CN114970044B - Rolling bearing fault diagnosis method and system based on threshold convolutional neural network - Google Patents
Rolling bearing fault diagnosis method and system based on threshold convolutional neural network Download PDFInfo
- Publication number
- CN114970044B CN114970044B CN202210701053.0A CN202210701053A CN114970044B CN 114970044 B CN114970044 B CN 114970044B CN 202210701053 A CN202210701053 A CN 202210701053A CN 114970044 B CN114970044 B CN 114970044B
- Authority
- CN
- China
- Prior art keywords
- threshold
- rolling bearing
- module
- neural network
- fault diagnosis
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/17—Mechanical parametric or variational design
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/10—Noise analysis or noise optimisation
Abstract
The invention belongs to the field of fault type identification and diagnosis of rotary machinery, and particularly relates to a rolling bearing fault diagnosis method and system based on a threshold convolutional neural network, wherein the method comprises the following main points: 1. training vibration history data of faults contained in the rolling bearing to obtain a rolling bearing fault diagnosis model based on a threshold convolutional neural network; 2. collecting vibration data of the rolling bearing in real time; 3. monitoring vibration data in real time by using a trained fault diagnosis model, and identifying whether faults exist; 4. if the fault exists, fault characteristics are extracted and classified by using a fault diagnosis model, and a fault identification result is output. The invention can solve the problem that the performance is reduced when the convolutional neural network model performs feature extraction on the vibration data containing noise, has higher fault diagnosis accuracy, accelerates the diagnosis time, accords with the criteria of simplicity, practicability and strong portability in the industry, and has higher application value.
Description
Technical Field
The invention relates to the technical field of fault type identification and diagnosis of rotary machinery, in particular to a rolling bearing fault diagnosis method and system based on a threshold convolutional neural network.
Background
There are many mechanical devices in the control field that play an important role in the fields of wind power generation, power supply, aerospace, etc. Rotary machines are an integral part of mechanical equipment, the stable operation of which is an important prerequisite for ensuring production and human safety, however, they generally operate in harsh environments, increasing the probability of failure of the rotary machine, which, once it fails, causes losses and increases maintenance costs. Therefore, it is very important to improve the fault diagnosis precision of the rolling bearing under the noise environment, the equipment repair and maintenance cost can be effectively reduced, the running economy of mechanical equipment is improved, and the safety production is ensured.
Disclosure of Invention
Aiming at the problems in the background, the invention provides a rolling bearing fault diagnosis method and a rolling bearing fault diagnosis system based on a threshold convolutional neural network, which are used for solving the problem that the performance is reduced when the convolutional neural network model performs feature extraction on vibration data containing noise, and realizing intelligent monitoring on the fault type of the rolling bearing.
In order to achieve the above purpose, the present invention provides the following technical solutions: a rolling bearing fault diagnosis method based on a threshold convolutional neural network, the diagnosis method comprising the steps of:
1) Training vibration history data of faults contained in the rolling bearing to obtain a rolling bearing fault diagnosis model based on a threshold convolutional neural network;
2) Collecting vibration data of the rolling bearing in real time;
3) Monitoring vibration data in real time by using a trained fault diagnosis model, and identifying whether faults exist;
4) If the fault exists, fault characteristics are extracted and classified by using a fault diagnosis model, and a fault identification result is output.
Preferably, the input data is subjected to feature extraction by the modules 1, 2, 3 and 4.
Preferably, the step 1 includes:
step 1.1, module 1 and module 2 consist of a convolution layer, a rectifying linear unit and a maximum pooling layer, respectively. Sliding the input data by using a filter according to the step length and performing convolution operation to extract data characteristics; the rectification linear unit performs sparsification on the data characteristics, wherein R=max (0, x), x represents input data, and R represents the rectification linear unit; and the maximum pooling operation compresses the thinned characteristics. The forward learning mathematical model is built as follows:
y 1 ＝f 1 (x,k 1 ) (1)
y 2 ＝f 2 (y 1 ,k 2 )＝f 2 (f 1 (x),k 2 ) (2)
wherein k is 1 And k 2 Representing the number of filters, f 1 And f 2 Respectively, the convolution, rectification and maximum pooling operations of the modules 1 and 2, y 1 And y 2 The outputs of module 1 and module 2, respectively.
