CN112488179A - Rotary machine fault diagnosis method based on GRU - Google Patents
Rotary machine fault diagnosis method based on GRU Download PDFInfo
- Publication number
- CN112488179A CN112488179A CN202011355499.XA CN202011355499A CN112488179A CN 112488179 A CN112488179 A CN 112488179A CN 202011355499 A CN202011355499 A CN 202011355499A CN 112488179 A CN112488179 A CN 112488179A
- Authority
- CN
- China
- Prior art keywords
- fault diagnosis
- gru
- fault
- rotary machine
- vibration signal
- 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
- 238000003745 diagnosis Methods 0.000 title claims abstract description 42
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000012549 training Methods 0.000 claims abstract description 27
- 238000012360 testing method Methods 0.000 claims abstract description 11
- 230000008569 process Effects 0.000 claims abstract description 9
- 238000012545 processing Methods 0.000 claims abstract description 6
- 239000011159 matrix material Substances 0.000 claims description 16
- 230000006870 function Effects 0.000 claims description 14
- 210000002569 neuron Anatomy 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 230000003247 decreasing effect Effects 0.000 claims description 2
- 230000009466 transformation Effects 0.000 claims description 2
- 238000000605 extraction Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 239000003365 glass fiber Substances 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
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
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24137—Distances to cluster centroïds
- G06F18/2414—Smoothing the distance, e.g. radial basis function networks [RBFN]
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield 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
- G06N3/04—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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T1/00—General purpose image data processing
- G06T1/20—Processor architectures; Processor configuration, e.g. pipelining
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses a rotary machine fault diagnosis method based on GRU, which comprises the following steps: 1) collecting and acquiring a vibration signal from a rotary mechanical system; 2) converting the acquired original one-dimensional vibration signal into a two-dimensional image, and dividing the two-dimensional image into a training data set and a testing data set; 3) inputting a training data set into a linear layer to increase the dimensionality of each row of data in the image and improve the feature capacity; 4) processing the output of the linear layer by using a gating circulation unit GRU, and learning representative characteristics from the time sequence data; 5) using the learned representative features as an input classifier to obtain a fault diagnosis model for fault state identification; 6) and inputting the test data set into the trained fault diagnosis model for fault identification, and completing the fault diagnosis process. The fault diagnosis method does not need to manually extract features, has strong noise resistance, can reduce the influence of extreme working conditions on fault diagnosis, and improves the fault classification precision of the rotary machine.
Description
Technical Field
The invention relates to a fault diagnosis technology, in particular to a rotary machine fault diagnosis method based on GRU.
Background
With the sophistication of modern industrial system equipment, rotary machine fault diagnosis has become a focus of attention in both academic and industrial areas. Key components such as rolling bearings and gears in rotary machinery are prone to failure under extreme working conditions, system performance is affected, and huge loss is caused. Therefore, it is necessary to develop an effective fault diagnosis method based on a rotary machine.
The fault diagnosis generally comprises three steps of vibration signal acquisition, feature extraction and fault classification, wherein the last two steps are more critical to the performance of the fault diagnosis method. The traditional fault diagnosis method based on data driving relies on manual feature extraction, requires expert knowledge, and directly influences the final classification precision. The fault diagnosis method based on deep learning can complete the processes of feature extraction and fault classification through the deep learning model, and avoids manual feature extraction.
The CNN and the RNN in the deep learning model show excellent performance in fault diagnosis. RNNs are better at processing time series data than CNNs due to their memory. However, RNN has a long-term dependence problem, and to improve this problem, LSTM and GRU have been proposed by introducing a gating mechanism to control the accumulation rate of information. The GRU combines the input gates in the LSTM into an update gate, and the combined result leads to the reduction of parameter quantity and easier convergence; and no unit state exists, so that the neural network structure can be simplified and the learning efficiency can be improved. The GRU is a simpler network than the LSTM, and can learn a representative feature by making full use of time information of time-series data.
Disclosure of Invention
The invention aims to solve the technical problem of providing a rotary machine fault diagnosis method based on GRU (glass-fiber reinforced Unit) aiming at the defects in the prior art.
The technical scheme adopted by the invention for solving the technical problems is as follows: a rotary machine fault diagnosis method based on GRU comprises the following steps:
1) collecting and acquiring a vibration signal from a rotary mechanical system; marking the fault state of the vibration signal;
2) converting the acquired original one-dimensional vibration signal into a two-dimensional image, and dividing the two-dimensional image into a training data set and a testing data set;
3) inputting a training data set into a linear layer to increase the dimensionality of each row of data in the image and improve the feature capacity;
4) processing the output of the linear layer by using a gating circulation unit GRU, and learning representative characteristics from the time sequence data;
5) using the learned representative features as an input classifier to obtain a fault diagnosis model for fault state identification;
6) and inputting the test data set into the trained fault diagnosis model for fault identification, and completing the fault diagnosis process.
According to the scheme, the fault state in the step 1) comprises a normal state and a fault state corresponding to a preset fault type.
According to the scheme, the step 2) converts the acquired original one-dimensional vibration signal into a two-dimensional image, which specifically comprises the following steps:
2.1) randomly extracting the length N from the original vibration signal2The vibration signal samples of (a);
2.2) filling the continuous line segment signals with the length of N in the lines of the constructed image in sequence;
2.3) normalizing the sample data of the vibration signal to be in a range of 0-255;
2.4) completing the conversion from the original vibration signal to N2And (4) converting image pixels.
According to the scheme, in the step 3), the training data set is input into a linear layer to increase the dimensionality of each row of data in the image, the linear layer is a full connection layer for linear transformation, and the increase of the data dimensionality is determined according to experiments.
According to the scheme, in the step 4), the GRU includes two gates: reset gate rtAnd an update gate zt(ii) a When the input sequence is xtThe output of the two gated and hidden states is as follows:
rt＝σ(Wrxt+Urht-1+br)
zt＝σ(Wzxt+Uzht-1+bz)
ht＝zt⊙ht-1+(1-zt)⊙tanh(Whxt+Uh(rt⊙ht-1)+bh)
wherein, Wr、Wz、WhRespectively representing the weight matrix between the reset gate and the input, the weight matrix between the update gate and the input, the weight matrix between the hidden state and the input at the present moment, Ur、Uz、UhRespectively representing a weight matrix between the reset gate and the previous hidden layer, a weight matrix between the update gate and the previous hidden layer, a weight matrix between the hidden state at the current moment and the previous hidden layer, ht-1σ is the Logistic function for the hidden state at the previous time.
According to the scheme, the GRU adopts a one-way structure, residual connection is added into the GRU, and the step length T is set to be N.
According to the scheme, the multilayer perceptron is adopted as the classifier in the step 5), the multilayer perceptron comprises two hidden layers and an output layer, the number of neurons of each hidden layer is set according to the dimension of each line of data in the step 3), the number of neurons of the output layer is set according to the number of the fault state marks in the step 1), and a ReLU function is used as an activation function.
According to the scheme, in the step 5), a batch normalization and discarding method is added to each hidden layer of the multilayer perceptron, and the discarding rate is set to be 0.5.
According to the scheme, in the process of training the classifier in the step 5), a learning rate attenuation strategy is adopted, and the parameters are set as follows: the number of training times was set to 35, the initial empirical learning rate was used in the first 25 training stages, set to 0.0002, and then the learning rate was decreased by an attenuation rate of 0.1; batch size 24; selecting an RMSProp algorithm as an optimizer; selecting a cross entropy loss function as an objective function, which is specifically defined as follows:
where m is the number of samples, θ is the training parameter of the model, x(i)And y(i)For the ith sample and its corresponding label, h (-) represents the proposed classifier model.
The invention has the following beneficial effects:
1. compared with the traditional fault diagnosis method based on data driving, the method does not need to manually extract features and does not depend on expert knowledge.
2. The invention increases the dimensionality of input data through the linear layer, improves the feature capacity and enables the GRU to learn more potential features.
3. According to the invention, the residual error connection and learning rate attenuation strategy is applied to the model, so that the stability and prediction precision of the training process can be effectively improved.
4. The invention has strong noise resistance through the combination of the linear layer, the GRU and the multilayer perceptron, can reduce the influence of extreme working conditions on fault diagnosis and improve the fault classification precision of the rotary machine.
