index is: Among them, D(y ik (t)) is the improved Kolmogorov complexity, P i is the probability that the observation system is located at the i-th unit at the τth time, ε is the number of symbols, N represents the length of the symbol time series, y ik (t) is the modal component.

5. The method for diagnosing hoisting machinery faults based on multivariate adaptive singular spectrum analysis according to claim 1 is characterized in that: The method of using a convolutional neural network to identify fault patterns based on the feature matrix includes: A cross entropy loss function is used to optimize the training process of the convolutional neural network.

6. The method for diagnosing hoisting machinery faults based on multivariate adaptive singular spectrum analysis according to claim 1, characterized in that: The method further comprises: The precision, recall and F1 measurement indicators are used to verify the fault identification performance.

7. The method for diagnosing hoisting machinery faults based on multivariate adaptive singular spectrum analysis according to claim 1 is characterized in that: The method further comprises: The intra-class distance, inter-class distance and the ratio of inter-class to intra-class distance are used to evaluate the effectiveness of the multi-channel fault information extraction.

8. A hoisting machinery fault diagnosis device based on multivariate adaptive singular spectrum analysis, characterized in that: The device comprises: Data acquisition module, used to collect multi-channel vibration data; A multivariate adaptive singular spectrum analysis module, used for performing multivariate adaptive singular spectrum decomposition on the multi-channel vibration data to obtain multi-channel modal components and residual signals; A multi-channel fault information determination module, used to calculate the complexity index of the multi-channel modal component and determine the characteristic matrix of the multi-channel fault information; The hoisting machinery fault diagnosis module is used to perform fault pattern recognition based on the feature matrix using a convolutional neural network to determine the type of hoisting machinery fault.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images