Preferably, the step 1 further includes:
step 1.2, module 3 and module 4 are respectively composed of a convolution layer and a rectifying linear unit, and the forward learning mathematical model is established as follows:
y 3 ＝f 3 (y 2 ,k 3 )＝f 3 (f 2 (f 1 (x,k 1 ),k 2 ),k 3 ) (3)
y 4 ＝f 4 (y 3 ,k 4 )＝f 4 (f 3 (f 2 (f 1 (x,k 1 ),k 2 ),k 3 ),k 4 ) (4)
wherein k is 3 And k 4 Representing the number of filters, f 3 And f 4 Respectively, means the convolution and rectification linear unit operation of the module 3 and the module 4, y 3 And y 4 The outputs of modules 3 and 4, respectively.
Preferably, the step 1 further includes:
step 1.3, denoising the data characteristics extracted by the threshold convolutional neural network by using a full threshold Function (FT), wherein the FT function expression is as follows:
wherein x is FT Representing the characteristics of the input, y representing the output result, t representing the threshold value, alpha 1 Represent the weight, alpha 1 >1, useful characteristic information can be extracted in a larger proportion, 0<β 1 <And 1, the characteristic information with smaller contribution can be reserved to a certain extent, and the noise removing function is realized.
Preferably, the step 1 further includes:
step 1.4, through the rolling bearing fault diagnosis model training based on a threshold convolutional neural network, automatically acquiring a threshold t in FT, mainly comprising two calculation branches, wherein the steps are as follows:
step 1.4.1, the first branch carries out average operation of each channel on the output characteristics of the last convolution layer in the threshold convolution neural network;
step 1.4.2, carrying out normalization processing on the output characteristics by the other branch;
and step 1.4.3, multiplying the output vectors of the two branches to obtain a threshold t.
Preferably, the step 1 further includes:
step 1.5, a rolling bearing fault diagnosis model based on a threshold convolutional neural network updates parameters through reverse learning, and gradient values are as follows:
wherein D (y) is a derivative function of FT; alpha 2 Is the FT slope that enhances the effective information gradient; beta 2 Is FT slope, beta for reducing noise gradient 2 The value is smaller and approaches to 0, so that the phenomenon of gradient explosion or disappearance of the rolling bearing fault diagnosis model based on the threshold convolutional neural network in the training process can be effectively avoided; a slope of 1 indicates that no processing is done on the data.
The back propagation process of the rolling bearing fault diagnosis model based on the threshold convolution neural network is as follows:
where delta is the loss value of the last layer, when y 4 At > t, the gradient value of FT is alpha 3 (α 3 >1) To enhance the overall gradient and highlight the effective information; when-t < y 4 When < t, the gradient value of FT is beta 3 (0<β 3 <1) To reduce overall gradients, feature quantity; when y is 4 And when the value of the gradient of FT is less than-t, setting the gradient value of FT to be 1, and not performing FT operation.
Preferably, the step 1 further includes:
step 1.6, processing the output characteristics of the threshold convolutional neural network by adopting global average pooling, and obtaining the data characteristics as follows:
f pooling ＝[z 1 ,z 2 ,…,z m ] T (8)
wherein f pooling Representing the output value, z, of the global average pooling layer m Representing the average value calculated for each characteristic data, wherein m is the number of fault types; the classification function performs fault type classification on the output value of the global average pooling layer, and converts the output value into a range of [0,1 ]]And the probability distribution with 1 is obtained by taking the category corresponding to the maximum probability as an output result, and the calculation process is as follows:
wherein p is s Representing a classification function and e representing an exponential function.
A rolling bearing fault diagnosis system based on a threshold convolutional neural network, which supports the rolling bearing fault diagnosis method based on the threshold convolutional neural network, comprising:
the model training unit is used for training vibration history data of faults contained in the rolling bearing to obtain a rolling bearing fault diagnosis model based on a threshold convolutional neural network;
the data acquisition unit is used for acquiring vibration data of the rolling bearing in real time;
the processing unit is used for adopting the rolling bearing fault diagnosis model based on the threshold convolutional neural network and obtained in the model training unit to monitor vibration data in real time;
the identification unit is used for identifying whether the vibration data monitored in real time in the processing unit has faults or not, and if the faults exist, fault feature extraction and classification are carried out by using a fault diagnosis model;
and the output and display unit is used for outputting the bearing fault type output by the identification unit and displaying the bearing fault type.
