CN118035889A - Method, device, equipment, storage medium and product for fault classification of rotary equipment - Google Patents

Method, device, equipment, storage medium and product for fault classification of rotary equipment Download PDF

Info

Publication number
CN118035889A
CN118035889A CN202410173763.XA CN202410173763A CN118035889A CN 118035889 A CN118035889 A CN 118035889A CN 202410173763 A CN202410173763 A CN 202410173763A CN 118035889 A CN118035889 A CN 118035889A
Authority
CN
China
Prior art keywords
fault
attention
output information
sample
graph
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202410173763.XA
Other languages
Chinese (zh)
Inventor
闫循石
时振刚
周燕
赵晶晶
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tsinghua University
Original Assignee
Tsinghua University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tsinghua University filed Critical Tsinghua University
Priority to CN202410173763.XA priority Critical patent/CN118035889A/en
Publication of CN118035889A publication Critical patent/CN118035889A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/2431—Multiple classes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/253—Fusion techniques of extracted features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0499—Feedforward networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Biomedical Technology (AREA)
  • Health & Medical Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

The application relates to a fault classification method, a fault classification device, a fault classification storage medium and a fault classification product for rotary equipment. The method comprises the following steps: extracting characteristics of fault information of the rotary equipment to be detected to obtain fault characteristics of the fault information; inputting fault characteristics into a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model to obtain fault types of the rotating equipment to be tested; the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model is obtained by training an initial multi-scale multi-head self-attention GCN model based on a fusion graph feature sample, wherein the fusion graph feature sample is obtained by fusing target output information and a plurality of graph feature samples, the target output information is determined by the initial multi-scale multi-head self-attention GCN model based on a fault information sample, and the graph feature sample is determined by the fault information sample, an adjacent matrix and the initial multi-scale multi-head self-attention GCN model. By adopting the method, the fault detection precision can be improved.

Description

Method, device, equipment, storage medium and product for fault classification of rotary equipment
Technical Field
The present application relates to the field of machine learning technologies, and in particular, to a fault classification method, apparatus, device, storage medium, and product for a rotating device.
Background
Common rotating equipment such as motors, engines, etc. are prone to failure during operation, and therefore, it is necessary to perform failure detection on the rotating equipment.
With the development of machine learning technology, learning models such as support vector machines and graph convolution neural networks are increasingly applied to fault detection of rotating equipment.
When the traditional graph convolution neural network is trained, because a single graph network is used for detection, enough information is difficult to extract from limited training samples, and therefore the fault detection precision of the graph convolution neural network model obtained through training is low.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a fault classification method, apparatus, device, medium, and product of a rotary device capable of improving fault detection accuracy.
In a first aspect, the present application provides a method for fault classification of a rotating device. The method comprises the following steps:
extracting characteristics of fault information of the rotary equipment to be detected to obtain fault characteristics of the fault information;
Inputting fault characteristics into a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model to obtain fault types of the rotating equipment to be tested; the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model is obtained by training an initial multi-scale multi-head self-attention GCN model based on a fusion graph feature sample, wherein the fusion graph feature sample is obtained by fusing target output information and a plurality of graph feature samples, the target output information is determined by the initial multi-scale multi-head self-attention GCN model based on a fault information sample, and the graph feature sample is determined by the fault information sample, an adjacent matrix and the initial multi-scale multi-head self-attention GCN model.
In one embodiment, an initial multi-scale multi-head self-attention GCN model includes a plurality of first sub-models and a multi-layer perceptron; the method further comprises the steps of:
Inputting fault information samples and adjacent matrixes corresponding to the first sub-models into the first sub-models aiming at the first sub-models to obtain graph feature samples output by the first sub-models;
inputting the fault information sample into a multi-layer sensor to obtain first output information; the target output information includes first output information;
Obtaining a fusion graph characteristic sample according to the first output information and each graph characteristic sample;
training the initial multi-scale multi-head self-attention GCN model according to the fusion map feature sample to obtain a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model.
In one embodiment, the first sub-model comprises a first graph convolution layer and a second graph convolution layer, and the initial multi-scale multi-head self-attention GCN model comprises a connection layer and a linear layer; obtaining a fusion graph feature sample according to the first output information and each graph feature sample, including:
Aiming at each first sub-model, inputting a fault information sample and an adjacent matrix corresponding to each first sub-model into a first graph convolution layer of the first sub-model to obtain a first intermediate graph characteristic;
Inputting first intermediate graph features obtained by the first graph convolution layers of the first sub-models into the connecting layer to obtain second intermediate graph features;
inputting the second intermediate graph characteristics into a linear layer to obtain second output information;
obtaining a fusion graph characteristic sample according to the first output information, the second output information and each graph characteristic sample; the target output information includes first output information and second output information.
In one embodiment, a fused graph feature sample is obtained according to the first output information, the second output information and each graph feature sample; the target output information includes first output information and second output information, including:
Obtaining the first output information, the second output information and the weighted value of each graph feature sample according to the first output information, the second output information, the first weight corresponding to the first output information, the second weight corresponding to the second output information and the third weight corresponding to each graph feature sample;
and obtaining a fusion map feature sample based on the weighted value.
In one embodiment, the number of fusion map feature samples is a plurality; the initial multi-scale multi-head self-attention GCN model comprises a multi-head self-attention fusion model and a full connection layer; training an initial multi-scale multi-head self-attention GCN model according to the fusion map feature sample to obtain a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model, wherein the method comprises the following steps:
Connecting the fusion map feature samples by utilizing a multi-head self-attention fusion model to obtain map feature vectors;
obtaining a predicted fault category based on the graph feature vector by using the full connection layer;
And training the initial multi-scale multi-head self-attention GCN model according to the error between the predicted fault category and the real fault category corresponding to the fault information sample to obtain a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model.
In one embodiment, training the initial multi-scale multi-head self-attention GCN model to obtain a target multi-scale multi-head self-attention seeking-convolved neural network GCN model according to an error between a predicted fault category and a true fault category corresponding to a fault information sample, including:
Determining a loss value according to an error between the predicted fault category and a real fault category corresponding to the fault information sample;
Under the condition that the loss value is larger than a preset threshold value, adjusting parameters of the initial multi-scale multi-head self-attention GCN model according to the loss value to obtain an intermediate GCN model, training the initial multi-scale multi-head self-attention GCN model until the loss value is smaller than the preset threshold value, and taking the intermediate GCN model corresponding to the loss value smaller than the preset threshold value as a target multi-scale multi-head self-attention power seeking convolutional neural network GCN model.
In a second aspect, the application further provides a fault classification device of the rotating equipment. The device comprises:
the feature extraction module is used for extracting features of fault information of the rotary equipment to be detected to obtain fault features of the fault information;
The fault classification module is used for inputting fault characteristics into a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model to obtain fault types of the rotating equipment to be detected; the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model is obtained by training an initial multi-scale multi-head self-attention GCN model based on a fusion graph feature sample, wherein the fusion graph feature sample is obtained by fusing target output information and a plurality of graph feature samples, the target output information is determined by the initial multi-scale multi-head self-attention GCN model based on a fault information sample, and the graph feature sample is determined by the fault information sample, an adjacent matrix and the initial multi-scale multi-head self-attention GCN model.
In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor implementing the steps of the method of any of the above-mentioned first aspects when the processor executes the computer program.
In a fourth aspect, the present application also provides a computer-readable storage medium. A computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the method of any of the above-mentioned first aspects.
In a fifth aspect, the present application also provides a computer program product. A computer program product comprising a computer program which, when executed by a processor, implements the steps of the method of any of the first aspects described above.
According to the fault classification method, the device, the equipment, the storage medium and the product of the rotating equipment, the target output information and the multiple image feature samples are fused to obtain the fused image feature samples, then the initial multi-scale multi-head self-attention GCN model is trained to obtain the target multi-scale multi-head self-attention convolutional neural network GCN model based on the fused image feature samples, and because the fused image feature samples comprise the target output information and the multiple image feature samples, the information contained in the fused image feature samples is richer than that contained in the single image feature sample in the traditional technology, the prediction precision of the target multi-scale multi-head self-attention convolutional neural network GCN model obtained based on the training of the fused image feature samples with richer information is higher, and further, the fault detection precision can be improved.
Drawings
FIG. 1 is an application environment diagram of a method of fault classification for a rotating device in one embodiment;
FIG. 2 is a flow diagram of a method of fault classification for a rotating device in one embodiment;
FIG. 3 is a schematic diagram of a patterning process in one embodiment;
Fig. 4 is a schematic diagram of a model corresponding to k=2 and a model corresponding to k=4 in one embodiment;
FIG. 5 is a flow chart of training a target multi-scale multi-head self-care force diagram convolutional neural network GCN model in one embodiment;
FIG. 6 is a schematic flow diagram of training an initial multi-scale multi-head self-attention GCN model according to a fusion map feature sample to obtain a target multi-scale multi-head self-attention seeking graph convolutional neural network GCN model in one embodiment;
FIG. 7 is a flow chart of a method of fault classification for a rotating device in an exemplary embodiment;
FIG. 8 is a schematic diagram of a target multi-scale multi-headed self-care seeking-convolved neural network GCN model in one illustrative embodiment;
FIG. 9 is a schematic diagram of a multi-headed self-attention fusion model in accordance with an exemplary embodiment;
FIG. 10 is a schematic diagram of a fault classification device of a rotary apparatus in one embodiment;
FIG. 11 is an internal block diagram of a server in one embodiment;
Fig. 12 is an internal structural diagram of a terminal in one embodiment.
Detailed Description
The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.
The fault classification method of the rotating equipment provided by the embodiment of the application can be applied to an application environment shown in fig. 1. The computer equipment 102 performs feature extraction on fault information of the rotating equipment to be detected to obtain fault features of the fault information; inputting fault characteristics into a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model to obtain fault types of the rotating equipment to be tested; the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model is obtained by training an initial multi-scale multi-head self-attention GCN model based on a fusion graph feature sample, wherein the fusion graph feature sample is obtained by fusing target output information and a plurality of graph feature samples, the target output information is determined by the initial multi-scale multi-head self-attention GCN model based on a fault information sample, and the graph feature sample is determined by the fault information sample, an adjacent matrix and the initial multi-scale multi-head self-attention GCN model. The computer device 102 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, internet of things devices, and portable wearable devices, and the internet of things devices may be smart speakers, smart televisions, smart air conditioners, smart vehicle devices, and the like. The portable wearable device may be a smart watch, smart bracelet, headset, or the like.
In one embodiment, as shown in fig. 2, a fault classification method of a rotating device is provided, and the method is applied to the computer device 102 in fig. 1 for illustration, and includes the following steps:
And 202, extracting characteristics of fault information of the rotary equipment to be detected to obtain fault characteristics of the fault information.
The rotating device to be tested may be any one of a motor, an engine, and the like, which is not limited in this embodiment.
Optionally, firstly, acquiring fault signals of the rotating equipment to be tested by using sensors such as a vibration sensor, a current sensor and the like installed in the rotating equipment to be tested, and segmenting the fault signals of the rotating equipment to be tested to obtain fault information of the rotating equipment to be tested; and then, extracting the characteristics of the fault information of the rotary equipment to be detected to obtain the fault characteristics of the fault information. The feature extraction method may be a fast fourier transform method or a normalization method, or may be other methods, which is not limited in this embodiment.
Step 204, inputting fault characteristics into a target multi-scale multi-head self-care force diagram convolutional neural network GCN model to obtain fault types of the rotating equipment to be tested; the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model is obtained by training an initial multi-scale multi-head self-attention GCN model based on a fusion graph feature sample, wherein the fusion graph feature sample is obtained by fusing target output information and a plurality of graph feature samples, the target output information is determined by the initial multi-scale multi-head self-attention GCN model based on a fault information sample, and the graph feature sample is determined by the fault information sample, an adjacent matrix and the initial multi-scale multi-head self-attention GCN model.
Alternatively, patterning is required when using the initial multi-scale multi-headed self-attention GCN model, since there is no displayed correlation between the fault information samples. The method comprises the steps of taking a certain fault information sample as a central node, calculating the pairwise distances between the central node and other fault information samples by using a KNN (K Neighbor) algorithm, taking K fault information samples nearest to the central node as Neighbor nodes related to the central node, and describing the association relation between the fault information samples on the graph by using an adjacency matrix A, wherein the formula of the adjacency matrix A is shown in a formula (1).
(1)
In formula (1), x i is a center node, and x j is a neighboring node associated with the center node.
Alternatively, different K values may be taken to construct different graphs, and the number of constructed graphs may be an integer greater than or equal to 2, which is not limited in this embodiment. Assume that the number of the constructed graphs is 2,K, and the values are 2 and 4, respectively, and the patterning process is shown in fig. 3. When k=2, the failure information sample X 1 is taken as a central node, the neighboring nodes associated with X 1 are X 2 and X 3, and the adjacency matrix is A1. When k=4, the failure information sample X 5 is taken as the center node, the neighboring node associated with X 5 is X 1、X2、X3、X4, and the adjacency matrix is A2.
Fig. 4 shows a schematic diagram of a corresponding model at k=2 and a corresponding model at k=4, where the corresponding model at k=2 is denoted as a first sub-model 1, the first sub-model 1 comprises two stacks, respectively, stack 1-1 and stack 1-2,K =4, and the corresponding model at k=2 is denoted as a first sub-model 2, and the first sub-model 2 comprises two stacks, respectively, stack 2-1 and stack 2-2.
The fault information sample and the adjacent matrix A1 are input into the first sub-model 1 to obtain a graph characteristic sample 1, and the fault information sample and the adjacent matrix A2 are input into the first sub-model 2 to obtain a graph characteristic sample 2. And inputting the fault information sample into a target network in the initial multi-scale multi-head self-attention GCN model to obtain target output information. The target network in the initial multi-scale multi-head self-attention GCN model can be a multi-layer perceptron, or a sub-network formed by a connecting layer and a linear layer, or the multi-layer perceptron and the sub-network, that is, the target output information can be first output information output by the multi-layer perceptron, or the target output information can also be second output information output by the sub-network, or the target output information can comprise the first output information and the second output information. This embodiment is not limited thereto.
And fusing the target output information and the multiple graph feature samples to obtain fused graph feature samples, and training an initial multi-scale multi-head self-attention GCN model based on the fused graph feature samples to obtain a target multi-scale multi-head self-attention seeking convolutional neural network GCN model. The fusion may be weighted summation or other calculation method, which is not limited in this embodiment.
And inputting the fault characteristics into a target multiscale multi-head self-attention force diagram convolutional neural network GCN model to obtain the fault type of the rotary equipment to be tested.
In the fault classification method of the rotating equipment, the fault information of the rotating equipment to be detected is subjected to feature extraction to obtain the fault feature of the fault information; inputting fault characteristics into a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model to obtain fault types of the rotating equipment to be tested; the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model is obtained by training an initial multi-scale multi-head self-attention GCN model based on a fusion graph feature sample, wherein the fusion graph feature sample is obtained by fusing target output information and a plurality of graph feature samples, the target output information is determined by the initial multi-scale multi-head self-attention GCN model based on a fault information sample, and the graph feature sample is determined by the fault information sample, an adjacent matrix and the initial multi-scale multi-head self-attention GCN model. The fusion graph characteristic sample comprises target output information and a plurality of graph characteristic samples, and contains more information than the single graph characteristic sample in the prior art, so that the prediction precision of the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model obtained based on the fusion graph characteristic sample training with more abundant information is higher, and further, the fault detection precision can be improved.
In one embodiment, the initial multi-scale multi-headed self-attention GCN model includes a plurality of first sub-models and a multi-layer perceptron, and the process of training to obtain the target multi-scale multi-headed self-attention seeking-to-convolve neural network GCN model is shown in FIG. 5, and includes:
Step 502, for each first sub-model, inputting the fault information sample and the adjacent matrix corresponding to each first sub-model into the first sub-model to obtain a graph feature sample output by the first sub-model.
Optionally, assuming that the initial multi-scale multi-head self-attention GCN model includes 2 first sub-models, namely a first sub-model 1 and a first sub-model 2, inputting the fault information sample and an adjacent matrix corresponding to the first sub-model 1 into the first sub-model 1 to obtain a graph feature sample 1 output by the first sub-model. And inputting the fault information sample and the adjacent matrix corresponding to the first sub-model 2 into the first sub-model 2 to obtain a graph characteristic sample 2 output by the first sub-model.
Step 504, inputting the fault information sample to the multi-layer sensor to obtain first output information; the target output information includes first output information.
Optionally, the fault information sample is input to the multi-layer sensor to obtain first output information, where the first output information is shown in formula (2).
(2)
In the formula (2),For the corresponding characteristics of the fault information sample,/>Is the first output information.
And step 506, obtaining a fusion graph characteristic sample according to the first output information and each graph characteristic sample.
Optionally, determining a first weight corresponding to the first output information, a third weight corresponding to each graph feature sample, adding products of the first weight and the first output information and products of each graph feature sample and the third weight, and taking the added result as a fusion graph feature sample.
And step 508, training the initial multi-scale multi-head self-attention GCN model according to the fusion map feature sample to obtain a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model.
Alternatively, the number of feature samples of the fusion map may be one or more, which is not limited in this embodiment. If the number of the fusion graph feature samples is one, a predicted fault class is obtained based on the fusion graph feature samples by utilizing a full-connection layer in the initial multi-scale multi-head self-attention GCN model, and the initial multi-scale multi-head self-attention GCN model is trained according to errors between the predicted fault class and a real fault class corresponding to the fault information samples to obtain the target multi-scale multi-head self-attention power-attempt convolutional neural network GCN model.
In the embodiment, for each first sub-model, a fault information sample and an adjacent matrix corresponding to each first sub-model are input into the first sub-model to obtain a graph characteristic sample output by the first sub-model; inputting the fault information sample into a multi-layer sensor to obtain first output information; the target output information includes first output information; obtaining a fusion graph characteristic sample according to the first output information and each graph characteristic sample; training the initial multi-scale multi-head self-attention GCN model according to the fusion map feature sample to obtain a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model. The first output information and the plurality of graph feature samples are fused to obtain a fused graph feature sample, and then the initial multi-scale multi-head self-attention GCN model is trained to obtain the target multi-scale multi-head self-attention seeking convolutional neural network GCN model based on the fused graph feature sample.
In one embodiment, the first sub-model includes a first graph convolution layer and a second graph convolution layer, and the initial multi-scale multi-headed self-attention GCN model includes a connection layer and a linear layer; obtaining a fusion graph feature sample according to the first output information and each graph feature sample, including:
Aiming at each first sub-model, inputting a fault information sample and an adjacent matrix corresponding to each first sub-model into a first graph convolution layer of the first sub-model to obtain a first intermediate graph characteristic;
Inputting first intermediate graph features obtained by the first graph convolution layers of the first sub-models into the connecting layer to obtain second intermediate graph features;
And inputting the second intermediate graph characteristic into the linear layer to obtain second output information.
Optionally, assuming that the initial multi-scale multi-head self-attention GCN model includes 2 first sub-models, namely a first sub-model 1 and a first sub-model 2, inputting an adjacent matrix corresponding to a fault information sample and the first sub-model 1 into a first graph convolution layer of the first sub-model 1 to obtain a first intermediate graph feature 1, inputting an adjacent matrix corresponding to a fault information sample and the first sub-model 2 into the first graph convolution layer of the first sub-model 2 to obtain a first intermediate graph feature 2, inputting the first intermediate graph feature 1 and the first intermediate graph feature 2 into a connection layer to obtain a second intermediate graph feature, inputting the second intermediate graph feature into a linear layer to obtain second output information, and a formula of the second output information is shown in a formula (3).
(3)
In the formula (3),For the first intermediate graph feature 1,/>Features 2, w and b are trainable parameters for the first intermediate graph.
Obtaining a fusion graph characteristic sample according to the first output information, the second output information and each graph characteristic sample; the target output information includes first output information and second output information.
Alternatively, the first output information, the second output information, and each graph feature sample may be weighted and summed, and the result of the weighted and summed may be used as a fused graph feature sample.
In the embodiment, for each first sub-model, a fault information sample and an adjacent matrix corresponding to each first sub-model are input to a first graph convolution layer of the first sub-model to obtain a first intermediate graph feature; inputting first intermediate graph features obtained by the first graph convolution layers of the first sub-models into the connecting layer to obtain second intermediate graph features; inputting the second intermediate graph characteristics into a linear layer to obtain second output information; obtaining a fusion graph characteristic sample according to the first output information, the second output information and each graph characteristic sample; the target output information includes first output information and second output information. The fusion graph feature samples comprise first output information, second output information and each graph feature sample, and the contained information is richer than that of a single graph feature sample in the prior art.
In one embodiment, a fused graph feature sample is obtained according to the first output information, the second output information and each graph feature sample; the target output information includes first output information and second output information, including:
Obtaining the first output information, the second output information and the weighted value of each graph feature sample according to the first output information, the second output information, the first weight corresponding to the first output information, the second weight corresponding to the second output information and the third weight corresponding to each graph feature sample;
and obtaining a fusion map feature sample based on the weighted value.
Optionally, formulas for calculating the first weight corresponding to the first output information, the second weight corresponding to the second output information, and the third weight corresponding to each graph feature sample are shown in formulas (4) and (5).
(4)
(5)
In the formulas (4) and (5), assuming that the number of the graph characteristic samples is 2, the values of i are 1,2, 3, 4,Z 1, Z 2, Z 3, Z 4,And/>Third weight corresponding to each graph characteristic sample,/>For the first weight corresponding to the first output information,/>And the second weight is corresponding to the second output information. Wherein/>The sum of (2) is 1.
And obtaining the weighted values of the first output information, the second output information and the characteristic samples of each graph according to the first output information, the second output information, the first weight corresponding to the first output information, the second weight corresponding to the second output information and the third weight corresponding to the characteristic samples of each graph, wherein a calculation formula is shown in a formula (6), and the weighted values are used as fusion characteristic samples.
(6)
In the formula (6) of the present invention,And representing a fusion map feature sample.
In this embodiment, according to the first output information, the second output information, each graph feature sample, the first weight corresponding to the first output information, the second weight corresponding to the second output information, and the third weight corresponding to each graph feature sample, the weighted values of the first output information, the second output information, and each graph feature sample are obtained; and obtaining a fusion map feature sample based on the weighted value. By giving different weights to the first output information, the second output information and the characteristic samples of each graph, the first output information, the second output information and the characteristic samples of each graph can be adaptively fused, and the phenomenon of overcomplete caused by a deep network is avoided while the information is enriched.
In one embodiment, the number of fusion map feature samples is a plurality; the initial multi-scale multi-head self-attention GCN model comprises a multi-head self-attention fusion model and a full connection layer; training the initial multi-scale multi-head self-attention GCN model according to the fusion graph characteristic sample to obtain a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model, wherein the flow is shown in figure 6 and comprises the following steps:
And step 602, connecting the feature samples of each fusion map by using a multi-head self-attention fusion model to obtain a map feature vector.
Optionally, in order to make the characterization capability of the fusion map feature sample stronger, multiple sets of trainable parameters can be setBased on formulas (4), (5) and (6), a plurality of fusion map feature samples/>, are calculatedWherein, the method comprises the steps of, wherein,. And connecting a plurality of fusion map feature samples by using a multi-head self-attention fusion model, and obtaining a map feature vector as shown in a formula (7).
(7)
In the formula (7) of the present invention,Is a graph feature vector.
Step 604, obtaining the predicted fault class based on the graph feature vector by using the full connection layer.
Optionally, the graph feature vector is subjected to linear transformation and Softmax function processing by using the full connection layer, so as to obtain a predicted fault class, as shown in a formula (8).
(8)
In the formula (8), the expression "a",Representing the graph feature vectors, W and b representing the trainable parameters, and y representing the predicted fault class.
Step 606, training the initial multi-scale multi-head self-attention GCN model according to the error between the predicted fault category and the real fault category corresponding to the fault information sample to obtain a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model.
Optionally, describing an error between the predicted fault category and the real fault category corresponding to the fault information sample by using a loss function, and training the initial multi-scale multi-head self-attention GCN model according to a loss value corresponding to the error until the loss value meets a preset condition to obtain the target multi-scale multi-head self-attention power-driven convolutional neural network GCN model. Here, the loss function may be cross entropy, or may be another function, which is not limited in this embodiment. The preset condition here may be that the loss value is smaller than a preset threshold value, or that the loss value is equal to a preset threshold value, which is not limited in this embodiment.
In the embodiment, a multi-head self-attention fusion model is utilized to connect the feature samples of each fusion graph to obtain a graph feature vector; obtaining a predicted fault category based on the graph feature vector by using the full connection layer; and training the initial multi-scale multi-head self-attention GCN model according to the error between the predicted fault category and the real fault category corresponding to the fault information sample to obtain a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model. By introducing a multi-head self-attention fusion model, the characteristic sample of the fusion map has stronger characterization capability, and the accuracy of the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model is further improved.
In one embodiment, training the initial multi-scale multi-head self-attention GCN model to obtain a target multi-scale multi-head self-attention seeking convolutional neural network GCN model according to an error between a predicted fault class and a true fault class corresponding to a fault information sample, including:
and determining a loss value according to the error between the predicted fault class and the real fault class corresponding to the fault information sample.
Optionally, the error between the predicted fault class and the true fault class corresponding to the fault information sample is described using cross entropy, and the loss value is determined according to the cross entropy.
Under the condition that the loss value is larger than a preset threshold value, adjusting parameters of the initial multi-scale multi-head self-attention GCN model according to the loss value to obtain an intermediate GCN model, training the initial multi-scale multi-head self-attention GCN model until the loss value is smaller than the preset threshold value, and taking the intermediate GCN model corresponding to the loss value smaller than the preset threshold value as a target multi-scale multi-head self-attention power seeking convolutional neural network GCN model.
Optionally, if the loss value is greater than a preset threshold, adjusting parameters of the initial multi-scale multi-head self-attention GCN model to obtain an intermediate GCN model, determining a graph feature sample output by each first sub-model in the intermediate GCN model based on the fault information sample, an adjacent matrix corresponding to each first sub-model in the intermediate GCN model and each first sub-model in the intermediate GCN model, determining target output information corresponding to the intermediate GCN model based on the fault information sample and the intermediate GCN model, fusing the graph feature sample output by each first sub-model in the intermediate GCN model with the target output information corresponding to the intermediate GCN model to obtain a fused feature sample corresponding to the intermediate GCN model, connecting the fused feature samples corresponding to the plurality of intermediate GCN models by using the multi-head self-attention fusion model to obtain a graph feature vector corresponding to the intermediate GCN model, obtaining a predicted fault class corresponding to the intermediate GCN model based on the graph feature vector corresponding to the intermediate GCN model by using the full connection layer, determining a new loss value according to an error between the predicted fault class corresponding to the intermediate GCN model and a real fault class corresponding to the fault information, and if the new preset loss value is smaller than the multi-scale self-attention GCN model, and taking the multi-scale self-attention neural network model as the target loss value.
In the embodiment, a loss value is determined according to an error between a predicted fault class and a real fault class corresponding to a fault information sample; under the condition that the loss value is larger than a preset threshold value, adjusting parameters of the initial multi-scale multi-head self-attention GCN model according to the loss value to obtain an intermediate GCN model, training the initial multi-scale multi-head self-attention GCN model until the loss value is smaller than the preset threshold value, and taking the intermediate GCN model corresponding to the loss value smaller than the preset threshold value as a target multi-scale multi-head self-attention power seeking convolutional neural network GCN model. By comparing the magnitude relation between the loss value and the preset threshold value, the parameters of the initial multi-scale multi-head self-attention GCN model are continuously adjusted, so that the predicted value of the model is closer to the true value, and the accuracy of the target multi-scale multi-head self-attention power-seeking convolutional neural network GCN model is improved.
In an exemplary embodiment, a fault classification method of a rotating device is provided, and a flow is shown in fig. 7, including:
in step 701, for each first sub-model, the fault information sample and the adjacent matrix corresponding to each first sub-model are input into the first sub-model, so as to obtain a graph feature sample output by the first sub-model.
Step 702, inputting a fault information sample to a multi-layer sensor to obtain first output information; the target output information includes first output information.
Step 703, for each first sub-model, inputting the fault information sample and the adjacent matrix corresponding to each first sub-model to the first graph convolution layer of the first sub-model, so as to obtain a first intermediate graph feature.
Step 704, inputting the first intermediate graph features obtained by the first graph convolution layer of each first sub-model to the connection layer to obtain second intermediate graph features.
Step 705, inputting the second intermediate graph feature to the linear layer to obtain second output information.
Step 706, obtaining the first output information, the second output information and the weighted value of each graph feature sample according to the first output information, the second output information, the first weight corresponding to the first output information, the second weight corresponding to the second output information and the third weight corresponding to each graph feature sample.
Step 707, based on the weighted values, fusing the graph feature samples.
Step 708, connecting the feature samples of each fusion map by using the multi-head self-attention fusion model to obtain a map feature vector.
Step 709, obtaining the predicted fault class based on the graph feature vector by using the full connection layer.
And step 710, determining a loss value according to the error between the predicted fault class and the real fault class corresponding to the fault information sample.
Step 711, under the condition that the loss value is greater than the preset threshold, adjusting parameters of the initial multi-scale multi-head self-attention GCN model according to the loss value to obtain an intermediate GCN model, training the initial multi-scale multi-head self-attention GCN model until the loss value is less than the preset threshold, and taking the intermediate GCN model corresponding to the loss value less than the preset threshold as the target multi-scale multi-head self-attention power seeking convolutional neural network GCN model.
And step 712, extracting the characteristics of the fault information of the rotary equipment to be tested to obtain the fault characteristics of the fault information.
In step 713, the fault characteristics are input to the target multi-scale multi-head self-care force diagram convolutional neural network GCN model to obtain the fault class of the rotating device to be tested.
The fault classification method of the rotating equipment carries out fusion on the target output information and a plurality of graph characteristic samples to obtain a fusion graph characteristic sample, trains an initial multi-scale multi-head self-attention GCN model to obtain a target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model based on the fusion graph characteristic sample, and because the fusion graph characteristic sample comprises the target output information and the plurality of graph characteristic samples, the information contained in the method is richer than that contained in a single graph feature sample in the traditional technology, so that the prediction accuracy of the target multi-scale multi-head self-care force-seeking convolutional neural network GCN model obtained based on the training of the fusion graph feature sample with richer information is higher, and further, the accuracy of fault detection can be improved.
In one exemplary embodiment, a schematic diagram of a target multi-scale multi-headed self-care seeking-convolved neural network GCN model is provided, as shown in FIG. 8. And taking the fault information sample, the adjacent matrix 1 and the adjacent matrix 2 as inputs, and obtaining the predicted fault category through processing each module in the target multiscale multi-head self-care force diagram convolutional neural network GCN model.
In one exemplary embodiment, a schematic diagram of a multi-headed self-attention fusion model is provided, as shown in FIG. 9. And taking the graph characteristic sample 1, the graph characteristic sample 2, the first output information and the second output information as inputs, and obtaining the graph characteristic vector through processing of a linear layer, a Tanh function and a Softmax function.
It should be understood that, although the steps in the flowcharts related to the above embodiments are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.
Based on the same inventive concept, the embodiment of the application also provides a fault classification device of the rotary equipment for realizing the fault classification method of the rotary equipment. The implementation of the solution provided by the apparatus is similar to the implementation described in the above method, so the specific limitation in the embodiments of the fault classification apparatus for one or more rotating devices provided below may refer to the limitation of the fault classification method for a rotating device hereinabove, and will not be repeated herein.
In one embodiment, as shown in fig. 10, there is provided a fault classification apparatus 1000 of a rotary device, including: a feature extraction module 1020, a fault classification module 1040, wherein:
The feature extraction module 1020 is configured to perform feature extraction on fault information of the rotating device to be tested to obtain fault features of the fault information;
The fault classification module 1040 is configured to input fault features to a target multi-scale multi-head self-care force diagram convolutional neural network GCN model to obtain a fault class of the rotating device to be tested; the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model is obtained by training an initial graph convolution neural network GCN model based on a fusion graph feature sample, wherein the fusion graph feature sample is obtained by fusing target output information and a plurality of graph feature samples, the target output information is determined by the initial multi-scale multi-head self-attention GCN model based on a fault information sample, and the graph feature sample is determined by the fault information sample, an adjacent matrix and the initial multi-scale multi-head self-attention GCN model.
In one embodiment, the fault classification apparatus 1000 of the rotating device further includes:
The first training module is used for inputting the fault information sample and the adjacent matrix corresponding to each first sub-model into the first sub-model to obtain a graph characteristic sample output by the first sub-model; inputting the fault information sample into a multi-layer sensor to obtain first output information; the target output information includes first output information; obtaining a fusion graph characteristic sample according to the first output information and each graph characteristic sample; training the initial multi-scale multi-head self-attention GCN model according to the fusion map feature sample to obtain a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model.
In one embodiment, the fault classification apparatus 1000 of the rotating device further includes:
The second training module is used for inputting the fault information sample and the adjacent matrix corresponding to each first sub-model into the first graph convolution layer of the first sub-model to obtain first intermediate graph characteristics; inputting first intermediate graph features obtained by the first graph convolution layers of the first sub-models into the connecting layer to obtain second intermediate graph features; inputting the second intermediate graph characteristics into a linear layer to obtain second output information; obtaining a fusion graph characteristic sample according to the first output information, the second output information and each graph characteristic sample; the target output information includes first output information and second output information.
In one embodiment, the fault classification apparatus 1000 of the rotating device further includes:
The third training module is used for obtaining the first output information, the second output information and the weighted value of each graph characteristic sample according to the first output information, the second output information, the characteristic sample of each graph, the first weight corresponding to the first output information, the second weight corresponding to the second output information and the third weight corresponding to the characteristic sample of each graph; and obtaining a fusion map feature sample based on the weighted value.
In one embodiment, the fault classification apparatus 1000 of the rotating device further includes:
The fourth training module is used for connecting the feature samples of each fusion graph by utilizing the multi-head self-attention fusion model to obtain graph feature vectors; obtaining a predicted fault category based on the graph feature vector by using the full connection layer; and training the initial multi-scale multi-head self-attention GCN model according to the error between the predicted fault category and the real fault category corresponding to the fault information sample to obtain a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model.
In one embodiment, the fault classification apparatus 1000 of the rotating device further includes:
The fifth training module is used for determining a loss value according to errors between the predicted fault category and the real fault category corresponding to the fault information sample; under the condition that the loss value is larger than a preset threshold value, adjusting parameters of the initial multi-scale multi-head self-attention GCN model according to the loss value to obtain an intermediate GCN model, training the initial multi-scale multi-head self-attention GCN model until the loss value is smaller than the preset threshold value, and taking the intermediate GCN model corresponding to the loss value smaller than the preset threshold value as a target multi-scale multi-head self-attention power seeking convolutional neural network GCN model.
The respective modules in the fault classification device of the rotating apparatus described above may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a server, and the internal structure of which may be as shown in fig. 11. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured 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, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is for storing data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program, when executed by a processor, implements a method of fault classification for a rotating device.
In one embodiment, a computer device is provided, which may be a terminal, and the internal structure thereof may be as shown in fig. 12. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured 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 computer programs in the non-volatile storage media. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless mode can be realized through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. The computer program, when executed by a processor, implements a method of fault classification for a rotating device. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, can also be keys, a track ball or a touch pad arranged on the shell of the computer equipment, and can also be an external keyboard, a touch pad or a mouse and the like.
It will be appreciated by those skilled in the art that the structures shown in fig. 11 and 12 are merely block diagrams of portions of structures associated with the inventive arrangements and are not limiting of the computer device to which the inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided comprising a memory and a processor, the memory having stored therein a computer program, the processor when executing the computer program performing the steps of:
extracting characteristics of fault information of the rotary equipment to be detected to obtain fault characteristics of the fault information;
Inputting fault characteristics into a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model to obtain fault types of the rotating equipment to be tested; the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model is obtained by training an initial multi-scale multi-head self-attention GCN model based on a fusion graph feature sample, wherein the fusion graph feature sample is obtained by fusing target output information and a plurality of graph feature samples, the target output information is determined by the initial multi-scale multi-head self-attention GCN model based on a fault information sample, and the graph feature sample is determined by the fault information sample, an adjacent matrix and the initial multi-scale multi-head self-attention GCN model.
In an embodiment, there is also provided a computer device comprising a memory and a processor, the memory having stored therein a computer program, the processor implementing the steps of the method embodiments described above when the computer program is executed.
In one embodiment, a computer readable storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of:
extracting characteristics of fault information of the rotary equipment to be detected to obtain fault characteristics of the fault information;
Inputting fault characteristics into a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model to obtain fault types of the rotating equipment to be tested; the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model is obtained by training an initial multi-scale multi-head self-attention GCN model based on a fusion graph feature sample, wherein the fusion graph feature sample is obtained by fusing target output information and a plurality of graph feature samples, the target output information is determined by the initial multi-scale multi-head self-attention GCN model based on a fault information sample, and the graph feature sample is determined by the fault information sample, an adjacent matrix and the initial multi-scale multi-head self-attention GCN model.
In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when executed by a processor, carries out the steps of the method embodiments described above.
In one embodiment, a computer program product is provided comprising a computer program which, when executed by a processor, performs the steps of:
extracting characteristics of fault information of the rotary equipment to be detected to obtain fault characteristics of the fault information;
Inputting fault characteristics into a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model to obtain fault types of the rotating equipment to be tested; the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model is obtained by training an initial multi-scale multi-head self-attention GCN model based on a fusion graph feature sample, wherein the fusion graph feature sample is obtained by fusing target output information and a plurality of graph feature samples, the target output information is determined by the initial multi-scale multi-head self-attention GCN model based on a fault information sample, and the graph feature sample is determined by the fault information sample, an adjacent matrix and the initial multi-scale multi-head self-attention GCN model.
In an embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the steps of the method embodiments described above.
Those skilled in the art will appreciate that implementing all or part of the above-described methods in accordance with the embodiments may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magneto-resistive random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (PHASE CHANGE Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in various forms such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc. The databases referred to in the embodiments provided herein may include at least one of a relational database and a non-relational database. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processor referred to in the embodiments provided in the present application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, or the like, but is not limited thereto.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The foregoing examples illustrate only a few embodiments of the application, which are described in detail and are not to be construed as limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of the application should be assessed as that of the appended claims.

Claims (10)

1. A method of fault classification for a rotating device, the method comprising:
extracting characteristics of fault information of the rotary equipment to be detected to obtain fault characteristics of the fault information;
Inputting fault characteristics into a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model to obtain fault types of the rotating equipment to be tested; the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model is obtained by training an initial multi-scale multi-head self-attention GCN model based on a fusion graph feature sample, wherein the fusion graph feature sample is obtained by fusing target output information and a plurality of graph feature samples, the target output information is determined by the initial multi-scale multi-head self-attention GCN model based on a fault information sample, and the graph feature sample is determined by the fault information sample, an adjacent matrix and the initial multi-scale multi-head self-attention GCN model.
2. The method of claim 1, wherein the initial multi-scale multi-headed self-attention GCN model comprises a plurality of first sub-models and a multi-layer perceptron; the method further comprises the steps of:
Inputting fault information samples and adjacent matrixes corresponding to the first sub-models into the first sub-models aiming at the first sub-models to obtain graph feature samples output by the first sub-models;
inputting the fault information sample into a multi-layer sensor to obtain first output information; the target output information includes first output information;
Obtaining a fusion graph characteristic sample according to the first output information and each graph characteristic sample;
training the initial multi-scale multi-head self-attention GCN model according to the fusion map feature sample to obtain a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model.
3. The method of claim 2, wherein the first sub-model comprises a first graph convolution layer and a second graph convolution layer, and the initial multi-scale multi-headed self-attention GCN model comprises a connection layer and a linear layer; obtaining a fusion graph feature sample according to the first output information and each graph feature sample, including:
Aiming at each first sub-model, inputting a fault information sample and an adjacent matrix corresponding to each first sub-model into a first graph convolution layer of the first sub-model to obtain a first intermediate graph characteristic;
Inputting first intermediate graph features obtained by the first graph convolution layers of the first sub-models into the connecting layer to obtain second intermediate graph features;
inputting the second intermediate graph characteristics into a linear layer to obtain second output information;
obtaining a fusion graph characteristic sample according to the first output information, the second output information and each graph characteristic sample; the target output information includes first output information and second output information.
4. A method according to claim 3, wherein a fused graph feature sample is obtained from the first output information, the second output information and each graph feature sample; the target output information includes first output information and second output information, including:
Obtaining the first output information, the second output information and the weighted value of each graph feature sample according to the first output information, the second output information, the first weight corresponding to the first output information, the second weight corresponding to the second output information and the third weight corresponding to each graph feature sample;
and obtaining a fusion map feature sample based on the weighted value.
5. The method according to any one of claims 2-4, wherein the number of fusion map feature samples is a plurality; the initial multi-scale multi-head self-attention GCN model comprises a multi-head self-attention fusion model and a full connection layer; training an initial multi-scale multi-head self-attention GCN model according to the fusion map feature sample to obtain a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model, wherein the method comprises the following steps:
Connecting the fusion map feature samples by utilizing a multi-head self-attention fusion model to obtain map feature vectors;
obtaining a predicted fault category based on the graph feature vector by using the full connection layer;
And training the initial multi-scale multi-head self-attention GCN model according to the error between the predicted fault category and the real fault category corresponding to the fault information sample to obtain a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model.
6. The method of claim 5, wherein training the initial multi-scale multi-headed self-attention GCN model to obtain the target multi-scale multi-headed self-attention seeking convolutional neural network GCN model based on errors between the predicted fault category and the true fault category corresponding to the fault information samples, comprises:
Determining a loss value according to an error between the predicted fault category and a real fault category corresponding to the fault information sample;
Under the condition that the loss value is larger than a preset threshold value, adjusting parameters of the initial multi-scale multi-head self-attention GCN model according to the loss value to obtain an intermediate GCN model, training the initial multi-scale multi-head self-attention GCN model until the loss value is smaller than the preset threshold value, and taking the intermediate GCN model corresponding to the loss value smaller than the preset threshold value as a target multi-scale multi-head self-attention power seeking convolutional neural network GCN model.
7. A fault classification device for a rotating apparatus, the device comprising:
the feature extraction module is used for extracting features of fault information of the rotary equipment to be detected to obtain fault features of the fault information;
The fault classification module is used for inputting fault characteristics into a target multi-scale multi-head self-attention force diagram convolutional neural network GCN model to obtain fault types of the rotating equipment to be detected; the target multi-scale multi-head self-attention force-seeking convolutional neural network GCN model is obtained by training an initial multi-scale multi-head self-attention GCN model based on a fusion graph feature sample, wherein the fusion graph feature sample is obtained by fusing target output information and a plurality of graph feature samples, the target output information is determined by the initial multi-scale multi-head self-attention GCN model based on a fault information sample, and the graph feature sample is determined by the fault information sample, an adjacent matrix and the initial multi-scale multi-head self-attention GCN model.
8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any one of claims 1 to 6 when the computer program is executed.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 6.
10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, realizes the steps of the method of any one of claims 1 to 6.
CN202410173763.XA 2024-02-07 2024-02-07 Method, device, equipment, storage medium and product for fault classification of rotary equipment Pending CN118035889A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410173763.XA CN118035889A (en) 2024-02-07 2024-02-07 Method, device, equipment, storage medium and product for fault classification of rotary equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410173763.XA CN118035889A (en) 2024-02-07 2024-02-07 Method, device, equipment, storage medium and product for fault classification of rotary equipment

Publications (1)

Publication Number Publication Date
CN118035889A true CN118035889A (en) 2024-05-14

Family

ID=90983614

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410173763.XA Pending CN118035889A (en) 2024-02-07 2024-02-07 Method, device, equipment, storage medium and product for fault classification of rotary equipment

Country Status (1)

Country Link
CN (1) CN118035889A (en)

Similar Documents

Publication Publication Date Title
US11681913B2 (en) Method and system with neural network model updating
CN116010226A (en) Software system reliability simulation evaluation method and device and computer equipment
CN114549849A (en) Image recognition method and device, computer equipment and storage medium
Yan et al. Rotating machinery fault diagnosis based on a novel lightweight convolutional neural network
CN116894802B (en) Image enhancement method, device, computer equipment and storage medium
CN110830939B (en) Positioning method based on improved CPN-WLAN fingerprint positioning database
CN116932935A (en) Address matching method, device, equipment, medium and program product
CN118035889A (en) Method, device, equipment, storage medium and product for fault classification of rotary equipment
CN115063658A (en) Object detection method, device, apparatus, storage medium and program product
CN117521770A (en) Training method and device for node detection model and computer equipment
CN114399459A (en) Two-branch neural network defect detection method, device, equipment, medium and product
CN113159100B (en) Circuit fault diagnosis method, circuit fault diagnosis device, electronic equipment and storage medium
CN110796167A (en) Image classification method based on deep neural network of lifting scheme
CN115965856B (en) Image detection model construction method, device, computer equipment and storage medium
CN117976244B (en) Medicine interaction prediction method and device based on multidimensional characteristics
CN118295664A (en) Code generation method, code generation device, computer equipment, storage medium and product
JP2024520249A (en) Method, device, quantum chip, and electronic device for generating quantum state preparation circuit
CN118366083A (en) Video behavior detection method, device, computer equipment, storage medium and product
CN118736545A (en) Power equipment fault diagnosis method, device, computer equipment and storage medium
Shen et al. Towards Nonparametric Topological Layers in Neural Networks
CN117216649A (en) Sensitivity determination method and device for unbalance degree of cable and computer equipment
CN118730282A (en) Motor production line defect detection and diagnosis method, device and storage medium
CN114648678A (en) Countermeasure sample detection method, apparatus, computer device and storage medium
CN117435910A (en) Abnormal data detection method and device and computer equipment
CN118209914A (en) Probe calibration method, probe calibration device, computer equipment and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination