CN111832428B - Data enhancement method applied to cold rolling mill broken belt fault diagnosis - Google Patents
Data enhancement method applied to cold rolling mill broken belt fault diagnosis Download PDFInfo
- Publication number
- CN111832428B CN111832428B CN202010578466.5A CN202010578466A CN111832428B CN 111832428 B CN111832428 B CN 111832428B CN 202010578466 A CN202010578466 A CN 202010578466A CN 111832428 B CN111832428 B CN 111832428B
- Authority
- CN
- China
- Prior art keywords
- fault
- dimensional
- data
- image
- 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
- 238000003745 diagnosis Methods 0.000 title claims abstract description 76
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000005097 cold rolling Methods 0.000 title claims abstract description 40
- 238000012549 training Methods 0.000 claims abstract description 67
- 238000012360 testing method Methods 0.000 claims abstract description 16
- 238000012545 processing Methods 0.000 claims abstract description 8
- 238000013527 convolutional neural network Methods 0.000 claims description 11
- 239000011159 matrix material Substances 0.000 claims description 7
- 238000011176 pooling Methods 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 2
- 238000009851 ferrous metallurgy Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 17
- 230000015654 memory Effects 0.000 description 6
- 238000005096 rolling process Methods 0.000 description 6
- 210000004205 output neuron Anatomy 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000003860 storage Methods 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 210000002364 input neuron Anatomy 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 4
- 230000004913 activation Effects 0.000 description 3
- 229910000831 Steel Inorganic materials 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 239000010959 steel Substances 0.000 description 2
- 230000003042 antagnostic effect Effects 0.000 description 1
- 230000008485 antagonism Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- 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
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a data enhancement method applied to cold rolling mill belt breakage fault diagnosis, and belongs to the technical field of ferrous metallurgy and fault diagnosis. The method comprises the following steps: collecting time sequence signals of a plurality of characteristics related to belt breakage fault diagnosis in cold rolling, and processing the collected time sequence signals of the plurality of characteristics to generate a two-dimensional fault image set; dividing the fault image set into training data and test data; training the auxiliary classification generation countermeasure network by using the training data and the corresponding labels to obtain a generation model, wherein the trained generation model is used for generating a fault image required by the diagnosis of the broken belt fault. By adopting the method and the device, the training speed of the generated model is improved, and the quality of the generated fault image can be improved, so that the fault image required by the diagnosis of the broken belt fault can be generated in a directed way, and the problem of insufficient fault data in the diagnosis of the broken belt fault is solved.
Description
Technical Field
The invention relates to the technical field of ferrous metallurgy and fault diagnosis, in particular to a data enhancement method applied to fault diagnosis of a broken belt of a cold rolling mill.
Background
The modern strip steel cold rolling is a high-quality and high-efficiency full-automatic production line for flexible production according to orders, and strip breakage is one of the most common faults in the cold rolling production line. Once a belt break failure occurs, the light weight can lead to equipment damage. The rolling production efficiency is affected, and fire disaster is caused by the twisting of the belt, so that the personal safety is greatly threatened. The fault diagnosis of cold rolling broken belts can effectively prevent accidents, inhibit product quality from decreasing, and furthest exert flow operation potential, thus having important scientific significance.
The fault diagnosis method based on data driving is a common method in the field of fault diagnosis, and the data quality has a great influence on the accuracy of the method. In diagnosis of broken belt faults, a plurality of factors influencing broken belt lead to high data dimension, main features are difficult to extract in diagnosis, and model training speed is low. In addition, the normal running state data of the rolled piece in cold rolling is better obtained, but compared with normal running, the frequency of faults is not high, so that the fault data is relatively lack, and the fault data becomes an important factor for restricting the diagnosis and research of broken belt faults based on data driving.
Disclosure of Invention
The embodiment of the invention provides a data enhancement method applied to fault diagnosis of a broken belt of a cold rolling mill, which can improve the quality of a generated fault image while improving the training speed of a generated model so as to be convenient for directionally generating the fault image required by the fault diagnosis of the broken belt, thereby solving the problem of insufficient fault data in the fault diagnosis of the broken belt. The technical scheme is as follows:
in one aspect, a data enhancement method applied to cold rolling mill belt breakage fault diagnosis is provided, the method is applied to electronic equipment, and the method comprises the following steps:
collecting time sequence signals of a plurality of characteristics related to belt breakage fault diagnosis in cold rolling, and processing the collected time sequence signals of the plurality of characteristics to generate a two-dimensional fault image set;
dividing the fault image set into training data and test data;
training the auxiliary classification generation countermeasure network by using the training data and the corresponding labels to obtain a generation model, wherein the trained generation model is used for generating a fault image required by the diagnosis of the broken belt fault.