Compared with the prior art, the invention has the following beneficial effects:
the invention can effectively reduce noise of vibration data containing noise and extract beneficial characteristic information. Compared with other classical deep learning models, the rolling bearing fault diagnosis model based on the threshold convolutional neural network has few layers, does not use a complex optimization method, and meets the criteria of simplicity, practicability and strong portability in the industry. The method can improve the fault diagnosis accuracy, quicken the fault diagnosis time and has higher application value.
Drawings
FIG. 1 is a schematic diagram of a rolling bearing fault diagnosis method and system based on a threshold convolutional neural network;
FIG. 2 is a FT curve;
FIG. 3 is a FT derivative function curve;
FIG. 4 is a graph of rolling bearing fault diagnosis model test accuracy and loss values based on a threshold convolutional neural network.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
As shown in fig. 1, the present embodiment provides a rolling bearing fault diagnosis method based on a threshold convolutional neural network, including the following steps:
1) Training vibration history data of faults contained in the rolling bearing to obtain a rolling bearing fault diagnosis model based on a threshold convolutional neural network;
2) Collecting vibration data of the rolling bearing in real time;
3) Monitoring vibration data in real time by using a trained fault diagnosis model, and identifying whether faults exist;
4) If the fault exists, fault characteristics are extracted and classified by using a fault diagnosis model, and a fault identification result is output.
1. In this embodiment, the rolling bearing vibration history data in step 1 includes the motor rotation speed, load, displacement and frequency, so as to form four bearing states including a normal state, an outer ring failure, an inner ring failure and a ball failure. Specifically, the results are shown in Table 1.
TABLE 1 description of different states of rolling bearing
2. In this embodiment, in the method for diagnosing a rolling bearing fault based on a threshold convolutional neural network described in step 1, feature extraction is performed on input data through a module 1, a module 2, a module 3, and a module 4, and step 1 includes:
step 1.1, module 1 and module 2 consist of a convolution layer, a rectifying linear unit and a maximum pooling layer, respectively. Sliding the input data by using a filter according to the step length and performing convolution operation to extract data characteristics; the rectification linear unit performs sparsification on the data characteristics, wherein R=max (0, x), x represents input data, and R represents the rectification linear unit; and the maximum pooling operation compresses the thinned characteristics. The forward learning mathematical model is built as follows:
y 1 ＝f 1 (x,k 1 ) (10)
y 2 ＝f 2 (y 1 ,k 2 )＝f 2 (f 1 (x),k 2 ) (11)
wherein k is 1 And k 2 Representing the number of filters, k 1 ＝125，k 2 ＝15，f 1 And f 2 Respectively, the convolution, rectification and maximum pooling operations of the modules 1 and 2, y 1 And y 2 The outputs of module 1 and module 2, respectively;
step 1.2, module 3 and module 4 are respectively composed of a convolution layer and a rectifying linear unit, and the forward learning mathematical model is established as follows:
y 3 ＝f 3 (y 2 ,k 3 )＝f 3 (f 2 (f 1 (x,k 1 ),k 2 ),k 3 ) (12)
y 4 ＝f 4 (y 3 ,k 4 )＝f 4 (f 3 (f 2 (f 1 (x,k 1 ),k 2 ),k 3 ),k 4 ) (13)
wherein k is 3 And k 4 Representing the number of filters, k 3 ＝9，k 4 ＝3，f 3 And f 4 Respectively, means the convolution and rectification linear unit operation of the module 3 and the module 4, y 3 And y 4 The outputs of module 3 and module 4, respectively;
step 1.3, denoising the data characteristics extracted by the threshold convolutional neural network by using a full threshold Function (FT), wherein the FT function expression is as follows:
wherein x is FT Representing the characteristics of the input, y representing the output result, t representing the threshold value, alpha 1 Represent the weight, alpha 1 =1.1, useful characteristic information can be extracted in a larger proportion, β 1 The device is=0.1, can reserve characteristic information with smaller contribution to a certain degree, and has a denoising function. The function curves are shown in fig. 2, where t=1;
step 1.4, through the rolling bearing fault diagnosis model training based on a threshold convolutional neural network, automatically acquiring a threshold t in FT, mainly comprising two calculation branches, wherein the steps are as follows:
step 1.4.1, the first branch carries out average operation of each channel on the output characteristics of the last convolution layer in the threshold convolution neural network;
step 1.4.2, carrying out normalization processing on the output characteristics by the other branch;
step 1.4.3, multiplying the output vectors of the two branches to obtain a threshold t;
step 1.5, a rolling bearing fault diagnosis model based on a threshold convolutional neural network updates parameters through reverse learning, and gradient values are as follows:
wherein D (y) is a derivative function of FT; alpha 2 Is the FT slope, alpha, that enhances the effective information gradient 2 =1.1; beta is FT slope to reduce noise gradient, beta 2 The rolling bearing fault diagnosis model based on the threshold convolutional neural network can be effectively prevented from gradient explosion or disappearance in the training process; a slope of 1 indicates that no processing is done on the data. The function curves are shown in fig. 3, where t=1;
the back propagation process of the rolling bearing fault diagnosis model based on the threshold convolution neural network is as follows:
where delta is the loss value of the last layer, when y 4 At > t, the gradient value of FT is alpha 3 (α 3 =1) to enhance the global gradient, highlighting valid information; when-t < y 4 When < t, the gradient value of FT is beta 3 (β 3 =0.1) to attenuate overall gradients, reduce feature numbers; when y is 4 When the value is < -t, setting the gradient value of FT to be 1, and not performing FT operation;
step 1.6, processing the output characteristics of the threshold convolutional neural network by adopting global average pooling, and obtaining the data characteristics as follows:
f pooling ＝[z 1 ,z 2 ,…,z m ] T (17)
wherein f pooling Representing the output value, z, of the global average pooling layer m Representing the average value calculated for each characteristic data, wherein m is the number of fault types; the classification function performs fault type classification on the output value of the global average pooling layer, and converts the output value into a range of [0,1 ]]And the probability distribution with 1 is obtained by taking the category corresponding to the maximum probability as an output result, and the calculation process is as follows:
wherein p is s Representing a classification function and e representing an exponential function.
Fig. 4 records the development trend of the rolling bearing fault diagnosis model based on the threshold convolutional neural network in the training and testing process, and the loss value and the testing accuracy rate can be seen to be gradually reduced, and the testing accuracy rate is gradually increased.
Example 2
Corresponding to the above method embodiment, the present embodiment provides a rolling bearing fault diagnosis system based on a threshold convolutional neural network, which supports the rolling bearing fault diagnosis method based on a threshold convolutional neural network described in embodiment 1, and includes:
the model training unit is used for training vibration history data of faults contained in the rolling bearing to obtain a rolling bearing fault diagnosis model based on a threshold convolutional neural network;
the data acquisition unit is used for acquiring vibration data of the rolling bearing in real time;
the processing unit is used for adopting the rolling bearing fault diagnosis model based on the threshold convolutional neural network and obtained in the model training unit to monitor vibration data in real time;
the identification unit is used for identifying whether the vibration data monitored in real time in the processing unit has faults or not, and if the faults exist, fault feature extraction and classification are carried out by using a fault diagnosis model;
and the output and display unit is used for outputting the bearing fault type output by the identification unit and displaying the bearing fault type.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (1)
1. A rolling bearing fault diagnosis method based on a threshold convolutional neural network, characterized in that the diagnosis method comprises the following steps:
1) Training the vibration history data of the rolling bearing containing faults to obtain a rolling bearing fault diagnosis model based on a threshold convolutional neural network, and extracting characteristics of input data through a module 1, a module 2, a module 3 and a module 4;
step 1.1, a module 1 and a module 2 are respectively composed of a convolution layer, a rectifying linear unit and a maximum pooling layer, and input data are subjected to filter sliding and convolution operation according to step length, so that data characteristics are extracted; the rectification linear unit performs sparsification on the data characteristics, wherein R=max (0, x), x represents input data, and R represents the rectification linear unit; the maximum pooling operation compresses the thinned characteristics, and the forward learning mathematical model is established as follows:
