CN112396109A - Motor bearing fault diagnosis method based on recursion graph and multilayer convolution neural network - Google Patents
Motor bearing fault diagnosis method based on recursion graph and multilayer convolution neural network Download PDFInfo
- Publication number
- CN112396109A CN112396109A CN202011298826.2A CN202011298826A CN112396109A CN 112396109 A CN112396109 A CN 112396109A CN 202011298826 A CN202011298826 A CN 202011298826A CN 112396109 A CN112396109 A CN 112396109A
- Authority
- CN
- China
- Prior art keywords
- neural network
- convolutional neural
- dimensional
- motor bearing
- recursion
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06N3/045—Combinations of 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)
- Bioinformatics & Computational Biology (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses a motor bearing fault diagnosis method based on a recursion graph and a multilayer convolution neural network, which comprises the following steps: collecting one-dimensional vibration signals of the motor bearing under different working conditions; converting the one-dimensional vibration signal of the motor bearing into a two-dimensional recursion map by adopting a recursion map algorithm; all the recursion graphs are subjected to shearing and abbreviating pretreatment; fault classification labels are marked on all recursive icons, and a training set and a test set are divided; constructing a multilayer convolutional neural network model; inputting a training set to the model, performing iterative learning training, and optimizing parameters of the multilayer convolutional neural network until the multilayer convolutional neural network model converges; deploying the trained multilayer convolutional neural network model, inputting a test set, determining fault classification, evaluating diagnosis accuracy, adjusting the model structure, and optimizing parameters until the parameters are optimal; saving the optimal model; and acquiring one-dimensional vibration signals of the motor bearing under a new working condition, converting the signals into a two-dimensional recursion graph, preprocessing the two-dimensional recursion graph, inputting the two-dimensional recursion graph into an optimal model, and performing fault classification and identification on the operation of the motor bearing.
Description
Technical Field
The invention relates to the field of motor bearing fault diagnosis, in particular to a motor bearing fault diagnosis method based on a recursion diagram and a multilayer convolutional neural network.
Background
The bearing is a key part in the motor, and if the motor bearing breaks down, equipment damage, shutdown and production halt can be caused, and even life safety can be threatened. Due to the complex internal structure of the motor, the fault characteristics and the fault types of the bearing are in a nonlinear relation when the bearing fails. Most of the existing methods for motor fault diagnosis are used for collecting bearing vibration signals, manually extracting characteristics, needing to have rich signal processing experience and a large number of prior rules as supports, and having low diagnosis efficiency, so that a new method is required to be used for characteristic extraction and fault diagnosis.
Vibration signals of most equipment are non-stationary and non-linear, more and more non-linear information processing technologies are used in the field of fault diagnosis, such as power spectrogram analysis, multivariate time series analysis, fractal dimension analysis and the like, wherein a recursive graph analysis algorithm can extract dynamic information in system operation from a time series of the signals, can directly describe mainstream information on key signals, and is suitable for vibration signal characteristic expression.
In image identification processing, a scheme for carrying out fault diagnosis on a motor bearing by an artificial intelligence method based on a multilayer convolutional neural network, which is rapidly developed in recent years, is introduced. The convolutional neural network is a typical supervised type feedforward neural network, and the training goal of the convolutional neural network is to realize the learning of abstract features through alternating and overlapping convolution kernels and pooling operations. The multilayer convolutional neural network has a very good classification effect in automatic image feature extraction and image classification and identification.
Disclosure of Invention
The invention aims to introduce a recursive graph analysis technology and a multilayer convolutional neural network image processing technology into the field of motor bearing fault diagnosis and provides a motor bearing fault diagnosis method based on a recursive graph and a multilayer convolutional neural network for improving the fault diagnosis accuracy. The invention utilizes the recursive graph technology in the nonlinear theory and combines the image processing technology and the multilayer convolutional neural network in the modern artificial intelligence algorithm to diagnose and identify the motor bearing faults, thereby improving the fault diagnosis classification accuracy.
The purpose of the invention is realized by the following technical scheme.
The invention discloses a motor bearing fault diagnosis method based on a recursion graph and a multilayer convolution neural network, which comprises the following processes:
(1) collecting one-dimensional vibration signals of the motor bearing under different working conditions;
(2) converting the one-dimensional vibration signal of the motor bearing into a two-dimensional recursion map by adopting a recursion map algorithm in a nonlinear information processing theory;
(3) through an image processing technology, all recursive graphs are subjected to shearing and thumbnail preprocessing;
(4) fault classification labels are injected to all the preprocessed recursive icons, and the recursive icons are divided into a training set and a test set;
(5) constructing a multilayer convolutional neural network model
Designing an input layer, a convolution layer, a pooling layer, a full-link layer and an output layer of the multilayer convolutional neural network according to the dimension of the fault feature vector;
(6) inputting a training set to the constructed multilayer convolutional neural network model, performing iterative learning training, and optimizing parameters of the multilayer convolutional neural network until the trained multilayer convolutional neural network model converges;
(7) deploying the trained multilayer convolutional neural network model, inputting a test set, determining fault classification, evaluating diagnosis accuracy, adjusting the structure of the multilayer convolutional neural network model, and optimizing parameters until the parameters are optimal;
(8) saving the optimal multilayer convolution neural network model;
(9) and (3) acquiring a one-dimensional vibration signal of the motor bearing under the new working condition, processing according to the steps (2) and (3), inputting the processed signal into the optimal multilayer convolutional neural network model trained in the step (8), and performing fault classification and identification on the operation of the motor bearing.
Step (2) adopts a recursion graph algorithm in a nonlinear information processing theory to convert the one-dimensional vibration signal of the motor bearing into a two-dimensional recursion graph, and the specific process is as follows:
1) the one-dimensional vibration time sequence signal x of the motor bearing in different statesiI-1, 2 … n are reconstructed into a two-dimensional space, the reconstruction algorithm is as follows:
Xi＝{xi,xi+τ,…xi+(m-1)τ},i＝1,2,…,N-(m-1)τ (4)
in the formula: xiThe method comprises the following steps of (1) collecting phase points, wherein the number of the two-dimensional phase space is N, the embedding dimension is m, and the time delay is tau;
2) drawing a recursive graph, determined by a recursive matrix, in which each element Rij(ε) is calculated by the following formula:
Rij(ε)＝Θ(ε-||Xi-Xj||),i,j＝1,2,…,N (5)
in the formula, N is phase point XiNumber of (2), XiAnd XjIs two-dimensional phase space arbitrary two points, | Xi-XjL is the norm of the distance between any two points in the two-dimensional phase space, and L is taken2Norm, ε is the distance threshold, Θ (x) is the Heaviside function:
mixing XiAnd XjSubstituting into formula (3) to obtain 0-1 matrix corresponding to NXN distance matrix, and calculating from RijAnd taking the value of 0 or 1 to draw a two-dimensional recursion graph.
And (3) performing shearing and thumbnail preprocessing on all the recursive graphs through an image processing technology, wherein the specific process is as follows:
1) converting the generated two-dimensional recursive graph into a two-dimensional gray recursive graph;
2) cutting a two-dimensional gray level recursive graph image;
3) and abbreviating the cut two-dimensional gray-scale recursion image.
Step (4), marking fault classification labels, dividing the labels into a training set and a test set, and using the labels as the input of the multilayer convolutional neural network, wherein the method comprises the following steps: the training set is 60% of the total samples, and the testing set is 40% of the total samples.
Designing an input layer, a convolution layer, a pooling layer, a full-connection layer and an output layer of the multilayer convolutional neural network according to the dimension of the fault feature vector:
1) an input layer for unifying the inputted images;
2) c1 convolution layer, inputting 32 x 32 image, convolution kernel 5 x 5, convolution kernel type 20, step length 1;
3) s2 pooling layer, inputting 28 x 28 images, wherein the pooling window is 2 x 2, the pooling mode is maximum pooling, and the step length is 2;
4) c3 convolution layer, inputting 14 × 14 images, 5 × 5 convolution kernels, 40 convolution kernel types and 1 step length;
5) s4 pooling layer, inputting 10 x 10 images, pooling window is 2 x 2, pooling mode is maximum pooling, step length is 2;
6) f5 full connection layer, adopting softmax algorithm to predict result;
7) and the output layer outputs the prediction result.
Evaluating the diagnosis accuracy in the step (7), adjusting the structure of the multilayer convolutional neural network model, and optimizing parameters: and (5) changing the structure and parameters of the multilayer convolutional neural network in the step (5) by evaluating the diagnosis accuracy, retraining and continuously improving the model diagnosis accuracy.
Compared with the prior art, the technical scheme of the invention has the following beneficial effects:
(1) the invention provides a new solution idea by introducing a recursive graph algorithm in a nonlinear information processing theory into motor bearing fault diagnosis.
(2) The invention provides a method for processing the recursive graph by graying, cutting, abbreviating and the like in the image processing technology, which is beneficial to improving the subsequent image processing and identifying speed.
(3) The invention provides a method for carrying out feature extraction and fault identification on a recursion graph for converting vibration signals of a motor bearing by using a multilayer convolution neural network in the field of artificial intelligence, and the diagnosis accuracy is improved.
Drawings
FIG. 1 is a motor bearing data acquisition laboratory bench;
FIG. 2 is a flow chart of the method of the present invention;
FIG. 3 is a recursive diagram and a subsequent image processing process for converting vibration signals under normal and various fault states when a motor bearing operates.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
The motor bearing fault diagnosis method based on the recursion diagram and the multilayer convolution neural network, as shown in fig. 2, comprises the following processes:
(1) and collecting one-dimensional vibration signals of the motor bearing under different working conditions.
(2) And converting the one-dimensional vibration signal of the motor bearing into a two-dimensional recursion map by adopting a recursion map algorithm in a nonlinear information processing theory. The specific process comprises the following steps:
1) the one-dimensional vibration time sequence signal x of the motor bearing in different statesiI-1, 2 … n are reconstructed into a two-dimensional space, the reconstruction algorithm is as follows:
Xi＝{xi,xi+τ,…xi+(m-1)τ},i＝1,2,…,N-(m-1)τ (7)
in the formula: xiThe method is characterized in that the method is a phase point set, the number of points in a two-dimensional phase space is N, the embedding dimension is m, the time delay is tau, and the method is selected by a mutual information method and is uniformly set.
2) Drawing a recursive graph, determined by a recursive matrix, in which each element Rij(ε) is calculated by the following formula:
Rij(ε)＝Θ(ε-||Xi-Xj||),i,j＝1,2,…,N (8)
in the formula, N is phase point XiNumber of (2), XiAnd XjIs two-dimensional phase space arbitrary two points, | Xi-XjL is the norm of the distance between any two points in the two-dimensional phase space, and L is taken2Norm, ε is the distance threshold, Θ (x) is the Heaviside function:
mixing XiAnd XjSubstituting into formula (3) to obtain 0-1 matrix corresponding to NXN distance matrix, and calculating from RijAnd taking the value of 0 or 1 to draw a two-dimensional recursion graph.
(3) All the recursive graphs are preprocessed by cutting, abbreviating and the like through an image processing technology.
The specific process comprises the following steps:
1) converting the generated two-dimensional recursive graph into a two-dimensional gray recursive graph;
2) cutting the two-dimensional gray level recursive graph image and keeping main information;
3) and the two-dimensional gray-scale recursive image after the cutting is abbreviated, so that the image processing speed is improved.
(4) And (3) injecting fault classification labels to all the preprocessed recursive icons, dividing the recursive icons into a training set and a test set, and using the training set and the test set as the input of the multilayer convolutional neural network, wherein the fault classification labels comprise: the training set is 60% of the total samples, and the testing set is 40% of the total samples.
(5) Constructing a multilayer convolutional neural network model
And designing an input layer, a convolution layer, a pooling layer, a full-connection layer and an output layer of the multilayer convolutional neural network according to the dimension of the fault feature vector.
1) An input layer for unifying the inputted images;
2) c1 convolution layer, inputting 32 x 32 image, convolution kernel 5 x 5, convolution kernel type 20, step length 1;
3) s2 pooling layer, inputting 28 x 28 images, wherein the pooling window is 2 x 2, the pooling mode is maximum pooling, and the step length is 2;
4) c3 convolution layer, inputting 14 × 14 images, 5 × 5 convolution kernels, 40 convolution kernel types and 1 step length;
5) s4 pooling layer, inputting 10 x 10 images, pooling window is 2 x 2, pooling mode is maximum pooling, step length is 2;
6) f5 full connection layer, adopting softmax algorithm to predict result;
7) and the output layer outputs the prediction result.
(6) Inputting a training set to the constructed multilayer convolutional neural network model, performing iterative learning training, and optimizing parameters of the multilayer convolutional neural network until the trained multilayer convolutional neural network model converges.
(7) Deploying the trained multilayer convolutional neural network model, inputting the test set into the trained multilayer convolutional neural network model for feature extraction, determining fault classification, changing the structure and parameters of the multilayer convolutional neural network in the step (5) by evaluating the diagnosis accuracy, retraining, continuously improving the model diagnosis accuracy, and optimizing the parameters until the parameters are optimal.
(8) And saving the optimal multilayer convolutional neural network model.
(9) And (3) acquiring a one-dimensional vibration signal of the motor bearing under the new working condition, processing according to the steps (2) and (3), inputting the processed signal into the optimal multilayer convolutional neural network model trained in the step (8), and performing fault classification and identification on the operation of the motor bearing.
The specific embodiment is as follows:
the implementation case comprises 5 bearing operating states of normal, inner and outer ring faults, rolling element faults and the like, and is verified by using a Kaiser university motor bearing experiment table data set, wherein the experiment table is shown in figure 1. The rotating speed of the motor is 1750rpm, the sampling frequency is 12kHz, and the vibration signal data of the bearing at the driving end are selected, wherein the data composition is shown in the table 1.
TABLE 1 Experimental data
As shown in fig. 2, the implementation steps of the present invention are as follows:
(1) the vibration signals of 5 motor bearing running states are collected, and the waveform of the collected one-dimensional signals is shown in figure 3.
(2) And converting the one-dimensional vibration signal of the motor bearing into a two-dimensional recursion map by adopting a recursion map algorithm in a nonlinear information processing theory.
(a) The one-dimensional vibration time sequence signal x of the motor bearing in different statesiI-1, 2 … n are reconstructed into a two-dimensional space, and the reconstruction algorithm is as formula (1).
(b) And drawing a recursive graph, wherein each element is determined by the formulas (2) and (3) and is determined by a recursive matrix.
(3) Image processing, mainly comprising:
(a) converting the generated two-dimensional recursive graph into a two-dimensional gray recursive graph;
(b) clipping the two-dimensional grayscale recursion map into an image of 320 × 320 pixels;
(c) the clipped 320 × 320 grayscale recursion map is reduced to 32 × 32.
The process is shown in figure 3.
(4) And (3) injecting fault classification labels to all the preprocessed recursive icons, dividing the recursive icons into a training set and a test set, and using the training set and the test set as the input of the multilayer convolutional neural network, wherein the fault classification labels comprise: the training set is 60% of the total samples, and the testing set is 40% of the total samples.
(5) Constructing and training a multi-layer convolutional neural model
(a) An input layer for unifying the inputted images;
(b) c1 convolution layer, 32 × 32 image is input, convolution kernel is 5 × 5, convolution kernel type is 20, step size is 1.
(c) And S2, pooling the layer, inputting 28 × 28 images, wherein the pooling window is 2 × 2, the pooling mode is maximum pooling, and the step size is 2.
(d) The C3 convolution layer receives a 14 × 14 image, and has 5 × 5 convolution kernels, 40 convolution kernel types, and a step size of 1.
(e) And S4 pooling layers, inputting 10 x 10 images, wherein the pooling window is 2 x 2, the pooling mode is maximum pooling, and the step size is 2.
(f) F5 full link layer, and adopting softmax algorithm to predict the result.
(g) And the output layer outputs the prediction result.
(6) Inputting a training set to the constructed multilayer convolutional neural network model, performing iterative learning training, and optimizing parameters of the multilayer convolutional neural network until the trained multilayer convolutional neural network model converges.
(7) Deploying the trained multilayer convolutional neural network model, inputting the test set into the trained multilayer convolutional neural network model for feature extraction, determining fault classification, changing the structure and parameters of the multilayer convolutional neural network in the step (5) by evaluating the diagnosis accuracy, retraining, continuously improving the model diagnosis accuracy, and optimizing the parameters until the parameters are optimal.
(8) And saving the optimal multilayer convolutional neural network model.
(9) And (3) acquiring a one-dimensional vibration signal of the motor bearing under the new working condition, processing according to the steps (2) and (3), inputting the processed signal into the optimal multilayer convolutional neural network model trained in the step (8), and performing fault classification and identification on the operation of the motor bearing.
Table 2 is an average value of the diagnosis and identification accuracy of the present invention in which 5 motor shaft operating states are subjected to multiple experiments, so that it can be seen that the present invention has a high diagnosis accuracy of the motor bearing failure.
TABLE 2
While the present invention has been described in terms of its functions and operations with reference to the accompanying drawings, it is to be understood that the invention is not limited to the precise functions and operations described above, and that the above-described embodiments are illustrative rather than restrictive, and that various changes and modifications may be effected therein by one skilled in the art without departing from the scope or spirit of the invention as defined by the appended claims.
Claims (6)
1. A motor bearing fault diagnosis method based on a recursion graph and a multilayer convolution neural network is characterized by comprising the following processes:
(1) collecting one-dimensional vibration signals of the motor bearing under different working conditions;
(2) converting the one-dimensional vibration signal of the motor bearing into a two-dimensional recursion map by adopting a recursion map algorithm in a nonlinear information processing theory;
(3) through an image processing technology, all recursive graphs are subjected to shearing and thumbnail preprocessing;
(4) fault classification labels are injected to all the preprocessed recursive icons, and the recursive icons are divided into a training set and a test set;
(5) constructing a multilayer convolutional neural network model
Designing an input layer, a convolution layer, a pooling layer, a full-link layer and an output layer of the multilayer convolutional neural network according to the dimension of the fault feature vector;
(6) inputting a training set to the constructed multilayer convolutional neural network model, performing iterative learning training, and optimizing parameters of the multilayer convolutional neural network until the trained multilayer convolutional neural network model converges;
(7) deploying the trained multilayer convolutional neural network model, inputting a test set, determining fault classification, evaluating diagnosis accuracy, adjusting the structure of the multilayer convolutional neural network model, and optimizing parameters until the parameters are optimal;
(8) saving the optimal multilayer convolution neural network model;
(9) and (3) acquiring a one-dimensional vibration signal of the motor bearing under the new working condition, processing according to the steps (2) and (3), inputting the processed signal into the optimal multilayer convolutional neural network model trained in the step (8), and performing fault classification and identification on the operation of the motor bearing.
2. The motor bearing fault diagnosis method based on the recursion map and the multilayer convolutional neural network as claimed in claim 1, wherein the step (2) adopts the recursion map algorithm in the nonlinear information processing theory to convert the one-dimensional vibration signal of the motor bearing into the two-dimensional recursion map, and the specific process comprises the following steps:
1) the one-dimensional vibration time sequence signal x of the motor bearing in different statesiI-1, 2 … n are reconstructed into a two-dimensional space, the reconstruction algorithm is as follows:
Xi＝{xi,xi+τ,…xi+(m-1)τ},i＝1,2,…,N-(m-1)τ (1)
in the formula: xiThe method comprises the following steps of (1) collecting phase points, wherein the number of the two-dimensional phase space is N, the embedding dimension is m, and the time delay is tau;
2) drawing a recursive graph, determined by a recursive matrix, in which each element Rij(ε) is calculated by the following formula:
Rij(ε)＝Θ(ε-||Xi-Xj||),i,j＝1,2,…,N (2)
in the formula, N is phase point XiNumber of (2), XiAnd XjIs two-dimensional phase space arbitrary two points, | Xi-XjL is the norm of the distance between any two points in the two-dimensional phase space, and L is taken2Norm, ε is the distance threshold, Θ (x) is the Heaviside function:
mixing XiAnd XjSubstituting into formula (3) to obtain 0-1 matrix corresponding to NXN distance matrix, and calculating from RijAnd taking the value of 0 or 1 to draw a two-dimensional recursion graph.
3. The motor bearing fault diagnosis method based on the recursion map and the multilayer convolutional neural network as claimed in claim 1, wherein the step (3) is to perform shearing and abbreviating preprocessing on all the recursion maps through an image processing technology, and the specific process is as follows:
1) converting the generated two-dimensional recursive graph into a two-dimensional gray recursive graph;
2) cutting a two-dimensional gray level recursive graph image;
3) and abbreviating the cut two-dimensional gray-scale recursion image.
4. The motor bearing fault diagnosis method based on the recursion map and the multilayer convolutional neural network as claimed in claim 1, wherein the step (4) of labeling fault classification labels, dividing the labels into a training set and a test set, and using the training set and the test set as inputs of the multilayer convolutional neural network comprises the following steps: the training set is 60% of the total samples, and the testing set is 40% of the total samples.
5. The motor bearing fault diagnosis method based on the recursion map and the multilayer convolutional neural network as claimed in claim 1, wherein the step (5) designs the input layer, the convolutional layer, the pooling layer, the full-link layer and the output layer of the multilayer convolutional neural network according to the fault feature vector dimension:
1) an input layer for unifying the inputted images;
2) c1 convolution layer, inputting 32 x 32 image, convolution kernel 5 x 5, convolution kernel type 20, step length 1;
3) s2 pooling layer, inputting 28 x 28 images, wherein the pooling window is 2 x 2, the pooling mode is maximum pooling, and the step length is 2;
4) c3 convolution layer, inputting 14 × 14 images, 5 × 5 convolution kernels, 40 convolution kernel types and 1 step length;
5) s4 pooling layer, inputting 10 x 10 images, pooling window is 2 x 2, pooling mode is maximum pooling, step length is 2;
6) f5 full connection layer, adopting softmax algorithm to predict result;
7) and the output layer outputs the prediction result.
6. The motor bearing fault diagnosis method based on the recursion map and the multilayer convolutional neural network as claimed in claim 1, wherein the diagnosis accuracy is evaluated in step (7), the structure of the multilayer convolutional neural network model is adjusted, and the parameters are optimized: and (5) changing the structure and parameters of the multilayer convolutional neural network in the step (5) by evaluating the diagnosis accuracy, retraining and continuously improving the model diagnosis accuracy.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011298826.2A CN112396109A (en) | 2020-11-19 | 2020-11-19 | Motor bearing fault diagnosis method based on recursion graph and multilayer convolution neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011298826.2A CN112396109A (en) | 2020-11-19 | 2020-11-19 | Motor bearing fault diagnosis method based on recursion graph and multilayer convolution neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112396109A true CN112396109A (en) | 2021-02-23 |
Family
ID=74607468
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011298826.2A Pending CN112396109A (en) | 2020-11-19 | 2020-11-19 | Motor bearing fault diagnosis method based on recursion graph and multilayer convolution neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112396109A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113060614A (en) * | 2021-03-16 | 2021-07-02 | 重庆市特种设备检测研究院 | Fault diagnosis method and device for elevator driving host and readable storage medium |
| CN113155464A (en) * | 2021-03-31 | 2021-07-23 | 燕山大学 | CNN model visual optimization method for bearing fault recognition |
| CN113203566A (en) * | 2021-04-06 | 2021-08-03 | 上海吞山智能科技有限公司 | Motor bearing fault diagnosis method based on one-dimensional data enhancement and CNN |
| CN113435453A (en) * | 2021-06-11 | 2021-09-24 | 西安交通大学 | Reciprocating compressor fault diagnosis method and system based on convolutional neural network |
| CN113432874A (en) * | 2021-05-31 | 2021-09-24 | 西安理工大学 | Rolling bearing fault diagnosis method based on 2DPCA-CWT and CNN |
| CN113469060A (en) * | 2021-07-02 | 2021-10-01 | 浙大城市学院 | Multi-sensor fusion convolution neural network aeroengine bearing fault diagnosis method |
| CN113607325A (en) * | 2021-10-09 | 2021-11-05 | 武汉地震工程研究院有限公司 | Intelligent monitoring method and system for looseness positioning of steel structure bolt group |
| CN113759320A (en) * | 2021-07-07 | 2021-12-07 | 中国人民解放军国防科技大学 | Radar active interference signal identification method based on recursive graph and deep learning |
| CN114037171A (en) * | 2021-11-17 | 2022-02-11 | 国网天津市电力公司 | Power load prediction method based on recursive graph and convolutional neural network |
| CN114036691A (en) * | 2021-11-26 | 2022-02-11 | 江苏科技大学 | Bearing degradation evaluation method and system based on recursive complex network |
| CN114800041A (en) * | 2022-05-06 | 2022-07-29 | 嘉兴南湖学院 | Cutter state monitoring method and monitoring device thereof |
| CN115062867A (en) * | 2022-07-26 | 2022-09-16 | 中车长春轨道客车股份有限公司 | Bearing fault prediction method based on Teager energy spectrum and improved CNN model |
| CN115546224A (en) * | 2022-12-06 | 2022-12-30 | 新乡学院 | Automatic fault identification and control method for motor operation process |
| CN116728291A (en) * | 2023-08-16 | 2023-09-12 | 湖南大学 | Robot polishing system state monitoring method and device based on edge calculation |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102880915A (en) * | 2012-09-06 | 2013-01-16 | 中山大学 | Method of forecasting electric quantity based on association mining of hot events |
| CN106289777A (en) * | 2016-08-01 | 2017-01-04 | 北京航空航天大学 | A kind of multi-state rolling bearing performance appraisal procedure based on geometry tolerance |
| CN108445751A (en) * | 2018-02-28 | 2018-08-24 | 天津大学 | Merge multiple target SSVEP ideas control methods and the application of recurrence plot and deep learning |
| CN110991240A (en) * | 2019-10-31 | 2020-04-10 | 湖南世优电力科技股份有限公司 | Single-phase fault identification and classification method for low-voltage neutral point ungrounded system |
| CN111222255A (en) * | 2020-02-25 | 2020-06-02 | 湖南世优电力科技股份有限公司 | Big data simulation system and method for electrical fault identification |
| CN111709292A (en) * | 2020-05-18 | 2020-09-25 | 杭州电子科技大学 | Compressor vibration fault detection method based on recursive graph and deep convolutional network |
-
2020
- 2020-11-19 CN CN202011298826.2A patent/CN112396109A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102880915A (en) * | 2012-09-06 | 2013-01-16 | 中山大学 | Method of forecasting electric quantity based on association mining of hot events |
| CN106289777A (en) * | 2016-08-01 | 2017-01-04 | 北京航空航天大学 | A kind of multi-state rolling bearing performance appraisal procedure based on geometry tolerance |
| CN108445751A (en) * | 2018-02-28 | 2018-08-24 | 天津大学 | Merge multiple target SSVEP ideas control methods and the application of recurrence plot and deep learning |
| CN110991240A (en) * | 2019-10-31 | 2020-04-10 | 湖南世优电力科技股份有限公司 | Single-phase fault identification and classification method for low-voltage neutral point ungrounded system |
| CN111222255A (en) * | 2020-02-25 | 2020-06-02 | 湖南世优电力科技股份有限公司 | Big data simulation system and method for electrical fault identification |
| CN111709292A (en) * | 2020-05-18 | 2020-09-25 | 杭州电子科技大学 | Compressor vibration fault detection method based on recursive graph and deep convolutional network |
Non-Patent Citations (3)
| Title |
|---|
| 曾全昊: "基于卷积神经网络的变压器有载分接开关故障识别", 《电力***自动化》 * |
| 金宁德 等: "《非线性信息处理技术》", 28 February 2017, 天津大学出版社 * |
| 陈敏: "《认知计算导论》", 31 May 2017, 华中科技大学出版社 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113060614A (en) * | 2021-03-16 | 2021-07-02 | 重庆市特种设备检测研究院 | Fault diagnosis method and device for elevator driving host and readable storage medium |
| CN113155464A (en) * | 2021-03-31 | 2021-07-23 | 燕山大学 | CNN model visual optimization method for bearing fault recognition |
| CN113203566A (en) * | 2021-04-06 | 2021-08-03 | 上海吞山智能科技有限公司 | Motor bearing fault diagnosis method based on one-dimensional data enhancement and CNN |
| CN113432874A (en) * | 2021-05-31 | 2021-09-24 | 西安理工大学 | Rolling bearing fault diagnosis method based on 2DPCA-CWT and CNN |
| CN113435453B (en) * | 2021-06-11 | 2023-04-07 | 西安交通大学 | Reciprocating compressor fault diagnosis method and system based on convolutional neural network |
| CN113435453A (en) * | 2021-06-11 | 2021-09-24 | 西安交通大学 | Reciprocating compressor fault diagnosis method and system based on convolutional neural network |
| CN113469060A (en) * | 2021-07-02 | 2021-10-01 | 浙大城市学院 | Multi-sensor fusion convolution neural network aeroengine bearing fault diagnosis method |
| CN113759320A (en) * | 2021-07-07 | 2021-12-07 | 中国人民解放军国防科技大学 | Radar active interference signal identification method based on recursive graph and deep learning |
| CN113607325A (en) * | 2021-10-09 | 2021-11-05 | 武汉地震工程研究院有限公司 | Intelligent monitoring method and system for looseness positioning of steel structure bolt group |
| CN114037171A (en) * | 2021-11-17 | 2022-02-11 | 国网天津市电力公司 | Power load prediction method based on recursive graph and convolutional neural network |
| CN114036691A (en) * | 2021-11-26 | 2022-02-11 | 江苏科技大学 | Bearing degradation evaluation method and system based on recursive complex network |
| CN114800041A (en) * | 2022-05-06 | 2022-07-29 | 嘉兴南湖学院 | Cutter state monitoring method and monitoring device thereof |
| CN115062867A (en) * | 2022-07-26 | 2022-09-16 | 中车长春轨道客车股份有限公司 | Bearing fault prediction method based on Teager energy spectrum and improved CNN model |
| CN115546224A (en) * | 2022-12-06 | 2022-12-30 | 新乡学院 | Automatic fault identification and control method for motor operation process |
| CN115546224B (en) * | 2022-12-06 | 2023-04-07 | 新乡学院 | Automatic fault identification and control method for motor operation process |
| CN116728291A (en) * | 2023-08-16 | 2023-09-12 | 湖南大学 | Robot polishing system state monitoring method and device based on edge calculation |
| CN116728291B (en) * | 2023-08-16 | 2023-10-31 | 湖南大学 | Robot polishing system state monitoring method and device based on edge calculation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112396109A (en) | Motor bearing fault diagnosis method based on recursion graph and multilayer convolution neural network | |
| Xu et al. | A hybrid deep-learning model for fault diagnosis of rolling bearings | |
| CN110361176B (en) | Intelligent fault diagnosis method based on multitask feature sharing neural network | |
| Pandhare et al. | Convolutional neural network based rolling-element bearing fault diagnosis for naturally occurring and progressing defects using time-frequency domain features | |
| CN110595780B (en) | Bearing fault identification method based on vibration gray level image and convolution neural network | |
| CN109858352B (en) | Fault diagnosis method based on compressed sensing and improved multi-scale network | |
| CN111651937A (en) | Method for diagnosing similar self-adaptive bearing fault under variable working conditions | |
| CN109992872B (en) | Mechanical equipment residual life prediction method based on stacked separation convolution module | |
| CN112926644A (en) | Method and system for predicting residual service life of rolling bearing | |
| CN116451150A (en) | Equipment fault diagnosis method based on semi-supervised small sample | |
| CN115791174B (en) | Rolling bearing abnormality diagnosis method, system, electronic equipment and storage medium | |
| CN114548190A (en) | Wind turbine fault diagnosis method based on self-adaptive residual error neural network | |
| CN114896733B (en) | Rolling bearing fault type identification method based on deep reinforcement learning | |
| CN112305388A (en) | On-line monitoring and diagnosing method for partial discharge fault of generator stator winding insulation | |
| Yeh et al. | Using convolutional neural network for vibration fault diagnosis monitoring in machinery | |
| CN113673442B (en) | Variable working condition fault detection method based on semi-supervised single classification network | |
| Zare et al. | Convolutional neural networks for wind turbine gearbox health monitoring | |
| Huo et al. | A class-level matching unsupervised transfer learning network for rolling bearing fault diagnosis under various working conditions | |
| CN117475191A (en) | Bearing fault diagnosis method for feature alignment domain antagonistic neural network | |
| CN110728377A (en) | Intelligent fault diagnosis method and system for electromechanical equipment | |
| Li et al. | A robust fault diagnosis method for rolling bearings based on deep convolutional neural network | |
| CN113052060B (en) | Bearing residual life prediction method and device based on data enhancement and electronic equipment | |
| Boudiaf et al. | Bearing fault diagnosis in induction motor using continuous wavelet transform and convolutional neural networks | |
| Li et al. | Intelligent CNN-based fault diagnosis of rotating machinery with small fault samples | |
| CN117493980B (en) | Bearing fault diagnosis method integrating feature extraction and sequencing causal discovery |
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 |