Drawings
The invention will be further described with reference to the accompanying drawings and examples, in which:
FIG. 1 is a schematic flow chart of a method of an embodiment of the present invention;
FIG. 2 is a schematic diagram of a model architecture of an embodiment of the present invention;
FIG. 3 is a schematic diagram of converting an original one-dimensional vibration signal into a two-dimensional image according to an embodiment of the present invention;
FIG. 4 is a confusion matrix of test results in an embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
As shown in fig. 1, a flowchart of a method for diagnosing fault of rotary machine based on GRU is shown in fig. 2;
the proposed GRU-based rotary machine fault diagnosis method is verified below with reference to fig. 1, taking a certain motor bearing fault data set as an example.
Step 1: acquiring a vibration signal from a rotating mechanical system;
under four load conditions of 0hp, 1hp, 2hp and 3hp, the rotating speeds are 1797r/min, 1772r/min, 1750r/min and 1730r/min respectively, and the driving end vibration signal is collected at the sampling frequency of 12 kHz. The selected data set comprises three bearing fault types including rolling element fault (BF), inner ring fault (IF) and outer ring fault (OF), each fault type comprises three fault diameters including 0.18mm, 0.36mm and 0.54mm, and therefore, the data set has ten states including normal state (NO).
Step 2: converting an original one-dimensional vibration signal into a two-dimensional image, and dividing the two-dimensional image into a training data set and a test data set;
a schematic diagram of converting an original one-dimensional vibration signal into a two-dimensional image is shown in fig. 3, and the specific steps are as follows:
step 2.1: randomly extracting vibration signal samples with the length of 4096 from an original signal;
step 2.2: filling the lines of the constructed image with continuous line segment signals with the length of 64 in sequence;
step 2.2: normalizing the sample data of the vibration signal to a range of 0-255, specifically as follows:
let S (i, j) denote the jth value of the ith signal segment, and the image pixel point P (i, j) of the constructed image should be as follows:
step 2.4: the image conversion of 64 x 64 pixels is completed.
960 images were generated for each fault type under four load conditions, so there were 9600 images in the data set. 8000 images are selected as a training data set, and the rest 1600 images are selected as a test data set. The details of the data set are shown in table 1.
Table 1 data set details
And step 3: inputting a training data set into a linear layer;
the dimension of each line of data in the image is increased by inputting a training data set into a linear layer, i.e. a vibration signal segment with a processing length of 64 is processed, and the input dimension is raised from 64 to 1024.
A batch normalization and discard method was added to the linear layer, setting the discard rate to 0.5.
And 4, step 4: processing the output of the linear layer by applying a GRU, and learning advanced features from the time series data;
the GRU adopts a one-way structure, the step length T is set to 64, the size of a hidden layer is set to 1024, and the number of the hidden layer is set to 1.
Residual error connection is added into GRU, representative characteristics are learned, and stability and prediction accuracy in the training process are improved.
And 5: the learned representative features are sent to a classification module to realize fault identification;
and (4) sending the learned representative features into a classifier, identifying and classifying the faults by adopting a multilayer perceptron, and training a fault diagnosis model.
The multi-layer perceptron used comprises two hidden layers, each with 1024 neurons, and an output layer with 10 neurons, and uses the ReLU function as the activation function.
And adding a batch normalization and discarding method to each hidden layer of the multi-layer perceptron, and setting the discarding rate to be 0.5.
In the process of training the fault diagnosis model, a learning rate attenuation strategy is adopted, the training times are set to be 35, the initial experience learning rate is used in the first 25 training stages and is set to be 0.0002, and then the learning rate is reduced by an attenuation rate of 0.1; batch size 24; selecting an RMSProp algorithm as an optimizer; selecting a cross entropy loss function as an objective function, which is specifically defined as follows:
where m is the number of samples, θ is the training parameter of the model, x(i)And y(i)For the ith sample and its corresponding label, h (-) represents the proposed model.
Step 6: and inputting the test data set into a trained fault diagnosis system for fault identification, and completing the fault diagnosis process.
The total of 10 experiments were performed, and the average value was used as the final recognition classification result.
The confusion matrix of the test results is shown in fig. 4, wherein the horizontal axis represents the predicted label and the vertical axis represents the actual label, and the numbers in the confusion matrix represent the percentage probability that a certain actual label is diagnosed as a certain label. As can be seen from the confusion matrix, the identification accuracy OF RF21 is 99.4%, the RF21 is misjudged as OF21 with the probability OF 0.6%, the identification accuracy OF the rest states is 100%, and the fault diagnosis system can accurately identify 10 states OF the bearing.
In conclusion, the effectiveness of the rotary machine fault diagnosis method based on the GRU is verified.
It will be understood that modifications and variations can be made by persons skilled in the art in light of the above teachings and all such modifications and variations are intended to be included within the scope of the invention as defined in the appended claims.
Claims (9)
1. A rotary machine fault diagnosis method based on GRU is characterized by comprising the following steps:
1) collecting and acquiring a vibration signal from a rotary mechanical system; marking the fault state of the vibration signal;
2) converting the acquired original one-dimensional vibration signal into a two-dimensional image, and dividing the two-dimensional image into a training data set and a testing data set;
3) inputting a training data set into a linear layer to increase the dimensionality of each row of data in the image and improve the feature capacity;
4) processing the output of the linear layer by using a gating circulation unit GRU, and learning representative characteristics from the time sequence data;
5) using the learned representative features as an input classifier to obtain a fault diagnosis model for fault state identification;
6) and inputting the test data set into the trained fault diagnosis model for fault identification, and completing the fault diagnosis process.
2. The fault diagnosis method for a rotary machine based on GRU of claim 1, wherein the fault state in step 1) includes a normal state and a fault state corresponding to a preset fault type.
3. The GRU-based rotary machine fault diagnosis method according to claim 1, wherein the step 2) converts the acquired original one-dimensional vibration signal into a two-dimensional image, specifically as follows:
2.1) randomly extracting the length N from the original vibration signal2The vibration signal samples of (a);
2.2) filling the continuous line segment signals with the length of N in the lines of the constructed image in sequence;
2.3) normalizing the sample data of the vibration signal to be in a range of 0-255;
2.4) completing the conversion from the original vibration signal to N2Image pixelThe conversion of (1).
4. The GRU-based rotary machine fault diagnosis method according to claim 1, wherein the step 3) inputs the training data set into a linear layer to increase the dimensionality of each line of data in the image, the linear layer being a fully connected layer that performs a linear transformation.
5. The method of claim 1, wherein in step 4), the GRU includes two gates: reset gate rtAnd an update gate zt(ii) a When the input sequence is xtThe output of the two gated and hidden states is as follows:
rt＝σ(Wrxt+Urht-1+br)
zt＝σ(Wzxt+Uzht-1+bz)
ht＝zt⊙ht-1+(1-zt)⊙tanh(Whxt+Uh(rt⊙ht-1)+bh)
wherein, Wr、Wz、WhRespectively representing the weight matrix between the reset gate and the input, the weight matrix between the update gate and the input, the weight matrix between the hidden state and the input at the present moment, Ur、Uz、UhRespectively representing a weight matrix between the reset gate and the previous hidden layer, a weight matrix between the update gate and the previous hidden layer, a weight matrix between the hidden state at the current moment and the previous hidden layer, ht-1σ is the Logistic function for the hidden state at the previous time.
6. The GRU-based rotary machine fault diagnosis method of claim 1, wherein the GRU uses a one-way structure, a residual join is added to the GRU, and the step length T is set to N.
7. The GRU-based rotary machine fault diagnosis method according to claim 1, wherein a multi-layer sensor is adopted as the classifier in the step 5), the multi-layer sensor comprises two hidden layers and an output layer, the number of neurons of each hidden layer is set according to the dimension of each line of data in the step 3), the number of neurons of the output layer is set according to the number of fault state markers in the step 1), and a ReLU function is used as the activation function.
8. The GRU-based rotary machine fault diagnosis method according to claim 7, wherein a batch normalization and discarding method is added to each hidden layer of the multilayer perceptron in step 5), and a discarding rate is set to 0.5.
9. The GRU-based rotary machine fault diagnosis method of claim 7, wherein in the step 5) during the training of the classifier, a learning rate attenuation strategy is adopted, and parameters are set as follows: the number of training times was set to 35, the initial empirical learning rate was used in the first 25 training stages, set to 0.0002, and then the learning rate was decreased by an attenuation rate of 0.1; batch size 24; selecting an RMSProp algorithm as an optimizer; selecting a cross entropy loss function as an objective function, which is specifically defined as follows:
where m is the number of samples, θ is the training parameter of the model, x(i)And y(i)For the ith sample and its corresponding label, h (-) represents the proposed classifier model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011355499.XA CN112488179A (en) | 2020-11-26 | 2020-11-26 | Rotary machine fault diagnosis method based on GRU |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011355499.XA CN112488179A (en) | 2020-11-26 | 2020-11-26 | Rotary machine fault diagnosis method based on GRU |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112488179A true CN112488179A (en) | 2021-03-12 |
Family
ID=74935913
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011355499.XA Pending CN112488179A (en) | 2020-11-26 | 2020-11-26 | Rotary machine fault diagnosis method based on GRU |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112488179A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113432877A (en) * | 2021-06-26 | 2021-09-24 | 郑州航空工业管理学院 | Complex rotating mechanical equipment fault diagnosis method based on visual feature fusion |
| CN114002597A (en) * | 2021-10-25 | 2022-02-01 | 浙江理工大学 | Motor fault diagnosis method and system based on GRU network stator current analysis |
| CN115446663A (en) * | 2022-10-14 | 2022-12-09 | 常州先进制造技术研究所 | Cutter wear state monitoring method based on deep learning network of physical guidance and application |
| CN117422935A (en) * | 2023-12-13 | 2024-01-19 | 深圳市鑫思诚科技有限公司 | Motorcycle fault non-contact diagnosis method and system |
| CN117830750A (en) * | 2024-03-04 | 2024-04-05 | 青岛大学 | Mechanical fault prediction method based on graph converter |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105931402A (en) * | 2016-06-27 | 2016-09-07 | 上海波汇科技股份有限公司 | Optical fiber perimeter intrusion monitoring method based on image recognition |
| CN107643181A (en) * | 2016-07-21 | 2018-01-30 | 北京航空航天大学 | A kind of rolling bearing variable working condition method for diagnosing faults based on image recognition |
| CN109827652A (en) * | 2018-11-26 | 2019-05-31 | 河海大学常州校区 | One kind being directed to Fibre Optical Sensor vibration signal recognition and system |
| CN109934335A (en) * | 2019-03-05 | 2019-06-25 | 清华大学 | High-speed railway track switch method for diagnosing faults based on interacting depth study |
| CN110597240A (en) * | 2019-10-24 | 2019-12-20 | 福州大学 | Hydroelectric generating set fault diagnosis method based on deep learning |
-
2020
- 2020-11-26 CN CN202011355499.XA patent/CN112488179A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105931402A (en) * | 2016-06-27 | 2016-09-07 | 上海波汇科技股份有限公司 | Optical fiber perimeter intrusion monitoring method based on image recognition |
| CN107643181A (en) * | 2016-07-21 | 2018-01-30 | 北京航空航天大学 | A kind of rolling bearing variable working condition method for diagnosing faults based on image recognition |
| CN109827652A (en) * | 2018-11-26 | 2019-05-31 | 河海大学常州校区 | One kind being directed to Fibre Optical Sensor vibration signal recognition and system |
| CN109934335A (en) * | 2019-03-05 | 2019-06-25 | 清华大学 | High-speed railway track switch method for diagnosing faults based on interacting depth study |
| CN110597240A (en) * | 2019-10-24 | 2019-12-20 | 福州大学 | Hydroelectric generating set fault diagnosis method based on deep learning |
Non-Patent Citations (6)
| Title |
|---|
| 杨菊花等: "基于CNN-GRU模型的道岔故障诊断算法研究", 《铁道学报》 * |
| 杨菊花等: "基于CNN-GRU模型的道岔故障诊断算法研究", 《铁道学报》, no. 07, 15 July 2020 (2020-07-15) * |
| 王文庆: "基于CNN与GRU的电液伺服阀故障预测研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技II辑》 * |
| 王文庆: "基于CNN与GRU的电液伺服阀故障预测研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技II辑》, 15 June 2020 (2020-06-15), pages 57 - 58 * |
| 郑洋: "风力发电系统的故障诊断方法研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技II辑》 * |
| 郑洋: "风力发电系统的故障诊断方法研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技II辑》, 15 December 2019 (2019-12-15), pages 50 - 55 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113432877A (en) * | 2021-06-26 | 2021-09-24 | 郑州航空工业管理学院 | Complex rotating mechanical equipment fault diagnosis method based on visual feature fusion |
| CN113432877B (en) * | 2021-06-26 | 2023-08-18 | 郑州航空工业管理学院 | Complex rotary mechanical equipment fault diagnosis method based on visual feature fusion |
| CN114002597A (en) * | 2021-10-25 | 2022-02-01 | 浙江理工大学 | Motor fault diagnosis method and system based on GRU network stator current analysis |
| CN115446663A (en) * | 2022-10-14 | 2022-12-09 | 常州先进制造技术研究所 | Cutter wear state monitoring method based on deep learning network of physical guidance and application |
| CN115446663B (en) * | 2022-10-14 | 2023-10-20 | 常州先进制造技术研究所 | Tool wear state monitoring method based on physical guidance deep learning network and application |
| CN117422935A (en) * | 2023-12-13 | 2024-01-19 | 深圳市鑫思诚科技有限公司 | Motorcycle fault non-contact diagnosis method and system |
| CN117422935B (en) * | 2023-12-13 | 2024-03-08 | 深圳市鑫思诚科技有限公司 | Motorcycle fault non-contact diagnosis method and system |
| CN117830750A (en) * | 2024-03-04 | 2024-04-05 | 青岛大学 | Mechanical fault prediction method based on graph converter |
| CN117830750B (en) * | 2024-03-04 | 2024-06-04 | 青岛大学 | Mechanical fault prediction method based on graph converter |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112488179A (en) | Rotary machine fault diagnosis method based on GRU | |
| CN109765053B (en) | Rolling bearing fault diagnosis method using convolutional neural network and kurtosis index | |
| WO2022077901A1 (en) | Bearing failure mode diagnosis method using small sample data sets, and system | |
| CN111721535B (en) | Bearing fault detection method based on convolution multi-head self-attention mechanism | |
| CN109858352B (en) | Fault diagnosis method based on compressed sensing and improved multi-scale network | |
| CN112964469B (en) | Online fault diagnosis method for rolling bearing under variable load of transfer learning | |
| CN113988126B (en) | Rolling bearing fault diagnosis method based on few-label data feature migration | |
| CN111397902B (en) | Rolling bearing fault diagnosis method based on feature alignment convolutional neural network | |
| CN111650453A (en) | Power equipment diagnosis method and system based on windowing characteristic Hilbert imaging | |
| CN110110768A (en) | Fault Diagnosis of Roller Bearings based on Concurrent Feature study and multi-categorizer | |
| CN112763214A (en) | Rolling bearing fault diagnosis method based on multi-label zero-sample learning | |
| CN113392881A (en) | Rotary machine fault diagnosis method based on transfer learning | |
| CN115290326A (en) | Rolling bearing fault intelligent diagnosis method | |
| CN112017204A (en) | Tool state image classification method based on edge marker graph neural network | |
| CN112001273A (en) | Fault diagnosis system and method combining convolution self-encoder and logistic regression | |
| CN115791174B (en) | Rolling bearing abnormality diagnosis method, system, electronic equipment and storage medium | |
| CN115358259A (en) | Self-learning-based unsupervised cross-working-condition bearing fault diagnosis method | |
| CN116702076A (en) | Small sample migration learning fault diagnosis method, system, computer and storage medium based on CNN feature fusion | |
| CN116842379A (en) | Mechanical bearing residual service life prediction method based on DRSN-CS and BiGRU+MLP models | |
| CN116861343A (en) | Bearing fault diagnosis method | |
| CN114925809B (en) | LSTM-based printer bearing fault diagnosis method and device | |
| CN115452376A (en) | Bearing fault diagnosis method based on improved lightweight deep convolution neural network | |
| CN112149804B (en) | Novel convolutional neural network model and application | |
| CN113723592A (en) | Fault diagnosis method based on wind power gear box monitoring system | |
| Chen et al. | Transfer learning with unsupervised domain adaptation method for bearing fault diagnosis |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210312 |
|
| RJ01 | Rejection of invention patent application after publication |