y 1 ＝f 1 (x,k 1 ) (1)
y 2 ＝f 2 (y 1 ,k 2 )＝f 2 (f 1 (x),k 2 ) (2)
wherein k is 1 And k 2 Representing the number of filters, f 1 And f 2 Respectively, the convolution and rectification lines of the module 1 and the module 2Sex units and max pooling operations, y 1 And y 2 The outputs of module 1 and module 2, respectively;
step 1.2, module 3 and module 4 are respectively composed of a convolution layer and a rectifying linear unit, and the forward learning mathematical model is established as follows:
y 3 ＝f 3 (y 2 ,k 3 )＝f 3 (f 2 (f 1 (x,k 1 ),k 2 ),k 3 ) (3)
y 4 ＝f 4 (y 3 ,k 4 )＝f 4 (f 3 (f 2 (f 1 (x,k 1 ),k 2 ),k 3 ),k 4 ) (4)
wherein k is 3 And k 4 Representing the number of filters, f 3 And f 4 Respectively, means the convolution and rectification linear unit operation of the module 3 and the module 4, y 3 And y 4 The outputs of module 3 and module 4, respectively;
step 1.3, denoising the data characteristics extracted by the threshold convolutional neural network by using a full threshold function FT, wherein the FT function expression is as follows:
wherein x is FT Representing the characteristics of the input, y representing the output result, t representing the threshold value, alpha 1 Represent the weight, alpha 1 >1, useful characteristic information can be extracted in a larger proportion, 0<β 1 <1, the characteristic information with smaller contribution can be reserved to a certain extent, and the noise removing function is realized;
step 1.4, through the rolling bearing fault diagnosis model training based on a threshold convolutional neural network, automatically acquiring a threshold t in FT, mainly comprising two calculation branches, wherein the steps are as follows:
step 1.4.1, the first branch carries out average operation of each channel on the output characteristics of the last convolution layer in the threshold convolution neural network;
step 1.4.2, carrying out normalization processing on the output characteristics by the other branch;
step 1.4.3, multiplying the output vectors of the two branches to obtain a threshold t;
step 1.5, a rolling bearing fault diagnosis model based on a threshold convolutional neural network updates parameters through reverse learning, and gradient values are as follows:
wherein D (y) is a derivative function of FT; alpha 2 Is the FT slope that enhances the effective information gradient; beta 2 Is FT slope, beta for reducing noise gradient 2 The value is smaller and approaches to 0, so that the phenomenon of gradient explosion or disappearance of the rolling bearing fault diagnosis model based on the threshold convolutional neural network in the training process can be effectively avoided; a slope of 1 means that no processing is done on the data,
the back propagation process of the rolling bearing fault diagnosis model based on the threshold convolution neural network is as follows:
where delta is the loss value of the last layer, when y 4 At > t, the gradient value of FT is alpha 3 ，α 3 >1, to enhance the overall gradient and highlight the effective information; when-t < y 4 When < t, the gradient value of FT is beta 3 ，0<β 3 <1, weakening the overall gradient and reducing the feature quantity; when y is 4 When the value is < -t, setting the gradient value of FT to be 1, and not performing FT operation;
step 1.6, processing the output characteristics of the threshold convolutional neural network by adopting global average pooling, and obtaining the data characteristics as follows:
f pooling ＝[z 1 ,z 2 ,…,z m ] T (8)
wherein f pooling Representing the output value, z, of the global average pooling layer m Representing the average of each feature dataThe value, m, is the number of fault types; the classification function performs fault type classification on the output value of the global average pooling layer, and converts the output value into a range of [0,1 ]]And the probability distribution with 1 is obtained by taking the category corresponding to the maximum probability as an output result, and the calculation process is as follows:
wherein p is s Representing a classification function, e representing an exponential function;
2) Collecting vibration data of the rolling bearing in real time;
3) Monitoring vibration data in real time by using a trained fault diagnosis model, and identifying whether faults exist;
4) If the fault exists, fault characteristics are extracted and classified by using a fault diagnosis model, and a fault identification result is output.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210701053.0A CN114970044B (en) | 2022-06-20 | 2022-06-20 | Rolling bearing fault diagnosis method and system based on threshold convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210701053.0A CN114970044B (en) | 2022-06-20 | 2022-06-20 | Rolling bearing fault diagnosis method and system based on threshold convolutional neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114970044A CN114970044A (en) | 2022-08-30 |
| CN114970044B true CN114970044B (en) | 2023-05-19 |
Family
ID=82964565
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210701053.0A Active CN114970044B (en) | 2022-06-20 | 2022-06-20 | Rolling bearing fault diagnosis method and system based on threshold convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114970044B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021016441A1 (en) * | 2019-07-23 | 2021-01-28 | Grail, Inc. | Systems and methods for determining tumor fraction |
| JP2021018683A (en) * | 2019-07-22 | 2021-02-15 | 富士通株式会社 | Information processing program, information processing method, and information processing device |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW201915727A (en) * | 2017-09-18 | 2019-04-16 | 元智大學 | Fault detection and classification method of multi-sensors |
| CN111964908B (en) * | 2020-08-17 | 2022-03-04 | 天津大学 | Bearing fault diagnosis method under variable working condition based on multi-scale weight distribution convolutional neural network model |
| CN112577747B (en) * | 2020-12-07 | 2023-08-08 | 东南大学 | Rolling bearing fault diagnosis method based on space pooling network |
| CN112729834B (en) * | 2021-01-20 | 2022-05-10 | 北京理工大学 | Bearing fault diagnosis method, device and system |
| CN114152442A (en) * | 2021-12-24 | 2022-03-08 | 杭州电子科技大学 | Rolling bearing cross-working condition fault detection method based on migration convolutional neural network |
-
2022
- 2022-06-20 CN CN202210701053.0A patent/CN114970044B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2021018683A (en) * | 2019-07-22 | 2021-02-15 | 富士通株式会社 | Information processing program, information processing method, and information processing device |
| WO2021016441A1 (en) * | 2019-07-23 | 2021-01-28 | Grail, Inc. | Systems and methods for determining tumor fraction |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114970044A (en) | 2022-08-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111046945B (en) | Fault type and damage degree diagnosis method based on combined convolutional neural network | |
| CN112254964A (en) | Rolling bearing fault diagnosis method based on rapid multi-scale convolution neural network | |
| CN112067294A (en) | Rolling bearing intelligent fault diagnosis method based on deep learning | |
| CN113008559B (en) | Bearing fault diagnosis method and system based on sparse self-encoder and Softmax | |
| CN114282579A (en) | Aviation bearing fault diagnosis method based on variational modal decomposition and residual error network | |
| CN112633098A (en) | Fault diagnosis method and system for rotary machine and storage medium | |
| CN114048787B (en) | Method and system for intelligently diagnosing bearing fault in real time based on Attention CNN model | |
| CN115828466A (en) | Fan main shaft component fault prediction method based on wide kernel convolution | |
| CN114755017B (en) | Variable-speed bearing fault diagnosis method of cross-domain data driving unsupervised field shared network | |
| CN114838936A (en) | Rolling bearing fault diagnosis method and system for sample unbalance under variable rotating speed working condition | |
| CN113916535B (en) | Bearing diagnosis method, system, equipment and medium based on time frequency and CNN | |
| Wang et al. | Fault diagnosis of industrial robots based on multi-sensor information fusion and 1D convolutional neural network | |
| CN113240022A (en) | Wind power gear box fault detection method of multi-scale single-classification convolutional network | |
| CN114970044B (en) | Rolling bearing fault diagnosis method and system based on threshold convolutional neural network | |
| Qi et al. | Feature classification method of frequency cepstrum coefficient based on weighted extreme gradient boosting | |
| CN114897292B (en) | Fault diagnosis method and system based on multi-scale residual convolution variational network | |
| Jin et al. | Multisource data fusion diagnosis method of rolling bearings based on improved multiscale CNN | |
| CN116611184A (en) | Fault detection method, device and medium for gear box | |
| CN115750341A (en) | Roots compressor fault diagnosis method and system | |
| CN115859791A (en) | Method and device for estimating residual life of aero-engine | |
| Istiaque et al. | Classification of Bearing Fault Using Radial Basis Neural Network and Genetic Algorithm | |
| CN114462446A (en) | Rolling bearing fault diagnosis method based on vibration signal and electronic equipment | |
| Li et al. | Deep learning based end-to-end rolling bearing fault diagnosis | |
| CN110836770A (en) | Automatic fault diagnosis method for gearbox based on neural network and characteristic frequency band | |
| Li et al. | Bearing fault diagnosis method using envelope analysis and euclidean distance |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |