CN112926642B - Multi-source domain self-adaptive intelligent mechanical fault diagnosis method and system - Google Patents

Multi-source domain self-adaptive intelligent mechanical fault diagnosis method and system Download PDF

Info

Publication number
CN112926642B
CN112926642B CN202110196455.5A CN202110196455A CN112926642B CN 112926642 B CN112926642 B CN 112926642B CN 202110196455 A CN202110196455 A CN 202110196455A CN 112926642 B CN112926642 B CN 112926642B
Authority
CN
China
Prior art keywords
source domain
data
domain
fault diagnosis
source
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
Application number
CN202110196455.5A
Other languages
Chinese (zh)
Other versions
CN112926642A (en
Inventor
李沂滨
徐丹雅
宋艳
高晟耀
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shandong University
Original Assignee
Shandong University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shandong University filed Critical Shandong University
Priority to CN202110196455.5A priority Critical patent/CN112926642B/en
Publication of CN112926642A publication Critical patent/CN112926642A/en
Application granted granted Critical
Publication of CN112926642B publication Critical patent/CN112926642B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • 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
    • 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
    • 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
    • Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computational Linguistics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Evolutionary Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

The utility model provides a multisource domain self-adaptive intelligent mechanical fault diagnosis method and a system, which are used for acquiring the running state data of mechanical equipment; extracting the characteristics of the acquired running state data by using a preset characteristic extractor; inputting the extracted features into a preset classifier to obtain a fault classification prediction result; the network parameters of the feature extractor are updated according to the classification loss of a preset multi-source domain discriminator, the maximum mean difference of a source domain data set and a target domain data set and the classification loss obtained by the classifier according to the source domain data set; the method and the device can effectively solve the problem that the training set learning model cannot adapt to the test set due to different feature distributions between the training set and the test set data under different operating conditions, and effectively improve the fault diagnosis classification accuracy of the mechanical equipment under new working conditions.

Description

Multi-source domain self-adaptive intelligent mechanical fault diagnosis method and system
Technical Field
The disclosure relates to the technical field of fault diagnosis of mechanical equipment, in particular to a multi-source domain self-adaptive intelligent mechanical fault diagnosis method and system.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
With the development of mechanical equipment towards automation and intellectualization, health state monitoring and maintenance are particularly important for guaranteeing high-precision and high-reliability operation of a mechanical system. Once a mechanical device fails, a catastrophic result will occur, and timely diagnosis of the failure is particularly important to ensure high-precision and high-reliability operation of the mechanical device.
The purpose of fault diagnosis is to detect a fault on a faulty system by monitoring and analyzing the state of the machine using the obtained measurement values. The basic method of fault diagnosis is to arrange sensors around a tested component, monitor vibration signals and detect by using a fault diagnosis algorithm, and the key link is to extract fault characteristics, wherein the quality of the extracted characteristics has important influence on the final fault diagnosis result. The traditional fault diagnosis method usually adopts manual feature extraction and requires an operator to have a certain professional knowledge background.
The rapid development of the internet of things enables the data scale of the operation state of a mechanical system to be continuously enlarged, and meanwhile, the progress of the machine learning and deep learning theory promotes the rapid development of an intelligent fault diagnosis algorithm based on data driving. On the one hand, existing methods are mostly based on the assumption that the training set and the test set have the same data distribution, however, when acquiring vibration data, the complexity of the industrial environment and the change of the working environment may change the distribution of the actual mechanical data and the training data, and the network trained by using the training set is not suitable for the actual fault diagnosis task. On the other hand, most fault diagnosis algorithms assume that enough labeled data are available for training a model, and in an actual engineering application scene, a machine usually works in a healthy state and rarely breaks down; meanwhile, the acquisition of label data costs a lot, the failure data is insufficient, and no label is generated, so that the existing method is not applicable and a reliable diagnosis model is not trained.
The domain self-adaptation is a method for solving the problem, and based on the idea of knowledge reuse, various fault conditions of mechanical equipment can be simulated in a laboratory, sufficient marking data are collected, and a diagnosis model trained by using bearing data of the laboratory can be suitable for bearing fault diagnosis in an engineering scene. Some existing methods mostly adopt a single-source domain adaptive method, that is, a single data source is used as a training set (source domain), and in an actual application scenario, there may be a situation that multiple data sources are available. In practical application, a test set (target domain) is not labeled, and it cannot be determined which data source is used for performing domain adaptation better. One approach is to combine all data sources into one data set as a training set, but this ignores the differences between data sets.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a multi-source domain self-adaptive intelligent mechanical fault diagnosis method and system, which can effectively solve the problem that a training set learning model cannot adapt to a test set due to different feature distributions among training set data and test set data under different operating conditions, and effectively improve the fault diagnosis classification accuracy of mechanical equipment under a new working condition.
In order to achieve the purpose, the following technical scheme is adopted in the disclosure:
the first aspect of the disclosure provides a multi-source domain self-adaptive intelligent mechanical fault diagnosis method.
A multi-source domain self-adaptive intelligent mechanical fault diagnosis method comprises the following steps:
acquiring running state data of mechanical equipment;
extracting the characteristics of the acquired running state data by using a preset characteristic extractor;
inputting the extracted features into a preset classifier to obtain a fault classification prediction result;
and updating the network parameters of the feature extractor according to the classification loss of a preset multi-source domain discriminator, the maximum mean difference of the source domain data set and the target domain data set and the classification loss obtained by the classifier according to the source domain data set.
A second aspect of the disclosure provides a multi-source domain adaptive intelligent mechanical fault diagnosis system.
A multi-source domain adaptive intelligent mechanical fault diagnosis system comprises:
a data acquisition module configured to: acquiring running state data of mechanical equipment;
a feature extraction module configured to: extracting the characteristics of the acquired running state data by using a preset characteristic extractor;
a classification prediction module configured to: inputting the extracted features into a preset classifier to obtain a fault classification prediction result;
and updating the network parameters of the feature extractor according to the classification loss of a preset multi-source domain discriminator, the maximum mean difference of the source domain data set and the target domain data set and the classification loss obtained by the classifier according to the source domain data set.
A third aspect of the present disclosure provides a computer-readable storage medium, on which a program is stored, which when executed by a processor implements the steps in the multi-source domain adaptive intelligent mechanical fault diagnosis method according to the first aspect of the present disclosure.
A fourth aspect of the present disclosure provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, where the processor executes the program to implement the steps in the multi-source-domain adaptive intelligent mechanical fault diagnosis method according to the first aspect of the present disclosure.
Compared with the prior art, the beneficial effect of this disclosure is:
1. according to the method, the system, the medium or the electronic equipment, the fault diagnosis characteristics are learned by using the plurality of data sources, the domain characteristic expression is learned through the characteristic extractor, the distribution between each source domain data and the target domain data is reduced by using the training mode based on confrontation and distance, the method is applied to the target domain for fault diagnosis, the problem that a training set learning model cannot adapt to a test set due to different characteristic distributions between the training set data and the test set data under different operating conditions can be effectively solved, the fault diagnosis classification accuracy of the mechanical equipment under a new working condition is effectively improved, and compared with the prior art, the fault diagnosis accuracy under different working conditions is remarkably improved.
2. The method, the system, the medium or the electronic equipment disclosed by the disclosure consider the problem that the distribution of the acquired data is different from that of the existing data set under the actual condition, simultaneously solve the problem of how to select the source domain when a plurality of source domains exist, can extract the domain invariant features by utilizing a plurality of data sources so as to carry out effective fault classification, and improve the classification accuracy.
3. According to the method, the system, the medium or the electronic equipment, the feature extractor based on the one-dimensional convolutional neural network can effectively extract feature information from a mechanical vibration sequence, the distribution difference of the extracted features of the source domain and the target domain can be reduced by combining with the multi-source domain discriminator through continuous iterative training, the domain invariant features between each source domain and each target domain are obtained, and effective feature classification is further performed.
Advantages of additional aspects of the disclosure will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the disclosure.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure and are not to limit the disclosure.
Fig. 1 is a schematic structural diagram of a fault diagnosis prediction model provided in embodiment 1 of the present disclosure.
Detailed Description
The present disclosure is further illustrated by the following examples in conjunction with the accompanying drawings.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present disclosure. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
The embodiments and features of the embodiments in the present disclosure may be combined with each other without conflict.
Example 1:
the embodiment 1 of the disclosure provides a multi-source domain self-adaptive intelligent mechanical fault diagnosis method, which comprises the following steps:
acquiring running state data of mechanical equipment;
extracting the characteristics of the acquired running state data by using a preset characteristic extractor;
inputting the extracted features into a preset classifier to obtain a fault classification prediction result;
and updating the network parameters of the feature extractor according to the classification loss of a preset multi-source domain discriminator, the maximum mean difference of the source domain data set and the target domain data set and the classification loss obtained by the classifier according to the source domain data set.
Specifically, the method comprises the following steps:
s1: and establishing a source domain data set of the fault diagnosis algorithm by utilizing historical fault information under different working conditions.
Where N source domain datasets are denoted S1,S2,...,SNAnd the source domain data set is tagged.
Dividing vibration data into samples with signal size of (2048, 1), labeling different fault states, transposing and normalizing the data into data with the data of 0-1, wherein a normalization formula is as follows:
Figure BDA0002946900600000051
wherein x iss×1Representing the converted input data, s represents the magnitude of the signal, 2048 in this embodiment, mmaxAnd mminRespectively represent xs×1Maximum and minimum values of (a).
The data distribution of the source domain is defined as
Figure BDA0002946900600000061
Represents data from a source domain j, where
Figure BDA0002946900600000062
The data is represented by a representation of,
Figure BDA0002946900600000063
to represent
Figure BDA0002946900600000068
A corresponding label.
The target domain data set is a data set to be predicted, the data processing mode is the same as that of the source domain data, but the target domain is not labeled, the samples do not need to be labeled, and the data distribution is represented as Dt(x, y), data
Figure BDA0002946900600000064
Is the target domain data without label.
Suppose that:the source domain and the target domain share a feature space and have different data distributions (
Figure BDA0002946900600000065
D(Xi)≠D(Xj)，i≠j，i，j∈N；D(Xi)≠D(Xt) I ∈ N). In most cases, the source domain samples are sufficient in size to build an effective and accurate classifier.
S2: and constructing a feature extractor based on a one-dimensional convolutional neural network.
The feature extractor is composed of four one-dimensional convolutional layers, and each convolutional layer comprises a one-dimensional batch standardization and ReLU activation layer and a maximum pooling layer. And finally, expanding the data of the fourth layer, and obtaining the output F (x) of the feature extractor through a full connection layer, a ReLU activation function and a Dropout layer, wherein the output of the one-dimensional convolution of each layer is as follows:
fi＝E(w*uj+b) (2)
wherein f isiFor the output of the i-th layer one-dimensional convolution, E represents an activation function, and the parameter w belongs to RkAnd b ∈ R is the parameter of the convolution neural network, k denotes the convolution kernel size, uj∈R1×kFor source domain data XsAnd "", indicates a convolution operation.
S3: and constructing a multi-source domain discriminator based on confrontation and distance.
The part consists of two modules, namely a discriminator and a maximum mean difference module.
Establishing N domain discriminators for N source domains
Figure BDA0002946900600000066
For data x from the ith source domain or target domain, obtaining characteristics F (x) after passing through a characteristic extractor, and the ith discriminator
Figure BDA0002946900600000067
Accepting F (x), distinguishing whether the feature F (x) is from the source domain i or the target domain; discriminator
Figure BDA0002946900600000071
Only data from the source domain i and data from the target domain are accepted, and all discriminators accept data from the target domain.
The classification loss of the discriminator is calculated as follows:
Figure BDA0002946900600000072
where E represents the cross entropy loss.
To further reduce feature variation and data distribution variation, the method comprises
Figure BDA0002946900600000073
And
Figure BDA0002946900600000074
as from a data set
Figure BDA0002946900600000075
And P (X)t) The maximum mean difference equation for measuring the difference in the distribution of the two data sets is as follows:
Figure BDA0002946900600000076
wherein P and Q represent
Figure BDA0002946900600000077
And XtF denotes the mapping of the reconstructed kernel Hilbert space, HkThe representation represents hilbert space with kernel k, where the present network uses a gaussian kernel, and H is a function that represents that it can distinguish between any two data distributions, two distributions being considered identical if the value of the maximum mean difference of the two partitions is small enough, otherwise they are considered different.
S4: and constructing a classifier C to obtain an intelligent fault diagnosis model, as shown in FIG. 1, and training a network by using a source domain and a target domain.
The source domain data with the labels is used for training a classifier, the features F (x) obtained after the source domain data passes through a feature extractor are used as the input of the classifier, and the classification loss of the classifier is calculated as follows:
Figure BDA0002946900600000078
where E represents the cross entropy loss.
Through repeated iterative confrontation training, the distribution difference between data is continuously reduced, and parameters are modified, so that the classification loss is minimized.
S5: and based on the trained network, carrying out fault classification on the vibration data of the target domain, inputting the data of the target domain into the trained network, wherein the output of the classifier is the prediction result.
Specific examples are as follows:
the multi-source domain self-adaptive intelligent fault vibration method described in the embodiment is described in detail by taking a bearing fault vibration data set of the university of Keyssierra as an example.
The Kessi bearing fault data set is divided into four different working conditions (the loads are respectively 0HP, 1HP, 2HP and 3HP), and for example, three of the working conditions (0HP, 1HP and 2HP) are used as source data sets, and one (3HP) is used as a target data set, and the fault condition of the target domain data set (one working condition) is predicted by using the source data sets (three working conditions).
S1: establishing a source domain data set by utilizing three working conditions, namely segmenting and standardizing data, and establishing a label corresponding to each data sample; and establishing a target domain data set by using a fourth working condition, wherein the data segmentation and standardization are also included, but the data set does not establish a corresponding label.
S2: and establishing a feature extractor based on a one-dimensional convolutional neural network.
S3: a multi-source domain discriminator is established, the source domain comprises three data sets, so that the domain discriminator has three.
S4: and constructing a classifier to obtain a multi-source domain self-adaptive intelligent fault diagnosis model.
The network training comprises the following specific steps:
s4.1: 64 data samples are taken from the source domain dataset 1 and 64 data samples are taken from the target domain dataset.
S4.2: the countermeasure loss L is calculated by the formula (2)DAThe maximum mean difference L of the source domain data set and the target domain data set is calculated using equation (3)MMDCalculating the classification difference L using the formula (4)c。
S4.3: updating network parameters by using the obtained confrontation loss, the maximum mean difference and the classification difference respectively, wherein the formula is as follows:
Figure BDA0002946900600000081
Figure BDA0002946900600000091
Figure BDA0002946900600000092
wherein
Figure BDA0002946900600000093
For tagged source domain datasets, XtFor unlabeled target domain data sets, θcRepresenting classifier network parameters, domain discriminator parameters are
Figure BDA0002946900600000094
θfNetwork parameters representing the feature extractor.
S4.4: and circulating the steps until the network is converged to obtain the trained network.
S5: and carrying out fault classification on the target domain data based on the trained network, wherein the output of the classifier is the prediction result.
Example 2:
the embodiment 2 of the present disclosure provides a multisource domain self-adaptive intelligent mechanical fault diagnosis system, including:
a data acquisition module configured to: acquiring running state data of mechanical equipment;
a feature extraction module configured to: extracting the characteristics of the acquired running state data by using a preset characteristic extractor;
a classification prediction module configured to: inputting the extracted features into a preset classifier to obtain a fault classification prediction result;
and updating the network parameters of the feature extractor according to the classification loss of a preset multi-source domain discriminator, the maximum mean difference of the source domain data set and the target domain data set and the classification loss obtained by the classifier according to the source domain data set.
The working method of the system is the same as the multisource domain self-adaptive intelligent mechanical fault diagnosis method provided by the embodiment 1, and details are not repeated here.
Example 3:
the embodiment 3 of the present disclosure provides a computer-readable storage medium, on which a program is stored, where the program, when executed by a processor, implements the steps in the multi-source domain adaptive intelligent mechanical fault diagnosis method according to embodiment 1 of the present disclosure, where the steps are:
acquiring running state data of mechanical equipment;
extracting the characteristics of the acquired running state data by using a preset characteristic extractor;
inputting the extracted features into a preset classifier to obtain a fault classification prediction result;
and updating the network parameters of the feature extractor according to the classification loss of a preset multi-source domain discriminator, the maximum mean difference of the source domain data set and the target domain data set and the classification loss obtained by the classifier according to the source domain data set.
The detailed steps are the same as those of the multisource domain self-adaptive intelligent mechanical fault diagnosis method provided by the embodiment 1, and are not repeated here.
Example 4:
the embodiment 4 of the present disclosure provides an electronic device, which includes a memory, a processor, and a program stored in the memory and capable of running on the processor, where the processor executes the program to implement the steps in the multi-source-domain adaptive intelligent mechanical fault diagnosis method according to embodiment 1 of the present disclosure, where the steps are:
acquiring running state data of mechanical equipment;
extracting the characteristics of the acquired running state data by using a preset characteristic extractor;
inputting the extracted features into a preset classifier to obtain a fault classification prediction result;
and updating the network parameters of the feature extractor according to the classification loss of a preset multi-source domain discriminator, the maximum mean difference of the source domain data set and the target domain data set and the classification loss obtained by the classifier according to the source domain data set.
The detailed steps are the same as those of the multisource domain self-adaptive intelligent mechanical fault diagnosis method provided by the embodiment 1, and are not repeated here.
As will be appreciated by one of skill in the art, embodiments of the present disclosure may be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.
The present disclosure is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the disclosure. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure, and various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.

Claims (9)

1. A multisource domain self-adaptive intelligent mechanical fault diagnosis method is characterized by comprising the following steps: the method comprises the following steps:
acquiring running state data of mechanical equipment;
extracting the characteristics of the acquired running state data by using a preset characteristic extractor;
inputting the extracted features into a preset classifier to obtain a fault classification prediction result;
the network parameters of the feature extractor are updated according to the classification loss of a preset multi-source domain discriminator, the maximum mean difference of a source domain data set and a target domain data set and the classification loss obtained by the classifier according to the source domain data set;
the multi-source domain discriminator comprises a multi-source domain discrimination module and a maximum mean difference module, wherein each source domain establishes a domain discriminator, and the data from the ith source domain or target domain is subjected to feature extraction to obtain features;
the ith discriminator receives the derived features and discriminates whether the derived features are from the ith source domain or the target domain, and the ith discriminator accepts only data from the ith source domain and data from the target domain.
2. The multi-source domain adaptive intelligent mechanical fault diagnosis method of claim 1, characterized in that:
the feature extractor is a feature extractor of a one-dimensional convolutional neural network and comprises four one-dimensional convolutional layers, each convolutional layer comprises a one-dimensional batch standardization layer, a ReLU activation layer and a maximum pooling layer, data of a fourth convolutional layer in the four one-dimensional convolutional layers are expanded, and the output of the feature extractor is obtained through a full connection layer, a ReLU activation function and a Dropout layer.
3. The multi-source domain adaptive intelligent mechanical fault diagnosis method of claim 1, characterized in that:
the source domain and the target domain share a feature space and the data distribution is different.
4. The multi-source domain adaptive intelligent mechanical fault diagnosis method of claim 1, characterized in that:
a maximum mean difference comprising:
Figure FDA0003565365730000011
wherein P and Q represent source domain data
Figure FDA0003565365730000021
And target domain data XtF denotes the mapping of the reconstructed kernel Hilbert space, HkThe representation denotes hilbert space with kernel k.
5. The multi-source domain adaptive intelligent mechanical fault diagnosis method of claim 1, characterized in that:
and performing transposition and normalization processing on the acquired running state data, and performing feature extraction by using the data subjected to transposition and normalization processing.
6. The multi-source domain adaptive intelligent mechanical fault diagnosis method of claim 1, characterized in that:
the data sources in the target domain data set and the data sources in the source domain data set belong to different working conditions.
7. The utility model provides a multisource territory self-adaptation intelligent machine fault diagnostic system which characterized in that: the method comprises the following steps:
a data acquisition module configured to: acquiring running state data of mechanical equipment;
a feature extraction module configured to: extracting the characteristics of the acquired running state data by using a preset characteristic extractor;
a classification prediction module configured to: inputting the extracted features into a preset classifier to obtain a fault classification prediction result;
the network parameters of the feature extractor are updated according to the classification loss of a preset multi-source domain discriminator, the maximum mean difference of a source domain data set and a target domain data set and the classification loss obtained by the classifier according to the source domain data set;
the multi-source domain discriminator comprises a multi-source domain discrimination module and a maximum mean value difference module, wherein each source domain establishes a domain discriminator, and the data from the ith source domain or target domain is subjected to feature extraction to obtain features;
the ith discriminator receives the derived features and discriminates whether the derived features are from the ith source domain or the target domain, and the ith discriminator accepts only data from the ith source domain and data from the target domain.
8. A computer-readable storage medium, on which a program is stored, wherein the program, when executed by a processor, implements the steps in the multi-source domain adaptive intelligent mechanical fault diagnosis method of any one of claims 1-6.
9. An electronic device comprising a memory, a processor, and a program stored on the memory and executable on the processor, wherein the processor implements the steps of the multi-source-domain adaptive intelligent mechanical fault diagnosis method of any one of claims 1-6 when executing the program.
CN202110196455.5A 2021-02-22 2021-02-22 Multi-source domain self-adaptive intelligent mechanical fault diagnosis method and system Active CN112926642B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110196455.5A CN112926642B (en) 2021-02-22 2021-02-22 Multi-source domain self-adaptive intelligent mechanical fault diagnosis method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110196455.5A CN112926642B (en) 2021-02-22 2021-02-22 Multi-source domain self-adaptive intelligent mechanical fault diagnosis method and system

Publications (2)

Publication Number Publication Date
CN112926642A CN112926642A (en) 2021-06-08
CN112926642B true CN112926642B (en) 2022-06-07

Family

ID=76170058

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110196455.5A Active CN112926642B (en) 2021-02-22 2021-02-22 Multi-source domain self-adaptive intelligent mechanical fault diagnosis method and system

Country Status (1)

Country Link
CN (1) CN112926642B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113673570A (en) * 2021-07-21 2021-11-19 南京旭锐软件科技有限公司 Training method, device and equipment for electronic device picture classification model
CN114997333B (en) * 2022-06-29 2024-04-23 清华大学 Fault diagnosis method and device for wind driven generator
CN115420490A (en) * 2022-09-01 2022-12-02 重庆大学 Mechanical fault intelligent diagnosis method based on relationship migration domain generalization network

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111898634B (en) * 2020-06-22 2022-08-16 西安交通大学 Intelligent fault diagnosis method based on depth-to-reactance-domain self-adaption
CN111898095B (en) * 2020-07-10 2024-04-19 佛山科学技术学院 Deep migration learning intelligent fault diagnosis method, device, storage medium and equipment
CN112308147B (en) * 2020-11-02 2024-02-09 西安电子科技大学 Rotary machinery fault diagnosis method based on multi-source domain anchor adapter integrated migration

Also Published As

Publication number Publication date
CN112926642A (en) 2021-06-08

Similar Documents

Publication Publication Date Title
CN112926642B (en) Multi-source domain self-adaptive intelligent mechanical fault diagnosis method and system
Zhao et al. Deep multi-scale convolutional transfer learning network: A novel method for intelligent fault diagnosis of rolling bearings under variable working conditions and domains
CN110647830B (en) Bearing fault diagnosis method based on convolutional neural network and Gaussian mixture model
CN116304820B (en) Bearing fault type prediction method and system based on multi-source domain transfer learning
CN114358124B (en) New fault diagnosis method for rotary machinery based on deep countermeasure convolutional neural network
CN114004252A (en) Bearing fault diagnosis method, device and equipment
CN114429152A (en) Rolling bearing fault diagnosis method based on dynamic index antagonism self-adaption
CN110757510A (en) Method and system for predicting remaining life of robot
CN115358259A (en) Self-learning-based unsupervised cross-working-condition bearing fault diagnosis method
CN114970715A (en) Variable working condition fault diagnosis method and system under small sample and unbalanced data constraint
CN113919540A (en) Method for monitoring running state of production process and related equipment
CN112257862A (en) Semi-supervised identification method based on relational network marker sample expansion
CN113884300A (en) Rolling bearing fault diagnosis method for deep anti-migration learning
CN113239610A (en) Domain self-adaptive rolling bearing fault diagnosis method based on Wasserstein distance
Flotzinger et al. Building inspection toolkit: Unified evaluation and strong baselines for damage recognition
CN116204781A (en) Rotary machine fault migration diagnosis method and system
Ekman et al. Task based semantic segmentation of soft X-ray CT images using 3D convolutional neural networks
CN116089812A (en) Fault diagnosis method based on semi-supervised contrast domain generalization intelligent model
Liu et al. A dual-branch balance saliency model based on discriminative feature for fabric defect detection
Zhu et al. Profile Abstract: An Optimization-based Subset Selection and Summarization Method for Profile Data Mining
CN115217534A (en) Method and system for monitoring service quality state of steam turbine
CN114548295A (en) Bearing fault classification system and method based on multi-scale domain adaptive network
Moghaddam et al. Operator's Experience-Level Classification Based on Micro-Assembly Steps for Smart Factories
Darveau Prognostics and Availability for Industrial Equipment Using High Performance Computing (HPC) and AI Technology
CN117634006B (en) BIM technology-based sleeve embedded engineering management system and method

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