CN114593917A - Small sample bearing fault diagnosis method based on triple model - Google Patents
Small sample bearing fault diagnosis method based on triple model Download PDFInfo
- Publication number
- CN114593917A CN114593917A CN202210220876.1A CN202210220876A CN114593917A CN 114593917 A CN114593917 A CN 114593917A CN 202210220876 A CN202210220876 A CN 202210220876A CN 114593917 A CN114593917 A CN 114593917A
- Authority
- CN
- China
- Prior art keywords
- model
- bearing
- sample
- triple
- fault
- 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
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000003745 diagnosis Methods 0.000 title claims abstract description 21
- 239000013598 vector Substances 0.000 claims abstract description 47
- 238000012549 training Methods 0.000 claims abstract description 44
- 238000012360 testing method Methods 0.000 claims abstract description 28
- 238000007781 pre-processing Methods 0.000 claims abstract description 5
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 4
- 238000012935 Averaging Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 4
- 238000005096 rolling process Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 241001391944 Commicarpus scandens Species 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
- G01M13/04—Bearings
- G01M13/045—Acoustic or vibration analysis
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- General Physics & Mathematics (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses a small sample bearing fault diagnosis method based on a three-tuple model, which comprises the following steps of: preprocessing the acquired one-dimensional time sequence signal of the original bearing to obtain a two-dimensional time-frequency image; randomly selecting a sample pair to input a triple model training and iteratively updating model parameters; extracting the feature vectors of the training samples and calculating the mean value of the feature vectors of various bearing faults; and judging the fault category of the test set sample so as to calculate the performance of the model. The method utilizes the triple model to learn the characteristics in and among the bearing fault classes, and can obtain higher bearing fault recognition rate under the condition of less training samples by using the model.
Description
Technical Field
The invention belongs to the technical field of machinery, and particularly relates to a small sample bearing fault diagnosis method based on a triple model.
Background
With the rapid development of industrialization, a large number of intelligent rotating machinery devices are adopted in numerous fields such as aerospace, electric power and manufacturing industries, bearings are used as important component parts of the rotating machinery, and the health condition of the bearings is concerned about the normal operation of the devices. However, the working environment of the equipment is severe, and the bearing is easy to break down, so that if the bearing can be found out in time, a large amount of manpower and material resources can be saved, and even serious accidents can be avoided. At present, with the development of artificial intelligence, deep learning is widely used in bearing fault diagnosis, and a good fault recognition rate is obtained, but most of the deep learning-based bearing fault diagnosis methods rely on a large number of training samples to obtain satisfactory results. However, in actual production practice, a sufficient number of sample training models cannot be obtained, so that a small-sample bearing fault diagnosis method based on a triple model is provided.
Disclosure of Invention
In view of the above problems, an object of the present invention is to provide a small sample bearing fault diagnosis method based on a triple model, which can obtain a high bearing fault recognition rate under the condition of a small number of bearing fault training samples.
In order to achieve the above object, the present invention comprises the steps of:
step 1: acquiring a vibration time sequence signal of a bearing, and dividing the vibration time sequence signal into a training set and a testing set;
step 2: respectively preprocessing signals of the training set and the test set, and converting one-dimensional time sequence signals into two-dimensional signals;
and step 3: inputting the processed training set samples into a triple model, randomly selecting three samples from the training set each time as the input of the triple model, respectively mapping the samples to a 128-dimensional vector space through a sub-model of the triple model to obtain three corresponding characteristic vectors, guiding the optimization direction of the triple model by using the similarity of the three characteristic vectors as the loss value of the model, and repeating the steps until the triple model is converged;
and 4, step 4: inputting the samples in the training set into the triple model to obtain the characteristic vector of each sample, and averaging the characteristic vectors of various fault samples to obtain the characteristic vector mean value of each type of bearing fault;
and 5: inputting the test set samples into the model to obtain the feature vector of each test set sample, measuring the similarity between the feature vector of each test sample and the mean value of the feature vector of each type of bearing fault obtained in the step 4, and considering the test set sample as the bearing fault if the feature similarity is the highest.
Further, in step 1, the vibration timing signals of the bearing include a healthy bearing signal, a bearing signal of a rolling element failure, a bearing signal of an outer ring failure, and a bearing signal of an inner ring failure.
Furthermore, in step 1, the vibration timing signal of the bearing is divided into two parts according to a time sequence, the former part obtains a training set sample by using a sliding window sequential sampling method, and the latter part obtains a test set sample by using the same method.
Further, in step 2, a one-dimensional time-frequency signal is converted into a two-dimensional time-frequency signal by using a fast fourier transform, and the time-frequency signal is stored as a 33x33 picture.
Further, in said step 3, samples of the triad model are input, two of which are the same bearing fault and the other is a different bearing fault.
Furthermore, in step 3, the triplet model is composed of three identical subnetworks, and weights of the three subnetworks are shared, and the subnetworks have a similar structure as the residual network.
Further, in step 3, the similarity between two feature vectors is measured by using euclidean distance:
wherein x isi,xjIn order to input the samples, the method,representing the dimensions of the xi sample feature vector.
Further, in step 3, the loss function of the triplet model is:
Loss(x1,x2,x3)=max(D(x1,x2)-D(x1,x3)+marain,0)
wherein Loss is Loss value, x1,x2,x3For inputting training samples, D (-) is Euclidean distance, and margin is an artificially set threshold value used for controlling the distance between the same fault and different fault feature vectors.
Further, in the step 3, Adam is used as an optimizer for training the model, and a learning rate exponential decay strategy is used to accelerate the training of the model.
Compared with the prior art, the invention has obvious advantages and beneficial effects. According to the technical scheme, the original time sequence signal is converted into the time-frequency signal through short-time Fourier transform, and the training sample is converted into the two-dimensional image more suitable for convolution; by learning both intra-class and inter-class faults for a bearing, the model can relate the class information of the bearing fault, not just the characteristic information of a single sample. Therefore, the invention can obtain higher fault recognition rate under the condition of a small number of training samples.
Drawings
FIG. 1 is a flow chart of the present invention
FIG. 2 illustrates the data set partitioning and data preprocessing
FIG. 3 is a schematic diagram of a three-tuple model framework
FIG. 4 is a schematic diagram of a three-tuple model subnetwork structure
Detailed Description
As shown in fig. 1 to 4, the method for diagnosing the fault of the small sample bearing based on the triple model of the present invention includes the following steps:
step 1: acquiring a vibration time sequence signal of a bearing, and dividing the vibration time sequence signal into a training set and a testing set;
step 2: respectively preprocessing signals of the training set and the test set, and converting one-dimensional time sequence signals into two-dimensional signals;
and step 3: inputting the processed training set samples into a triple model, randomly selecting three samples from the training set each time as the input of the triple model, respectively mapping the samples to a 128-dimensional vector space through a sub-model of the triple model to obtain three corresponding characteristic vectors, guiding the optimization direction of the triple model by using the similarity of the three characteristic vectors as the loss value of the model, and repeating the steps until the triple model is converged;
and 4, step 4: inputting the samples in the training set into the triple model to obtain the characteristic vector of each sample, and averaging the characteristic vectors of various fault samples to obtain the characteristic vector mean value of each type of bearing fault;
and 5: inputting the test set samples into the model to obtain the feature vector of each test set sample, measuring the similarity between the feature vector of each test sample and the mean value of the feature vector of each type of bearing fault obtained in the step 4, and considering the test set sample as the bearing fault if the feature similarity is the highest.
In the step 1, the vibration time sequence signals of the bearing comprise healthy bearing signals, rolling element fault bearing signals, outer ring fault bearing signals and inner ring fault bearing signals.
In the step 1, the vibration time sequence signal of the bearing is divided into two parts according to the time sequence, the former part adopts a method of sliding window sequential sampling to obtain a training set sample, and the latter part obtains a test set sample by the same method.
In the step 2, a one-dimensional time sequence signal is converted into a two-dimensional time frequency signal by using fast fourier transform, and the time frequency signal is stored as a 33x33 picture.
In step 3, a sample of the triad model is input, two of which are the same bearing fault and the other is a different bearing fault.
In step 3, the three-tuple model is composed of three identical sub-networks, the weights of the three sub-networks are shared, and the sub-networks have a structure similar to that of the residual error network.
In step 3, the similarity of the two feature vectors is measured by using the euclidean distance:
wherein x isi,xjIn order to input the samples, the method,representing the dimensions of the xi sample feature vector.
Further, in step 3, the loss function of the triplet model is:
Loss(x1,x2,x3)=max(D(x1,x2)-D(x1,x3)+marain,0)
wherein Loss is Loss value, x1,x2,x3For inputting training samples, D (-) is Euclidean distance, and margin is an artificially set threshold value used for controlling the distance between the same fault and different fault feature vectors.
In the step 3, the optimizer adopted in the model training is Adam, and the learning rate exponential decay strategy is used for accelerating the model training.
Examples
In the embodiment, a bearing data set of university in south and south of the Yangtze river is adopted, noise signals of four states of bearing health, bearing inner ring fault, bearing outer ring fault and bearing rolling element fault at the rotating speeds of 600, 800 and 1000 are respectively collected by the data set, and the sampling frequency is 50 kHz. In the embodiment, bearing data at the rotating speed of 800 is adopted to verify the small sample bearing fault diagnosis method based on the triple model. The method comprises the following specific steps:
step 1: acquiring a vibration time sequence signal of a bearing, and dividing the vibration time sequence signal into a training set and a testing set;
step 2: the signals of the training set and the test set are preprocessed respectively, and one-dimensional time sequence signals are converted into two-dimensional signals, as shown in table 1
TABLE 1 bearing failure data set
And step 3: randomly extracting 5, 10, 15, 30, 50, 80 and 100 samples from the bearing faults in the training set, and verifying the identification accuracy of the small sample bearing fault diagnosis method based on the triple model under different numbers of training samples. Inputting the processed training set samples into a triple model, randomly selecting three samples from the training set each time as the input of the triple model, respectively mapping the samples to a 128-dimensional vector space through a sub-model of the triple model to obtain three corresponding characteristic vectors, and guiding the optimization direction of the triple model by taking the similarity of the three characteristic vectors as the loss value of the model. Setting a model optimizer as Adam, a learning rate of 0.01 and a maximum iteration number of 40, and accelerating model training by using a learning rate exponential decay strategy;
and 4, step 4: inputting the samples in the training set into the triple model to obtain the characteristic vector of each sample, and averaging the characteristic vectors of various fault samples to obtain the characteristic vector mean value of each type of bearing fault;
and 5: inputting the test set samples into the model to obtain the feature vector of each test set sample, measuring the similarity between the feature vector of each test sample and the mean value of the feature vector of each type of bearing fault obtained in the step 4, and considering the test set sample as the bearing fault if the feature similarity is the highest.
Table 2 shows the average accuracy after repeating the experiment 5 times, and comparing the average accuracy with a 2DCNN (two-dimensional convolution model, the model structure is the same as the model structure of the present invention) model and a 1DCNN (one-dimensional convolution model) model, and the comparison shows the advantage of the present invention in bearing fault diagnosis under a small sample.
TABLE 2 comparison of bearing fault diagnosis methods for different training sample numbers
Claims (9)
1. A small sample bearing fault diagnosis method based on a three-unit model is characterized by comprising the following steps:
step 1: acquiring a vibration time sequence signal of a bearing, and dividing the vibration time sequence signal into a training set and a testing set;
step 2: respectively preprocessing signals of the training set and the test set, and converting one-dimensional time sequence signals into two-dimensional signals;
and step 3: inputting the processed training set samples into a triple model, randomly selecting three samples from the training set each time as the input of the triple model, respectively mapping the samples to a 128-dimensional vector space through a sub-model of the triple model to obtain three corresponding characteristic vectors, guiding the optimization direction of the triple model by using the similarity of the three characteristic vectors as the loss value of the model, and repeating the steps until the triple model is converged;
and 4, step 4: inputting the samples in the training set into the triple model to obtain the characteristic vector of each sample, and averaging the characteristic vectors of various fault samples to obtain the characteristic vector mean value of each type of bearing fault;
and 5: inputting the test set samples into the model to obtain the feature vector of each test set sample, measuring the similarity between the feature vector of each test sample and the mean value of the feature vector of each type of bearing fault obtained in the step 4, and considering the test set sample as the bearing fault if the feature similarity is the highest.
2. The small-sample bearing fault diagnosis method based on the triple model as set forth in claim 1, characterized in that: in the step 1, the vibration time sequence signals of the bearing comprise healthy bearing signals, bearing signals with fault rolling bodies, bearing signals with fault outer rings and bearing signals with fault inner rings.
3. The small-sample bearing fault diagnosis method based on the triple model as set forth in claim 1, characterized in that: in the step 1, the vibration time sequence signal of the bearing is divided into two parts according to the time sequence, the former part adopts the method of sliding window sequential sampling to obtain a training set sample, and the latter part obtains a test set sample by the same method.
4. The small-sample bearing fault diagnosis method based on the triple model as claimed in claim 1, characterized in that: in the step 2, a one-dimensional time sequence signal is converted into a two-dimensional time frequency signal by using fast fourier transform, and the time frequency signal is stored as a 33x33 picture.
5. The small-sample bearing fault diagnosis method based on the triple model as set forth in claim 1, characterized in that: in step 3, a sample of the triad model is input, two of which are the same bearing fault and the other is a different bearing fault.
6. The small-sample bearing fault diagnosis method based on the triple model as set forth in claim 1, characterized in that: in step 3, the triplet model is composed of three identical sub-networks, the weights of the three sub-networks are shared, and the sub-networks have a structure similar to that of the residual error network.
7. The small-sample bearing fault diagnosis method based on the triple model as set forth in claim 1, characterized in that: in step 3, the similarity of the two feature vectors is measured by using the euclidean distance:
8. The small-sample bearing fault diagnosis method based on the triple model as claimed in claim 1, characterized in that: in step 3, the loss function of the triplet model is:
Loss(x1,x2,x3)=max(D(x1,x2)―D(x1,x3)+marain,0)
wherein, Loss is Loss value, x1,x2,x3For inputting training samples, D (-) is Euclidean distance, and margin is an artificially set threshold value used for controlling the distance between the same fault and different fault feature vectors.
9. The small-sample bearing fault diagnosis method based on the triple model as claimed in claim 1, characterized in that: in the step 3, the optimizer adopted in the model training is Adam, and the learning rate exponential decay strategy is used for accelerating the model training.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210220876.1A CN114593917A (en) | 2022-03-08 | 2022-03-08 | Small sample bearing fault diagnosis method based on triple model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210220876.1A CN114593917A (en) | 2022-03-08 | 2022-03-08 | Small sample bearing fault diagnosis method based on triple model |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114593917A true CN114593917A (en) | 2022-06-07 |
Family
ID=81815478
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210220876.1A Pending CN114593917A (en) | 2022-03-08 | 2022-03-08 | Small sample bearing fault diagnosis method based on triple model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114593917A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN118035873A (en) * | 2024-02-29 | 2024-05-14 | 华中科技大学 | Fault diagnosis method for parallel convolution improved triple network |
Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
AU9432398A (en) * | 1997-10-06 | 1999-04-27 | Leo Pharmaceutical Products Ltd. A/S (Lovens Kemiske Fabrik Produktionsaktieselskab) | Novel vitamin d analogues |
CN107677472A (en) * | 2017-08-25 | 2018-02-09 | 中国矿业大学 | The bearing state noise diagnostics algorithm that network-oriented Variable Selection merges with Characteristic Entropy |
CN108760302A (en) * | 2018-05-08 | 2018-11-06 | 南京风电科技有限公司 | A kind of on-line monitoring and fault diagnosis system of wind power generating set bearing |
CN110146289A (en) * | 2019-05-28 | 2019-08-20 | 昆明理工大学 | A kind of rolling bearing Weak fault feature extracting method |
CN110555478A (en) * | 2019-09-05 | 2019-12-10 | 东北大学 | Fan multi-fault diagnosis method based on depth measurement network of difficult sample mining |
CN112304611A (en) * | 2020-09-14 | 2021-02-02 | 江苏大学 | Deep learning-based bearing fault diagnosis method |
US20210157312A1 (en) * | 2016-05-09 | 2021-05-27 | Strong Force Iot Portfolio 2016, Llc | Intelligent vibration digital twin systems and methods for industrial environments |
CN112924177A (en) * | 2021-04-02 | 2021-06-08 | 哈尔滨理工大学 | Rolling bearing fault diagnosis method for improved deep Q network |
CN113627317A (en) * | 2021-08-06 | 2021-11-09 | 安徽工业大学 | Motor bearing fault diagnosis method based on single sample learning |
CN114091553A (en) * | 2020-08-06 | 2022-02-25 | 长沙理工大学 | Diagnosis method for rolling bearing fault |
CN114444222A (en) * | 2022-01-14 | 2022-05-06 | 合肥工业大学 | Similarity calculation model, cross-category data diagnosis model, method and system |
CN114624027A (en) * | 2022-03-14 | 2022-06-14 | 安徽理工大学 | Bearing fault diagnosis method based on multi-input CNN |
-
2022
- 2022-03-08 CN CN202210220876.1A patent/CN114593917A/en active Pending
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
AU9432398A (en) * | 1997-10-06 | 1999-04-27 | Leo Pharmaceutical Products Ltd. A/S (Lovens Kemiske Fabrik Produktionsaktieselskab) | Novel vitamin d analogues |
US20210157312A1 (en) * | 2016-05-09 | 2021-05-27 | Strong Force Iot Portfolio 2016, Llc | Intelligent vibration digital twin systems and methods for industrial environments |
CN107677472A (en) * | 2017-08-25 | 2018-02-09 | 中国矿业大学 | The bearing state noise diagnostics algorithm that network-oriented Variable Selection merges with Characteristic Entropy |
CN108760302A (en) * | 2018-05-08 | 2018-11-06 | 南京风电科技有限公司 | A kind of on-line monitoring and fault diagnosis system of wind power generating set bearing |
CN110146289A (en) * | 2019-05-28 | 2019-08-20 | 昆明理工大学 | A kind of rolling bearing Weak fault feature extracting method |
CN110555478A (en) * | 2019-09-05 | 2019-12-10 | 东北大学 | Fan multi-fault diagnosis method based on depth measurement network of difficult sample mining |
CN114091553A (en) * | 2020-08-06 | 2022-02-25 | 长沙理工大学 | Diagnosis method for rolling bearing fault |
CN112304611A (en) * | 2020-09-14 | 2021-02-02 | 江苏大学 | Deep learning-based bearing fault diagnosis method |
CN112924177A (en) * | 2021-04-02 | 2021-06-08 | 哈尔滨理工大学 | Rolling bearing fault diagnosis method for improved deep Q network |
CN113627317A (en) * | 2021-08-06 | 2021-11-09 | 安徽工业大学 | Motor bearing fault diagnosis method based on single sample learning |
CN114444222A (en) * | 2022-01-14 | 2022-05-06 | 合肥工业大学 | Similarity calculation model, cross-category data diagnosis model, method and system |
CN114624027A (en) * | 2022-03-14 | 2022-06-14 | 安徽理工大学 | Bearing fault diagnosis method based on multi-input CNN |
Non-Patent Citations (4)
Title |
---|
PENG BS 等: "An intelligent fault diagnosis method for rotating machinery based on data fusion and deep residual neural network", 《APPLIED INTELLIGENCE》, vol. 52, no. 3, pages 3051 - 3065, XP037687667, DOI: 10.1007/s10489-021-02555-4 * |
张军 等: "融合注意力机制的二维CNN变工况轴承故障诊断", 《安徽理工大学学报《自然科学版)》, vol. 42, no. 2, pages 66 - 72 * |
张永宏;王逸飞;赵晓平;吴家新;王丽华;: "基于深度度量学习的电机故障诊断", 测控技术, no. 07, pages 36 - 43 * |
谢由生 等: "基于Triplet Network的小样本轴承、齿轮故障诊断方法", 《机电工程》, vol. 39, no. 8, pages 1111 - 1117 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN118035873A (en) * | 2024-02-29 | 2024-05-14 | 华中科技大学 | Fault diagnosis method for parallel convolution improved triple network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112418013B (en) | Complex working condition bearing fault diagnosis method based on meta-learning under small sample | |
CN112763214B (en) | Rolling bearing fault diagnosis method based on multi-label zero-sample learning | |
CN111709448A (en) | Mechanical fault diagnosis method based on migration relation network | |
CN111562108A (en) | Rolling bearing intelligent fault diagnosis method based on CNN and FCMC | |
CN105841961A (en) | Bearing fault diagnosis method based on Morlet wavelet transformation and convolutional neural network | |
CN111650453A (en) | Power equipment diagnosis method and system based on windowing characteristic Hilbert imaging | |
CN109946080B (en) | Mechanical equipment health state identification method based on embedded circulation network | |
CN109858352A (en) | A kind of method for diagnosing faults based on compressed sensing and the multiple dimensioned network of improvement | |
CN111275108A (en) | Method for performing sample expansion on partial discharge data based on generation countermeasure network | |
CN113569990B (en) | Strong noise interference environment-oriented performance equipment fault diagnosis model construction method | |
CN112926728B (en) | Small sample turn-to-turn short circuit fault diagnosis method for permanent magnet synchronous motor | |
CN110726898A (en) | Power distribution network fault type identification method | |
CN112364706A (en) | Small sample bearing fault diagnosis method based on class imbalance | |
CN112070104A (en) | Main transformer partial discharge identification method | |
CN114169377A (en) | G-MSCNN-based fault diagnosis method for rolling bearing in noisy environment | |
CN115859077A (en) | Multi-feature fusion motor small sample fault diagnosis method under variable working conditions | |
CN114491823B (en) | Train bearing fault diagnosis method based on improved generation countermeasure network | |
CN114593917A (en) | Small sample bearing fault diagnosis method based on triple model | |
CN115859702A (en) | Permanent magnet synchronous wind driven generator demagnetization fault diagnosis method and system based on convolutional neural network | |
CN115481657A (en) | Wind generating set communication slip ring fault diagnosis method based on electric signals | |
CN116186641A (en) | Cross-equipment rolling bearing small sample fault diagnosis method based on feature fusion | |
CN113758709A (en) | Rolling bearing fault diagnosis method and system combining edge calculation and deep learning | |
CN113705695A (en) | Power distribution network fault data identification method based on convolutional neural network | |
CN116821762A (en) | Mechanical fault diagnosis method based on multi-scale graph attention fusion network | |
CN115525866A (en) | Deep learning rolling bearing fault diagnosis method and system |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20220607 |
|
WD01 | Invention patent application deemed withdrawn after publication |