Further, the collecting the time sequence signals of the plurality of characteristics related to the diagnosis of the broken belt fault in the cold rolling, processing the collected time sequence signals of the plurality of characteristics, and generating the two-dimensional fault image set includes:
collecting time sequence signals of a plurality of characteristics related to belt breakage fault diagnosis in cold rolling;
reducing the dimension of the acquired time sequence signals of a plurality of characteristics through a stack self-coding network to obtain one-dimensional time sequence signals;
and generating a two-dimensional gray scale image from the one-dimensional time sequence signals through signal-image conversion to form a two-dimensional fault image set.
Further, the structure connection mode of the stack self-coding network is as follows: input layer-full connection layer.
Further, training the auxiliary classification generating countermeasure network by using the training data and the corresponding labels, and obtaining the generating model further includes:
turning, rotating and denoising the training data obtained by dividing to obtain training data of the auxiliary classification generation countermeasure network;
and inputting the training data of the obtained auxiliary classification generation countermeasure network and the corresponding label thereof into the auxiliary classification generation countermeasure network for training to obtain a generation model.
Further, the auxiliary classification generating an countermeasure network includes: a generator and a arbiter; wherein,
the generator is used for generating a fault image, and the fault image is a two-dimensional gray scale image;
the discriminator is used for judging the difference between the fault image generated by the generator and the fault image input to the auxiliary classification generation countermeasure network, and providing feedback for the generator.
Further, after training the auxiliary classification generation countermeasure network by using the training data and the corresponding labels thereof to obtain a generation model, the method further comprises:
generating a fault image required by diagnosis of the broken belt fault by using the generation model;
the generated fault image and training data obtained by original division are input into a two-dimensional convolutional neural network together for training, and a broken-band fault diagnosis model is obtained;
the trained belt breakage fault diagnosis model is used for carrying out belt breakage fault diagnosis and outputting belt breakage fault types.
Further, the structural connection mode of the two-dimensional convolutional neural network is as follows: two-dimensional convolution layer- & gt max pooling layer- & gt full connection layer- & gt SoftMax layer, wherein SoftMax represents a normalized exponential function.
Further, after the generated fault image and the training data obtained by the original division are input into the two-dimensional convolutional neural network together to be trained, the method further comprises the steps of:
and testing the trained broken belt fault diagnosis model by using the test data obtained by dividing.
In one aspect, an electronic device is provided, the electronic device including a processor and a memory, the memory storing at least one instruction, the at least one instruction being loaded and executed by the processor to implement the above-described data enhancement method applied to cold rolling mill belt breakage fault diagnosis.
In one aspect, a computer readable storage medium having stored therein at least one instruction loaded and executed by a processor to implement the above-described data enhancement method for cold rolling mill belt break fault diagnosis is provided.
The technical scheme provided by the embodiment of the invention has the beneficial effects that at least:
in the embodiment of the invention, time sequence signals of a plurality of characteristics related to the diagnosis of the broken belt fault in cold rolling are collected, and the collected time sequence signals of the plurality of characteristics are processed to generate a two-dimensional fault image set; dividing the fault image set into training data and test data; the training data and the corresponding labels are utilized to train the auxiliary classification generation countermeasure network, the training speed of the generated model is improved, and meanwhile, the quality of the generated fault image can be improved, so that the fault image required by the fault diagnosis of the broken belt can be generated in a directed mode, and the problem of insufficient fault data in the fault diagnosis of the broken belt is solved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a data enhancement method applied to cold rolling mill belt breakage fault diagnosis according to an embodiment of the present invention;
FIG. 2 is a detailed flow chart of a data enhancement method applied to the diagnosis of a broken belt fault of a cold rolling mill according to an embodiment of the present invention;
FIG. 3 is a graph showing the comparison of the loss function values before and after data enhancement according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of contrast between confusion matrices before and after data enhancement according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the embodiments of the present invention will be described in further detail with reference to the accompanying drawings.
As shown in fig. 1, an embodiment of the present invention provides a data enhancement method applied to a fault diagnosis of a broken belt of a cold rolling mill, where the method may be implemented by an electronic device, and the electronic device may be a terminal or a server, and the method includes:
s101, collecting time sequence signals of a plurality of characteristics related to belt breakage fault diagnosis in cold rolling, and processing the collected time sequence signals of the plurality of characteristics to generate a two-dimensional fault image set;
s102, dividing the fault image set into training data and test data;
and S103, training an auxiliary classification generation countermeasure network (auxiliary classifier generative adversarial networks, ACGANs) by using the training data and the corresponding labels thereof to obtain a generation model, wherein the trained generation model is used for generating a fault image required by the diagnosis of the broken belt fault.
The data enhancement method applied to the broken belt fault diagnosis of the cold rolling mill disclosed by the embodiment collects time sequence signals of a plurality of characteristics related to the broken belt fault diagnosis in cold rolling, processes the collected time sequence signals of the plurality of characteristics, and generates a two-dimensional fault image set; dividing the fault image set into training data and test data; the training data and the corresponding labels are utilized to train the auxiliary classification generation countermeasure network, the training speed of the generated model is improved, and meanwhile, the quality of the generated fault image can be improved, so that the fault image required by the fault diagnosis of the broken belt can be generated in a directed mode, and the problem of insufficient fault data in the fault diagnosis of the broken belt is solved.
In this embodiment, the label is a type of a broken belt fault, specifically referring to: broken belt of different frames is broken.
In this embodiment, the collecting the time sequence signals of the plurality of features related to the diagnosis of the belt breakage fault in the cold rolling, processing the collected time sequence signals of the plurality of features, and generating the two-dimensional fault image set includes:
a1, collecting time sequence signals of a plurality of characteristics related to diagnosis of belt breakage faults in cold rolling;
in this embodiment, as shown in fig. 2, the time series signals of 24 features related to the diagnosis of the belt breakage fault in the cold rolling are collected by N (for example, 24) sensors, so as to obtain a high-dimensional time series signal data matrix.
A2, reducing the dimension of the collected time sequence signals of a plurality of characteristics through a stacked auto-encoder network (SAE) to obtain one-dimensional time sequence signals;
in this embodiment, the obtained high-dimensional time sequence signal data matrix is put into a stack self-coding network (SAE) for training, and the high-dimensional time sequence signal data matrix is reduced to a one-dimensional time sequence signal, so as to achieve the effect of information fusion.
In this embodiment, the structure connection manner of the stack self-coding network (SAE) is as follows: input layer L0→full connection layer L1→full connection layer L2→full connection layer L3→full connection layer L4; wherein the number of output neurons of the input layer L0 is 30; the number of input neurons of the full-connection layer L1 is 30, and the number of output neurons is 10; the number of input neurons of the full-connection layer L2 is 10, and the number of output neurons is 1; the number of input neurons of the full-connection layer L3 is 1, and the number of output neurons is 10; the number of input neurons of the full connection layer L4 is 10, and the number of output neurons is 30.
A3, generating a two-dimensional gray scale image from the one-dimensional time sequence signals through signal-image conversion to form a two-dimensional fault image set.
In this embodiment, the specific manner of signal-image conversion is as shown in formula (1):
wherein P (m, n) represents the gray values of the m-th row and the n-th column in the generated two-dimensional gray map; n represents that the size of the generated image is n×n; form L (i) represents the gray value of the ith data point in L; max (L) and Min (L) respectively represent maximum and minimum values in L, L represents one-dimensional time sequence signal value after single sampling, and the length of the signal is N 2 The method comprises the steps of carrying out a first treatment on the surface of the The rounding function round (x) functions to round the data to ensure that the converted data takes an integer between 0 and 255.
In this embodiment, the one-dimensional time domain signal may be normalized, converted into a gray value, rounded, and subjected to signal interception and matrix transformation according to the image size by the formula (1) to obtain a two-dimensional gray map of the one-dimensional time sequence signal, thereby forming a two-dimensional fault image set.
In this embodiment, 80% of the failure image set constituting the two dimensions may be used as training data, and the remaining 20% may be used as test data.
In this embodiment, training the auxiliary classification generating countermeasure network by using the training data and the corresponding label thereof, and obtaining the generating model further includes:
b1, turning, rotating and noise adding are carried out on the training data obtained by division, and training data of an countermeasure network is generated by auxiliary classification;
in this embodiment, flipping includes horizontally flipping and/or vertically flipping the image.
In this embodiment, the rotation includes rotating the image 90 ° left or right.
In this embodiment, the noise adding process refers to adding random gaussian noise to the image, and adding gaussian noise specifically refers to directly adding the image matrix to the numbers randomly sampled from the gaussian distribution.
In this embodiment, each image in the training data is subjected to the above-mentioned processes of turning, rotation and noise adding, so that the training data with six times of the number of the original training data can be obtained, and the training data of the countermeasure network can be generated by obtaining the auxiliary classification with higher diversity.
And B2, inputting the obtained training data of the auxiliary classification generation countermeasure network and the corresponding label thereof into the auxiliary classification generation countermeasure network (ACGANs) for training to obtain a generation model.
In this embodiment, the Auxiliary Class Generation Antagonism Network (ACGANs) can process two-dimensional images, and thus, the ACGANs may also be referred to as 2D-ACGANs, as shown in fig. 2.
In this embodiment, the Auxiliary Classification Generating Antagonistic Networks (ACGANs) includes: the system comprises a generator and a discriminator, wherein the generator is used for generating a fault image, the discriminator is used for judging the difference between the fault image generated by the generator and the fault image input to the auxiliary classification generation countermeasure network, and providing feedback for the generator, and the fault image is a two-dimensional gray scale image. The generator includes 4 fractional step two-dimensional convolutional layers and 4 batch normalization layers, all using the ReLU function as an activation function except for the last output layer which uses the Tanh function as an activation function. The discriminator comprises 4 two-dimensional convolution layers and 4 batch normalization layers, and the activation function is used by the LeakyReLU function except the Sigmoid function used by the output layer.
In this embodiment, the two-dimensional convolution layer has fewer parameters and better capability of extracting features from the timing signal, and meanwhile, the ACGANs described in this embodiment considers the tag information of the fault data at the same time, and does not need to train multiple models, so that the training process is simple, and thus, the quality of the generated fault image can be improved while the training speed of the generated model is improved, so that the fault image required by the fault diagnosis of the broken belt can be generated in an oriented manner, and the problem of insufficient fault data in the fault diagnosis of the broken belt is solved.
In this embodiment, after training the auxiliary classification generating countermeasure network by using the training data and the corresponding label thereof to obtain the generating model, the method further includes:
generating a fault image required by diagnosis of the broken belt fault by using the generation model;
the generated fault image and the training data obtained by original division are input into a two-dimensional (2D) convolutional neural network (convolutional neural network, CNN) together for training, and a broken-band fault diagnosis model is obtained;
the trained belt breakage fault diagnosis model is used for carrying out belt breakage fault diagnosis and outputting belt breakage fault types.
In this embodiment, the structure of the two-dimensional convolutional neural network includes 11 layers in total, and a specific structure connection mode is as follows: two-dimensional convolution layer L1 (5×5×32) →maximum pooling layer L2 (2×2) →two-dimensional convolution layer L3 (3×3×64) →maximum pooling layer L4 (2×2) →two-dimensional convolution layer L5 (3×3×128) →maximum pooling layer L6 (2×2) →two-dimensional convolution layer L7 (3×3×256) →maximum pooling layer L8 (2×2) →full connection layer L9 (2560-768) →full connection layer L10 (768-10) →softMax (normalized exponential function) layer L11; wherein, three parameters in brackets of the two-dimensional convolution layer respectively represent the length of the convolution kernel, the width of the convolution layer and the number of the convolution kernels; the parameters in the max pooling layer brackets represent the size of their windows; the parameters in the full connection brackets represent the number of input parameters and the number of output parameters, respectively.
In this embodiment, after the generated fault image and the training data obtained by the original division are input together into the two-dimensional convolutional neural network to perform training, and the fault diagnosis model is obtained, the method further includes:
and testing the trained broken belt fault diagnosis model by using the test data obtained by dividing.
In order to verify the effectiveness of the data enhancement method applied to the fault diagnosis of the broken belt of the cold rolling mill, the relevant characteristic parameters of the fault of the broken belt of the cold rolling mill and the broken belt of the mill in a certain steel mill are collected, the specific composition of the relevant characteristic parameters is shown in a table 1, and a fault diagnosis experiment is carried out through the relevant characteristic parameters. In the experiment, fault data and normal data of 4 rolling mills are collected, 24 parameters are obtained in total, sampling is carried out at a sampling frequency of 12kHz, 4096 points are taken as a group, and 1000 groups of data are collected.
TABLE 1 characterization parameters relating to diagnosis of strip breakage failure in cold rolling
| 1 | 1 machine frame transmission side servo valve current |
| 2 | 1 machine frame operation side servo valve current |
| 3 | Deviation of rolling force (absolute value) |
| 4 | 2 deviation of rolling force of stand |
| 5 | 2 servo valve current at the transmission side of the frame |
| 6 | 2 machine frame operation side servo valve current |
| 7 | 3 servo valve current at transmission side of frame |
| 8 | 3 servo valve current at the operating side of the frame |
| 9 | 3 deviation of rolling force |
| 10 | 4 servo valve current at transmission side of frame |
| 11 | 4 servo valve current at operating side of stand |
| 12 | 4 deviation of rolling force |
| 13 | 1 actual tension value of the frame |
| 14 | Tension deviation value of 1 machine frame |
| 15 | 2 actual tension value of the frame |
| 16 | Tension deviation value of 2 machine frame |
| 17 | 3 actual tension value of the frame |
| 18 | Tension deviation value of 3 machine frame |
| 19 | Actual tension value of the frame |
| 20 | Tension deviation value of 4 machine frame |
| 21 | 1 motor current of the frame |
| 22 | Motor current of 2 frames |
| 23 | 3 motor current of frame |
| 24 | Motor current for 4 rack |
FIG. 3 (a) shows the variation trend of the loss function of the broken-band fault diagnosis model under no data enhancement, and the training loss function can be seen to be approximately converged to 0 finally, while the test loss function always oscillates about 0.5, which indicates that the model has an overfitting phenomenon due to insufficient training data; fig. 3 (b) shows a change trend of a loss function of the broken-band fault diagnosis model after the data enhancement method provided by the embodiment of the present invention, where the loss function value is approximately converged to 0 on both training data and test data, which indicates that the overfitting phenomenon substantially disappears after the data enhancement.
Fig. 4 (a) and (b) are respectively a comparison schematic diagram of confusion matrix for diagnosing fault of broken belt before and after data enhancement, wherein the data on the symmetry axis represents the proportion of the correctly identified number of the health state to all test data, and the rest data of each row is the proportion of the incorrectly identified number to other health states. It can be clearly seen that after data enhancement, the identification accuracy of the 3 frames is improved from 92.5% to 99%, 6.5% and the total average accuracy is also improved from 95% to 99.5%.
Fig. 5 is a schematic structural diagram of an electronic device 600 according to an embodiment of the present invention, where the electronic device 600 may have a relatively large difference due to different configurations or performances, and may include one or more processors (central processing units, CPU) 601 and one or more memories 602, where at least one instruction is stored in the memories 602, and the at least one instruction is loaded and executed by the processors 601 to implement the above-mentioned data enhancement method applied to the diagnosis of a strip breakage fault of a cold rolling mill.
In an exemplary embodiment, a computer readable storage medium, such as a memory comprising instructions executable by a processor in a terminal to perform the above-described data enhancement method applied to cold rolling mill belt break fault diagnosis is also provided. For example, the computer readable storage medium may be ROM, random Access Memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program for instructing relevant hardware, where the program may be stored in a computer readable storage medium, and the storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
The foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and scope of the invention are intended to be included within the scope of the invention.
Claims (6)
1. A data enhancement method applied to fault diagnosis of a broken belt of a cold rolling mill, comprising:
collecting time sequence signals of a plurality of characteristics related to belt breakage fault diagnosis in cold rolling, and processing the collected time sequence signals of the plurality of characteristics to generate a two-dimensional fault image set;
dividing the fault image set into training data and test data;
training an auxiliary classification generation countermeasure network by using the training data and the corresponding labels thereof to obtain a generation model, wherein the trained generation model is used for generating a fault image required by the diagnosis of the broken belt fault;
the method for generating the two-dimensional fault image set includes the steps of collecting time sequence signals of a plurality of characteristics related to diagnosis of the strip breakage fault in cold rolling, processing the collected time sequence signals of the plurality of characteristics, and generating the two-dimensional fault image set:
collecting time sequence signals of a plurality of characteristics related to belt breakage fault diagnosis in cold rolling;
reducing the dimension of the acquired time sequence signals of a plurality of characteristics through a stack self-coding network to obtain one-dimensional time sequence signals;
generating a two-dimensional gray scale image from the one-dimensional time sequence signal through signal-image conversion to form a two-dimensional fault image set;
the specific way of signal-image conversion is:
wherein P (m, n) represents the gray values of the m-th row and the n-th column in the generated two-dimensional gray map; n represents that the size of the generated image is n×n; form L (i) represents the gray value of the ith data point in L; max (L) and Min (L) respectively represent maximum and minimum values in L, L represents one-dimensional time sequence signal value after single sampling, and the length of the signal is N 2 The method comprises the steps of carrying out a first treatment on the surface of the Rounding function round (x) is used for rounding data to ensure that the value of the converted data is an integer between 0 and 255; the method comprises the steps of carrying out signal normalization on a one-dimensional time domain signal through signal-image conversion, converting the signal into a gray value, rounding, carrying out signal interception and matrix transformation according to the size of an image so as to obtain a two-dimensional gray level image of a one-dimensional time sequence signal, and forming a two-dimensional fault image set;
the structure connection mode of the stack self-coding network is as follows: input layer-full connection layer.
2. The data enhancement method applied to cold rolling mill belt breakage fault diagnosis according to claim 1, wherein training the auxiliary classification generation countermeasure network by using the training data and the corresponding label thereof, and obtaining the generation model further comprises:
turning, rotating and denoising the training data obtained by dividing to obtain training data of the auxiliary classification generation countermeasure network;
and inputting the training data of the obtained auxiliary classification generation countermeasure network and the corresponding label thereof into the auxiliary classification generation countermeasure network for training to obtain a generation model.
3. The data enhancement method applied to cold rolling mill belt break fault diagnosis according to claim 1, wherein the auxiliary classification generation countermeasure network comprises: a generator and a arbiter; wherein,
the generator is used for generating a fault image, and the fault image is a two-dimensional gray scale image;
the discriminator is used for judging the difference between the fault image generated by the generator and the fault image input to the auxiliary classification generation countermeasure network, and providing feedback for the generator.
4. The data enhancement method applied to cold rolling mill belt break fault diagnosis according to claim 1, wherein after training the auxiliary classification generation countermeasure network by using the training data and the corresponding label thereof, the method further comprises:
generating a fault image required by diagnosis of the broken belt fault by using the generation model;
the generated fault image and training data obtained by original division are input into a two-dimensional convolutional neural network together for training, and a broken-band fault diagnosis model is obtained;
the trained belt breakage fault diagnosis model is used for carrying out belt breakage fault diagnosis and outputting belt breakage fault types.
5. The data enhancement method applied to cold rolling mill broken belt fault diagnosis according to claim 4, wherein the structural connection mode of the two-dimensional convolutional neural network is as follows: two-dimensional convolution layer- & gt max pooling layer- & gt full connection layer- & gt SoftMax layer, wherein SoftMax represents a normalized exponential function.
6. The data enhancement method applied to the fault diagnosis of the broken belt of the cold rolling mill according to claim 4, wherein after the generated fault image and the training data obtained by the original division are input into the two-dimensional convolutional neural network together for training, the method further comprises:
and testing the trained broken belt fault diagnosis model by using the test data obtained by dividing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010578466.5A CN111832428B (en) | 2020-06-23 | 2020-06-23 | Data enhancement method applied to cold rolling mill broken belt fault diagnosis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010578466.5A CN111832428B (en) | 2020-06-23 | 2020-06-23 | Data enhancement method applied to cold rolling mill broken belt fault diagnosis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111832428A CN111832428A (en) | 2020-10-27 |
| CN111832428B true CN111832428B (en) | 2024-02-23 |
Family
ID=72897841
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010578466.5A Active CN111832428B (en) | 2020-06-23 | 2020-06-23 | Data enhancement method applied to cold rolling mill broken belt fault diagnosis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111832428B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112328588B (en) * | 2020-11-27 | 2022-07-15 | 哈尔滨工程大学 | Industrial fault diagnosis unbalanced time sequence data expansion method |
| CN113533945A (en) * | 2021-06-30 | 2021-10-22 | 桂林电子科技大学 | Analog circuit fault diagnosis method based on two-dimensional convolutional neural network |
Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102637020A (en) * | 2011-02-14 | 2012-08-15 | 武汉科技大学 | Integrated-type control system of roll steel heating furnace |
| JP2015021756A (en) * | 2013-07-16 | 2015-02-02 | 株式会社キーエンス | Three-dimensional image processor, three-dimensional image processing method, three-dimensional image processing program, computer-readable recording medium, and recording apparatus |
| GB201803083D0 (en) * | 2018-02-26 | 2018-04-11 | Nokia Technologies Oy | Artificial Neural Networks |
| WO2018082540A1 (en) * | 2016-11-07 | 2018-05-11 | 广州广电运通金融电子股份有限公司 | Bank note one-dimensional signal detection method and device |
| CN109580215A (en) * | 2018-11-30 | 2019-04-05 | 湖南科技大学 | A kind of wind-powered electricity generation driving unit fault diagnostic method generating confrontation network based on depth |
| CN109675935A (en) * | 2019-03-06 | 2019-04-26 | 北京科技大学 | A kind of IPCA operation of rolling on-line fault diagnosis method becoming control limit |
| CN109765053A (en) * | 2019-01-22 | 2019-05-17 | 中国人民解放军海军工程大学 | Utilize the Fault Diagnosis of Roller Bearings of convolutional neural networks and kurtosis index |
| CN109918999A (en) * | 2019-01-22 | 2019-06-21 | 西安交通大学 | Based on the mechanical equipment fault intelligent diagnosing method for generating model under a kind of Small Sample Database |
| CN109948693A (en) * | 2019-03-18 | 2019-06-28 | 西安电子科技大学 | Expand and generate confrontation network hyperspectral image classification method based on super-pixel sample |
| EP3509285A1 (en) * | 2018-01-05 | 2019-07-10 | Datamax-O'Neil Corporation | Methods, apparatuses, and systems for detecting printing defects and contaminated components of a printer |
| CN110068462A (en) * | 2019-05-14 | 2019-07-30 | 北京科技大学 | A kind of motor bearings method for diagnosing faults and device |
| CN110298397A (en) * | 2019-06-25 | 2019-10-01 | 东北大学 | The multi-tag classification method of heating metal image based on compression convolutional neural networks |
| WO2019241155A1 (en) * | 2018-06-11 | 2019-12-19 | Arterys Inc. | Simulating abnormalities in medical images with generative adversarial networks |
| CN110598530A (en) * | 2019-07-30 | 2019-12-20 | 浙江工业大学 | Small sample radio signal enhanced identification method based on ACGAN |
| KR20190141326A (en) * | 2018-06-14 | 2019-12-24 | 한국과학기술원 | Method and Apparatus for ECG Arrhythmia Classification using a Deep Convolutional Neural Network |
| CN110823574A (en) * | 2019-09-30 | 2020-02-21 | 安徽富煌科技股份有限公司 | Fault diagnosis method based on semi-supervised learning deep countermeasure network |
| CN110849626A (en) * | 2019-11-18 | 2020-02-28 | 东南大学 | Self-adaptive sparse compression self-coding rolling bearing fault diagnosis system |
| CN111076934A (en) * | 2019-12-24 | 2020-04-28 | 江苏大学 | Method for diagnosing potential fault of bearing based on S transformation |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5706647B2 (en) * | 2010-09-03 | 2015-04-22 | キヤノン株式会社 | Information processing apparatus and processing method thereof |
-
2020
- 2020-06-23 CN CN202010578466.5A patent/CN111832428B/en active Active
Patent Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102637020A (en) * | 2011-02-14 | 2012-08-15 | 武汉科技大学 | Integrated-type control system of roll steel heating furnace |
| JP2015021756A (en) * | 2013-07-16 | 2015-02-02 | 株式会社キーエンス | Three-dimensional image processor, three-dimensional image processing method, three-dimensional image processing program, computer-readable recording medium, and recording apparatus |
| WO2018082540A1 (en) * | 2016-11-07 | 2018-05-11 | 广州广电运通金融电子股份有限公司 | Bank note one-dimensional signal detection method and device |
| EP3509285A1 (en) * | 2018-01-05 | 2019-07-10 | Datamax-O'Neil Corporation | Methods, apparatuses, and systems for detecting printing defects and contaminated components of a printer |
| GB201803083D0 (en) * | 2018-02-26 | 2018-04-11 | Nokia Technologies Oy | Artificial Neural Networks |
| WO2019241155A1 (en) * | 2018-06-11 | 2019-12-19 | Arterys Inc. | Simulating abnormalities in medical images with generative adversarial networks |
| KR20190141326A (en) * | 2018-06-14 | 2019-12-24 | 한국과학기술원 | Method and Apparatus for ECG Arrhythmia Classification using a Deep Convolutional Neural Network |
| CN109580215A (en) * | 2018-11-30 | 2019-04-05 | 湖南科技大学 | A kind of wind-powered electricity generation driving unit fault diagnostic method generating confrontation network based on depth |
| CN109918999A (en) * | 2019-01-22 | 2019-06-21 | 西安交通大学 | Based on the mechanical equipment fault intelligent diagnosing method for generating model under a kind of Small Sample Database |
| CN109765053A (en) * | 2019-01-22 | 2019-05-17 | 中国人民解放军海军工程大学 | Utilize the Fault Diagnosis of Roller Bearings of convolutional neural networks and kurtosis index |
| CN109675935A (en) * | 2019-03-06 | 2019-04-26 | 北京科技大学 | A kind of IPCA operation of rolling on-line fault diagnosis method becoming control limit |
| CN109948693A (en) * | 2019-03-18 | 2019-06-28 | 西安电子科技大学 | Expand and generate confrontation network hyperspectral image classification method based on super-pixel sample |
| CN110068462A (en) * | 2019-05-14 | 2019-07-30 | 北京科技大学 | A kind of motor bearings method for diagnosing faults and device |
| CN110298397A (en) * | 2019-06-25 | 2019-10-01 | 东北大学 | The multi-tag classification method of heating metal image based on compression convolutional neural networks |
| CN110598530A (en) * | 2019-07-30 | 2019-12-20 | 浙江工业大学 | Small sample radio signal enhanced identification method based on ACGAN |
| CN110823574A (en) * | 2019-09-30 | 2020-02-21 | 安徽富煌科技股份有限公司 | Fault diagnosis method based on semi-supervised learning deep countermeasure network |
| CN110849626A (en) * | 2019-11-18 | 2020-02-28 | 东南大学 | Self-adaptive sparse compression self-coding rolling bearing fault diagnosis system |
| CN111076934A (en) * | 2019-12-24 | 2020-04-28 | 江苏大学 | Method for diagnosing potential fault of bearing based on S transformation |
Non-Patent Citations (5)
| Title |
|---|
| Fault Diagnosis of Aero-engine Bearing Using a Stacked Auto-Encoder Network;XueSen Lin 等;《 2018 IEEE 4th Information Technology and Mechatronics Engineering Conference (ITOEC)》;545-548 * |
| Multiobjective Coordinated Control Strategy for Grid-Connected Inverter under Unbalanced Voltage Conditions;Miao, L 等;《JOURNAL OF ENERGY ENGINEERING》;第146卷(第3期);1-7 * |
| 基于ACGAN的功率变换器参数性故障诊断方法;傅宏辉 等;《机械制造与自动化》;第48卷(第6期);159-163 * |
| 基于SAE的航空发动机气路故障诊断;杨洪富 等;《 航空计算技术 》;第48卷(第6期);47-51 * |
| 基于二维灰度图的数据增强方法在电机轴承故障诊断的应用研究;肖雄 等;《中国电机工程学报》;第41卷(第2期);738-748 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111832428A (en) | 2020-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Xiao et al. | Domain adaptive motor fault diagnosis using deep transfer learning | |
| CN111832428B (en) | Data enhancement method applied to cold rolling mill broken belt fault diagnosis | |
| Zhang et al. | Imbalanced fault diagnosis of rolling bearing using enhanced generative adversarial networks | |
| Zhang et al. | A multi-module generative adversarial network augmented with adaptive decoupling strategy for intelligent fault diagnosis of machines with small sample | |
| Kaya et al. | A new automatic bearing fault size diagnosis using time-frequency images of CWT and deep transfer learning methods | |
| CN112380767B (en) | Equipment fault diagnosis method and system based on improved generation countermeasure network | |
| CN111967573A (en) | Data processing method, device, equipment and computer readable storage medium | |
| CN115908842A (en) | Transformer partial discharge data enhancement and identification method | |
| CN115753101A (en) | Bearing fault diagnosis method based on weight adaptive feature fusion | |
| CN112747924A (en) | Bearing life prediction method based on attention mechanism and residual error neural network | |
| CN112464990A (en) | Method and device for sensing vibration data based on current and voltage sensor | |
| CN113221946B (en) | Method for diagnosing fault types of mechanical equipment | |
| CN112380932B (en) | Vibration signal characteristic value selection method and elevator health state evaluation or fault diagnosis method | |
| Liu et al. | Densely connected semi-Bayesian network for machinery fault diagnosis with non-ideal data | |
| CN114544172A (en) | Rotary machine fault diagnosis method, system and medium based on deep confidence network | |
| CN112365335B (en) | Method and device for processing credit data | |
| CN112598026A (en) | Continuous annealing production process fault diagnosis method based on unbalanced data deep learning | |
| CN112686367A (en) | Novel normalization mechanism | |
| CN112560894A (en) | Improved 3D convolutional network hyperspectral remote sensing image classification method and device | |
| CN114202110B (en) | Service fault prediction method and device based on RF-XGBOOST | |
| CN115546862A (en) | Expression recognition method and system based on cross-scale local difference depth subspace characteristics | |
| CN113671923A (en) | Continuous rolling process fault detection method based on data driving | |
| CN115638965A (en) | Rotating machine fault diagnosis method based on one-dimensional attention mechanism and convolutional neural network | |
| Li et al. | Fault Diagnosis of Rolling Bearing Based on Improved LeNet-5 CNN | |
| CN113887300A (en) | Method and device for detecting target, human face and human face key point and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |