Fault monitoring method and device for long short-term memory neural network, and computer equipment
By applying a fault monitoring method that combines long and short-term memory neural networks in chemical enterprises with empirical wavelet transformation and principal component analysis, the problem that traditional monitoring methods cannot effectively deal with complex patterns and weak fault characteristics in time series data is solved, achieving higher monitoring accuracy and faster fault recognition speed.
Patent Information
- Application Number
- CN202411887075.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-20
- Publication Date
- 2025-03-21
AI Technical Summary
Traditional monitoring methods cannot effectively handle complex patterns and weak fault characteristics in time series data, resulting in low monitoring accuracy.
Long short-term memory neural network (LSTM) was used to combine empirical wavelet transform (EWT) and principal component analysis (PCA). The specific steps include: obtaining the timing data of the device, performing empirical wavelet transformation processing to extract the feature matrix, reducing the dimensionality through principal component analysis to obtain fault characteristics, and inputting it into the fault monitoring model to obtain fault identification results.
By using empirical wavelet transform to extract useful signals and using principal component analysis to perform feature dimensionality reduction, it can improve fault identification speed while maintaining monitoring accuracy and reduce calculation burden.
Smart Images
Figure CN119669834A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of industrial simulation technology, and in particular to a long short-term memory neural network fault monitoring method, device and computer equipment. Background Art
[0002] With the rapid development of modern chemical industry products in various fields, automation and intelligent technologies have been widely used, and the production equipment and labor costs of chemical companies have continued to grow. Therefore, chemical companies have an increasing demand for automation equipment, and they continue to strengthen the level of automation and intelligence to ensure the smooth and safe operation of production equipment. However, this also makes the production process more complicated. In this case, even if a link in the factory fails, it may trigger a series of chain reactions, causing irreparable damage to the production equipment, and even causing serious economic losses and safety hazards. Therefore, in intelligent large-scale process simulation systems, early diagnosis of faults is crucial to ensure the stability and reliability of the system. Traditional monitoring methods cannot effectively handle complex patterns and weak fault characteristics in time series data, resulting in low monitoring accuracy. Summary of the invention
[0003] Based on this, it is necessary to provide a long short-term memory neural network fault monitoring method, device, computer equipment, computer readable storage medium and computer program product that can improve the monitoring accuracy in response to the above technical problems.
[0004] In a first aspect, the present application provides a long short-term memory neural network fault monitoring method. The method comprises:
[0005] Obtain the time series data of the device and generate a data matrix;
[0006] Performing empirical wavelet transform on the data matrix to obtain a feature matrix;
[0007] Performing principal component analysis on the feature matrix to obtain fault features;
[0008] The fault characteristics are input into a fault monitoring model to obtain a fault identification result.
[0009] In one embodiment, performing empirical wavelet transform on the data matrix to obtain a feature matrix comprises:
[0010] Performing fast Fourier transform calculation on the time series data to obtain a time series spectrum of the time series data;
[0011] According to the time series spectrum, a demarcation frequency is obtained;
[0012] According to the demarcation frequency, construct an empirical function; wherein the empirical function includes an empirical scale function and an empirical wavelet function;
[0013] Performing an inner product operation on the time series data and the empirical function to obtain a feature vector; wherein the feature vector includes a scale coefficient and a wavelet coefficient;
[0014] A feature matrix is generated according to the feature vector.
[0015] In one embodiment, performing principal component analysis on the feature matrix to obtain fault features includes:
[0016] The feature matrix is centrally processed to obtain a covariance matrix;
[0017] Performing eigenvalue decomposition on the covariance matrix to obtain eigenvalue parameters;
[0018] According to the characteristic parameters, a principal component matrix is constructed;
[0019] The feature matrix is projected onto the principal component matrix to obtain the fault feature.
[0020] In one embodiment, inputting the fault feature into a fault monitoring model to obtain a fault identification result includes:
[0021] Inputting the fault characteristics into a fault monitoring model to obtain a time step parameter;
[0022] Performing a linear transformation calculation on the time step parameter to obtain an attention score;
[0023] Normalizing the attention scores to obtain time step scores;
[0024] A fault recognition result is obtained according to the time step score and the preset attention weight.
[0025] In one embodiment, the step of inputting the fault feature into a fault monitoring model before obtaining a fault identification result includes:
[0026] The fault features are reformatted to obtain sequence data.
[0027] In one of the embodiments, the method further includes: generating prompt information according to the fault identification result.
[0028] In a second aspect, the present application also provides a long short-term memory neural network fault monitoring device. The device comprises:
[0029] The acquisition module is used to acquire the time series data of the device and generate a data matrix;
[0030] A processing module, used for performing empirical wavelet transform on the data matrix to obtain a feature matrix;
[0031] An analysis module, used for performing principal component analysis on the feature matrix to obtain fault features;
[0032] The result module is used to input the fault characteristics into the fault monitoring model to obtain the fault identification result.
[0033] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:
[0034] Obtain the time series data of the device and generate a data matrix;
[0035] Performing empirical wavelet transform on the data matrix to obtain a feature matrix;
[0036] Performing principal component analysis on the feature matrix to obtain fault features;
[0037] The fault characteristics are input into a fault monitoring model to obtain a fault identification result.
[0038] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:
[0039] Obtain the time series data of the device and generate a data matrix;
[0040] Performing empirical wavelet transform on the data matrix to obtain a feature matrix;
[0041] Performing principal component analysis on the feature matrix to obtain fault features;
[0042] The fault characteristics are input into a fault monitoring model to obtain a fault identification result.
[0043] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:
[0044] Obtain the time series data of the device and generate a data matrix;
[0045] Performing empirical wavelet transform on the data matrix to obtain a feature matrix;
[0046] Performing principal component analysis on the feature matrix to obtain fault features;
[0047] The fault characteristics are input into a fault monitoring model to obtain a fault identification result.
[0048] The above-mentioned long short-term memory neural network fault monitoring method, device and computer equipment perform empirical wavelet transform processing on time series data, separate different frequency components to establish a feature matrix; perform principal component analysis on the feature matrix to obtain fault features; input the fault features into the fault monitoring model to obtain fault identification results. This method uses empirical wavelet transform to process signals, which helps to extract useful signals from noise, and then uses principal component analysis to reduce feature dimensionality, which can reduce the computational burden while retaining key features, thereby improving the fault identification speed while maintaining monitoring accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0049] Figure 1 An application environment diagram of a long short-term memory neural network fault monitoring method in an embodiment;
[0050] Figure 2 A schematic diagram of a process of a long short-term memory neural network fault monitoring method in one embodiment;
[0051] Figure 3 A schematic diagram of a fault monitoring model of a long short-term memory neural network fault monitoring method in an embodiment;
[0052] Figure 4 A schematic diagram of a flow chart of a long short-term memory neural network fault monitoring method in another embodiment;
[0053] Figure 5 It is a structural block diagram of a long short-term memory neural network fault monitoring device in one embodiment;
[0054] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION
[0055] 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.
[0056] The long short-term memory neural network fault monitoring method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The server 104 obtains the time series data of the equipment in the process simulation system through the network, performs empirical wavelet transform processing and principal component analysis processing on the time series data to obtain fault characteristics; the fault characteristics are input into the fault monitoring model to obtain fault identification results. Among them, the terminal 102 can be but is 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, sensors, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server 104 can be implemented with an independent server or a server cluster consisting of multiple servers.
[0057] In one embodiment, Figure 2 As shown in the figure, a long short-term memory neural network fault monitoring method is provided, and the method is applied to Figure 1 Taking the server 104 in the example as an example, the following steps are included:
[0058] Step 202, obtaining the timing data of the device and generating a data matrix.
[0059] Specifically, some key parameters of the equipment in the process simulation system are monitored in real time and the temperature, pressure, flow rate, speed and other key parameters of the equipment are recorded. These generated continuous time series data are stored in a data matrix, where each row represents a time point and each column corresponds to a specific parameter of a device.
[0060] Step 204: Perform empirical wavelet transform on the data matrix to obtain a feature matrix.
[0061] Specifically, in order to adaptively decompose the complex signals in the time series data and separate different frequency components, the data matrix is processed using an empirical wavelet transform algorithm, the time series data is spectrally divided, and an adaptive wavelet filter group is established to effectively extract useful signal components from noise and obtain a feature matrix.
[0062] The method of performing empirical wavelet transform on the data matrix to obtain a characteristic matrix includes: performing fast Fourier transform calculation on the time series data to obtain a time series spectrum of the time series data; obtaining a demarcation frequency based on the time series spectrum; constructing an empirical function based on the demarcation frequency; performing inner product operation on the time series data and the empirical function to obtain a characteristic vector; and generating a characteristic matrix based on the characteristic vector.
[0063] Wherein, the time series data is composed of multiple time series signals; the empirical function includes an empirical scaling function and an empirical wavelet function; and the characteristic vector includes a scaling coefficient and a wavelet coefficient.
[0064] Specifically, the time series data is processed for missing values and abnormal values, and abnormal signals in the time series data are removed to ensure data quality and consistency. The time series data is subjected to fast Fourier transform calculation to obtain the time series spectrum of the measured time series signal. The number M of amplitude modulation-frequency modulation (AM-FM) signals to be divided is set according to the time series spectrum, and M-1 dividing frequencies are found; and an empirical scaling function is constructed according to the dividing frequencies. and the empirical wavelet function The time series data is inner-producted with the empirical scaling function and the empirical wavelet function to obtain scaling coefficients and wavelet coefficients. The scaling coefficients and wavelet coefficients are regarded as the eigenvectors of the time series data, and the eigenvectors of all time series signals are combined into a feature matrix, in which each column represents the feature of a time series signal and each row represents a coefficient at a specific scale or frequency.
[0065] Step 206: Perform principal component analysis on the feature matrix to obtain fault features.
[0066] Specifically, the principal component analysis algorithm is used to map the feature matrix to a new feature space through linear transformation, thereby effectively reducing the dimension of the feature vector while retaining the most critical fault features.
[0067] The method of performing principal component analysis on the feature matrix to obtain fault features includes: centralizing the feature matrix to obtain a covariance matrix by calculation; performing eigenvalue decomposition on the covariance matrix to obtain characteristic parameters; constructing a principal component matrix based on the characteristic parameters; and projecting the feature matrix onto the principal component matrix to obtain the fault features.
[0068] The characteristic parameters include characteristic values and characteristic vectors corresponding to the characteristic values.
[0069] Specifically, the feature matrix is centralized, that is, the mean of each column of the feature matrix is subtracted from the feature vector to ensure that the center of the principal component analysis is the origin, and the corresponding covariance matrix is calculated for the centralized feature matrix. The covariance matrix is subjected to eigenvalue decomposition to obtain feature parameters. Feature parameters whose cumulative contribution rate is greater than a preset value are selected, and the principal component matrix is constructed with the selected feature parameters. The original feature matrix is projected onto the selected principal component to obtain the data representation after dimensionality reduction.
[0070] Step 208: input the fault characteristics into a fault monitoring model to obtain a fault identification result.
[0071] The step of inputting the fault feature into a fault monitoring model to obtain a fault identification result includes: reshaping the format of the fault feature to obtain sequence data.
[0072] Specifically, the data of the fault characteristics are reshaped into a three-dimensional array format to ensure that the feature matrix processed by the empirical wavelet transform and the principal component analysis is formatted into sequence data, which is suitable for use as the input of the attention mechanism long short-term memory neural network (Attention-LSTM). On this basis, a corresponding label is prepared for each time series signal to indicate the signal state, which can be a normal state or an abnormal state.
[0073] In one embodiment, the three-dimensional array format is input value, implicit value and model layer number. The format of the input value is input_size; the format of the implicit value is hidden_size; the format of the model layer number is num_layers.
[0074] The step of inputting the fault feature into the fault monitoring model to obtain the fault identification result includes: inputting the fault feature into the fault monitoring model to obtain a time step parameter; performing a linear transformation calculation on the time step parameter to obtain an attention score; normalizing the attention score to obtain a time step score; and obtaining the fault identification result based on the time step score and a preset attention weight.
[0075] Specifically, the fault monitoring model is divided into five layers: input layer, long short-term memory neural network (LSTM) layer, attention mechanism layer, fully connected layer, and output layer. The role of the long short-term memory neural network (LSTM) layer is to achieve high-level feature learning; the role of the attention mechanism layer is to highlight key information; the role of the fully connected layer is to integrate local features to achieve the final prediction.
[0076] In one embodiment, Figure 3 As shown, Figure 3 Schematic diagram of a fault monitoring model of a long short-term memory neural network fault monitoring method in an embodiment.
[0077] The fault monitoring model is as follows:
[0078] Input layer: accepts sequence data in the format of a three-dimensional array.
[0079] Long Short-Term Memory Neural Network (LSTM) layer: Set to return the output of all time steps so that each time series signal corresponds to one time step.
[0080] Fully connected layer: Maps the input feature matrix to a new feature space for optimization to improve the performance of the model. The attention score is obtained by calculating the activation function Relu, which enhances the nonlinearity of the neural network model and can represent more complex patterns. The softmax function is applied to normalize the attention score to obtain the time step score.
[0081] Attention layer: multiplies the time step score by the preset attention weight so that the model will focus on the time series signals with higher time step weights when making the final prediction.
[0082] Output layer: used to output fault identification results.
[0083] The method further includes: generating prompt information according to the fault identification result.
[0084] Specifically, it is detected whether there is a device fault in the fault identification result; if there is a device fault, a fault warning prompt is generated; if there is no device fault, a device normal prompt is generated.
[0085] In the above-mentioned long short-term memory neural network fault monitoring method, the time series data is processed by empirical wavelet transform, and different frequency components are separated to establish a feature matrix; the feature matrix is subjected to principal component analysis to obtain fault features; the fault features are input into the fault monitoring model to obtain fault identification results. This method uses empirical wavelet transform to process signals, which helps to extract useful signals from noise, and then uses principal component analysis to reduce feature dimensionality, which can reduce the computational burden while retaining key features, thereby improving the fault identification speed while maintaining monitoring accuracy.
[0086] In one embodiment, Figure 4 As shown, Figure 4 A flow chart of a long short-term memory neural network fault monitoring method in another embodiment includes the following steps:
[0087] Step 402: Offline modeling.
[0088] Specifically, historical time series data is obtained, and data preprocessing, empirical wavelet transform signal decomposition (EWT signal decomposition) and principal component analysis dimensionality reduction (PCA dimensionality reduction) are performed on the historical time series data to generate training set samples. The training set samples are input into the LSTM model with integrated attention mechanism for training to generate a fault monitoring model.
[0089] Step 404: Online fault monitoring.
[0090] Specifically, real-time time series data is obtained, and the real-time time series data is subjected to data preprocessing, empirical wavelet transform signal decomposition (EWT signal decomposition) and principal component analysis dimensionality reduction (PCA dimensionality reduction) to generate fault features. The fault features are input into the LSTM monitoring model with integrated attention mechanism, i.e., the fault monitoring model, to obtain the fault identification result. According to the fault identification result, it is determined whether a fault occurs; if a fault occurs, a fault warning prompt is generated; if no fault occurs, the data at the next moment is continuously monitored.
[0091] In this embodiment, the empirical wavelet transform is used to decompose complex signals, separate different frequency components, and then the principal component analysis is used to reduce the feature dimension. By integrating different algorithms and technologies, the fault diagnosis process is made more accurate and reliable, the fault mode is more accurately identified and classified, the interpretability of the model is enhanced, and the key time points for classification decisions can be automatically identified for real-time data, the accuracy of fault monitoring is improved, and false positives and missed positives are reduced.
[0092] 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.
[0093] Based on the same inventive concept, the embodiment of the present application also provides a long short-term memory neural network fault monitoring device for implementing the long short-term memory neural network fault monitoring method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more long short-term memory neural network fault monitoring device embodiments provided below can refer to the limitations of the long short-term memory neural network fault monitoring method above, and will not be repeated here.
[0094] In one embodiment, Figure 5 As shown, a long short-term memory neural network fault monitoring device is provided, including: an acquisition module 510, a processing module 520, an analysis module 530 and a result module 540, wherein:
[0095] The acquisition module 510 is used to acquire the time series data of the device and generate a data matrix.
[0096] The processing module 520 is used to perform empirical wavelet transform on the data matrix to obtain a feature matrix.
[0097] The analysis module 530 is used to perform principal component analysis on the feature matrix to obtain fault features.
[0098] The result module 540 is used to input the fault characteristics into the fault monitoring model to obtain a fault identification result.
[0099] The processing module 520 is also used to perform fast Fourier transform calculation on the time series data to obtain the time series spectrum of the time series data; obtain the demarcation frequency based on the time series spectrum; construct an empirical function based on the demarcation frequency; perform inner product operation on the time series data and the empirical function to obtain a eigenvector; wherein the eigenvector includes a scale coefficient and a wavelet coefficient; and generate a characteristic matrix based on the eigenvector.
[0100] The analysis module 530 is also used to centralize the feature matrix and calculate the covariance matrix; perform eigenvalue decomposition on the covariance matrix to obtain feature parameters; construct a principal component matrix based on the feature parameters; and project the feature matrix onto the principal component matrix to obtain the fault features.
[0101] The result module 540 is also used to input the fault feature into the fault monitoring model to obtain a time step parameter; perform linear transformation calculation on the time step parameter to obtain an attention score; normalize the attention score to obtain a time step score; and obtain a fault identification result based on the time step score and a preset attention weight.
[0102] The result module 540 is also used to reformat the fault features to obtain sequence data.
[0103] The result module 540 is further configured to generate prompt information according to the fault identification result.
[0104] Each module in the above-mentioned long short-term memory neural network fault monitoring device can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.
[0105] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 6As shown. The computer device includes a processor, a memory and a network interface 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, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store process simulation data. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a long short-term memory neural network fault monitoring method is implemented.
[0106] 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.
[0107] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, any one of the long short-term memory neural network fault monitoring methods in the above embodiments is implemented.
[0108] Obtain the time series data of the device and generate a data matrix;
[0109] Performing empirical wavelet transform on the data matrix to obtain a feature matrix;
[0110] Performing principal component analysis on the feature matrix to obtain fault features;
[0111] The fault characteristics are input into a fault monitoring model to obtain a fault identification result.
[0112] 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, any one of the long short-term memory neural network fault monitoring methods in the above embodiments is implemented.
[0113] Obtain the time series data of the device and generate a data matrix;
[0114] Performing empirical wavelet transform on the data matrix to obtain a feature matrix;
[0115] Performing principal component analysis on the feature matrix to obtain fault features;
[0116] The fault characteristics are input into the fault monitoring model to obtain a fault identification result. In one embodiment, a computer program product is provided, including a computer program, which, when executed by a processor, implements the following steps:
[0117] Obtain the time series data of the device and generate a data matrix;
[0118] Performing empirical wavelet transform on the data matrix to obtain a feature matrix;
[0119] Performing principal component analysis on the feature matrix to obtain fault features;
[0120] The fault characteristics are input into a fault monitoring model to obtain a fault identification result.
[0121] 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.
[0122] 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.
[0123] 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.
[0124] 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 long short-term memory neural network fault monitoring method, characterized in that: The method comprises: Obtain the time series data of the device and generate a data matrix; Performing empirical wavelet transform on the data matrix to obtain a feature matrix; Performing principal component analysis on the feature matrix to obtain fault features; The fault characteristics are input into a fault monitoring model to obtain a fault identification result.
2. The method according to claim 1, characterized in that: The step of performing empirical wavelet transform on the data matrix to obtain a feature matrix comprises: Performing fast Fourier transform calculation on the time series data to obtain a time series spectrum of the time series data; According to the time series spectrum, a demarcation frequency is obtained; According to the demarcation frequency, construct an empirical function; wherein the empirical function includes an empirical scale function and an empirical wavelet function; Performing an inner product operation on the time series data and the empirical function to obtain a feature vector; wherein the feature vector includes a scale coefficient and a wavelet coefficient; A feature matrix is generated according to the feature vector.
3. The method according to claim 2, characterized in that The performing principal component analysis on the feature matrix to obtain fault features comprises: The feature matrix is centrally processed to obtain a covariance matrix; Performing eigenvalue decomposition on the covariance matrix to obtain eigenvalue parameters; According to the characteristic parameters, a principal component matrix is constructed; The feature matrix is projected onto the principal component matrix to obtain the fault feature.
4. The method according to claim 1, characterized in that: Inputting the fault feature into the fault monitoring model to obtain the fault identification result comprises: Inputting the fault characteristics into a fault monitoring model to obtain a time step parameter; Performing a linear transformation calculation on the time step parameter to obtain an attention score; Normalizing the attention scores to obtain time step scores; A fault recognition result is obtained according to the time step score and the preset attention weight.
5. The method according to any one of claims 1 to 4, characterized in that: The step of inputting the fault feature into the fault monitoring model to obtain the fault identification result includes: The fault features are reformatted to obtain sequence data.
6. The method according to claim 1, characterized in that The method further comprises: Generate prompt information according to the fault identification result.
7. A long short-term memory neural network fault monitoring device, characterized in that: The device comprises: The acquisition module is used to acquire the time series data of the device and generate a data matrix; A processing module, used for performing empirical wavelet transform on the data matrix to obtain a feature matrix; An analysis module, used for performing principal component analysis on the feature matrix to obtain fault features; The result module is used to input the fault characteristics into the fault monitoring model to obtain the fault identification result.
8. 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 6 are implemented.
9. 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 6 are implemented.
10. A computer program product, comprising a computer program, 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 6 are implemented.