CN116561655A - Fault diagnosis method, device, equipment and storage medium for motor current signal - Google Patents
Fault diagnosis method, device, equipment and storage medium for motor current signal Download PDFInfo
- Publication number
- CN116561655A CN116561655A CN202310504843.4A CN202310504843A CN116561655A CN 116561655 A CN116561655 A CN 116561655A CN 202310504843 A CN202310504843 A CN 202310504843A CN 116561655 A CN116561655 A CN 116561655A
- Authority
- CN
- China
- Prior art keywords
- fault
- current signal
- loss
- motor current
- source domain
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 98
- 238000003745 diagnosis Methods 0.000 title claims abstract description 95
- 238000003860 storage Methods 0.000 title claims abstract description 16
- 238000013145 classification model Methods 0.000 claims abstract description 78
- 238000012545 processing Methods 0.000 claims abstract description 14
- 230000008485 antagonism Effects 0.000 claims abstract description 10
- 238000012549 training Methods 0.000 claims description 64
- 238000005259 measurement Methods 0.000 claims description 14
- 230000006870 function Effects 0.000 claims description 12
- 238000000691 measurement method Methods 0.000 claims description 10
- 238000000605 extraction Methods 0.000 abstract description 19
- 238000004891 communication Methods 0.000 description 12
- 238000010586 diagram Methods 0.000 description 7
- 230000008901 benefit Effects 0.000 description 6
- 238000011176 pooling Methods 0.000 description 4
- 229910000831 Steel Inorganic materials 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000004590 computer program Methods 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 239000010959 steel Substances 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 230000005484 gravity Effects 0.000 description 2
- 238000013508 migration Methods 0.000 description 2
- 230000005012 migration Effects 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000013526 transfer learning Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 238000002405 diagnostic procedure Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005070 sampling 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
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/34—Testing dynamo-electric machines
-
- 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
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- 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
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
- Y04S10/52—Outage or fault management, e.g. fault detection or location
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The application discloses a motor current signal fault diagnosis method, a device, equipment and a storage medium, wherein the motor current signal fault diagnosis method comprises the following steps: acquiring a motor current signal; inputting the motor current signal into a preset signal classification model, and performing fault identification processing on the motor current signal based on the signal classification model to obtain fault information; the signal classification model comprises a feature extractor and a fault discriminator, wherein the feature extractor and the fault discriminator are trained based on a semantic alignment packet countermeasure method. According to the method, the characteristic extractor and the fault discriminator in the pre-trained signal classification model are used for carrying out semantic alignment grouping antagonism on the motor current signals, so that samples which come from different fields and have the same category labels can be mapped nearby in the characteristic space, the characteristic extraction capacity of the model is improved, and the accuracy of fault diagnosis is improved.
Description
Technical Field
The present disclosure relates to the field of fault diagnosis technologies, and in particular, to a fault diagnosis method, apparatus, device, and storage medium for a motor current signal.
Background
The SCARA robot is an important member in industrial production automation, and has 20 functions of high-precision accessory processing, carrying, assembling, sorting and the like. SCARA robots are favored by automated enterprises by virtue of their high rotational speeds and high precision processing advantages. However, the screw rod of the SCARA robot is easy to have the faults of clamping stagnation, steel ball shortage and the like under the running conditions of high rotating speed and long time. Thus, accurate fault diagnosis of SCARA robots is a current challenge to be solved.
The main method for engineering practice in the related art is migration learning. The diagnosis method of the transfer learning mainly comprises a characteristic distance measurement-based method, a field self-adaption method based on the countermeasure learning and the like. However, due to the flexible link between the motor and the screw rod, the fault characteristics contained in the motor current are extremely weak, so that the fault characteristics are difficult to extract by a self-adaptive method in the characteristic distance measurement and countermeasure field, and the fault diagnosis accuracy is low.
Disclosure of Invention
The main purpose of the application is to provide a fault diagnosis method, device, equipment and storage medium for motor current signals, and aims to solve the technical problem of low accuracy of fault diagnosis in the prior art.
To achieve the above object, the present application provides a fault diagnosis method of a motor current signal, the fault diagnosis method of the motor current signal including:
Acquiring a motor current signal;
inputting the motor current signal into a preset signal classification model, and performing fault identification processing on the motor current signal based on the signal classification model to obtain fault information;
the signal classification model comprises a feature extractor and a fault discriminator, wherein the feature extractor and the fault discriminator are trained based on a semantic alignment packet countermeasure method.
Optionally, before the step of acquiring the motor current signal, the method includes:
acquiring a source domain current signal sample and a fault result label of the source domain current signal sample;
based on the source domain current signal sample, pre-training to obtain an initial feature extractor;
grouping the source domain current signal samples into a preset first number of semantic alignment subgroups to obtain grouped source domain current signal samples;
pre-training to obtain an initial fault discriminator based on the grouped source domain current signal samples;
and performing iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample and a fault result label of the source domain current signal sample to obtain a signal classification model meeting the precision condition.
Optionally, the step of performing iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample and the fault result label of the source domain current signal sample to obtain a signal classification model meeting the accuracy condition includes:
inputting source domain current signal samples in the semantic alignment group to the initial feature extractor, extracting first fault features of the source domain current signal samples based on the initial feature extractor, and determining predicted fault information based on the first fault features;
determining a fault classification loss based on the predicted fault information and the fault result tag;
inputting the first fault characteristic to the initial fault discriminator to obtain group countermeasure loss;
determining a target loss based on the fault classification loss and the group countermeasure loss;
judging whether the target loss meets a loss standard indicated by a preset loss threshold range or not;
and if the target loss does not meet the loss standard indicated by the preset loss threshold range, returning to input the source domain current signal sample in the semantic alignment group to the initial feature extractor, extracting a first fault feature of the source domain current signal sample based on the initial feature extractor to obtain predicted fault information, and stopping training until the target loss meets the loss standard indicated by the preset loss threshold range to obtain a signal classification model meeting accuracy conditions.
Optionally, the step of performing iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample and the fault result label of the source domain current signal sample to obtain a signal classification model meeting the accuracy condition includes:
obtaining a target domain current signal sample;
and performing iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample, the fault result label of the source domain current signal sample and the target domain current signal sample and combining a Maximum Mean Difference (MMD) distance measurement method to obtain a signal classification model meeting the precision condition.
Optionally, the step of iteratively training the initial feature extractor and the initial fault discriminator to obtain a signal classification model with a precision condition met based on the source domain current signal sample, the fault result label of the source domain current signal sample, and the target domain current signal sample in combination with a maximum mean difference MMD distance measurement method includes:
inputting source domain current signal samples in the semantic alignment group to the initial feature extractor, extracting first fault features of the source domain current signal samples based on the initial feature extractor, and determining predicted fault information based on the first fault features;
Determining a fault classification loss based on the predicted fault information and the fault result tag;
inputting the first fault characteristic to the initial fault discriminator to obtain group countermeasure loss;
inputting the target domain current signal sample to the initial feature extractor to obtain a second fault feature of the target domain;
calculating the distance loss of the first fault feature and the second fault feature by adopting a Maximum Mean Difference (MMD) distance measurement function;
calculating the sum of the fault classification loss, the group antagonism loss and the distance loss to obtain a target loss;
judging whether the target loss meets a loss standard indicated by a preset loss threshold range or not;
and if the target loss does not meet the loss standard indicated by the preset loss threshold range, returning to input the source domain current signal sample in the semantic alignment group to the initial feature extractor, extracting a first fault feature of the source domain current signal sample based on the initial feature extractor to obtain predicted fault information, and stopping training until the target loss meets the loss standard indicated by the preset loss threshold range to obtain a signal classification model meeting accuracy conditions.
Optionally, the step of inputting the target domain current signal sample to the initial feature extractor to obtain a second fault feature of the target domain includes:
performing fast Fourier transform on the target domain current signal sample to obtain a transformed target domain current signal sample;
normalizing the transformed target domain current signal sample to obtain a normalized target domain current signal sample;
and inputting the normalized target domain current signal sample to the initial feature extractor to obtain a second fault feature of the target domain.
Optionally, the signal classification model employs a lightweight convolutional network, wherein the lightweight convolutional network is constructed from a depth Depthwise convolution, a point-by-point Pointwise convolution, and a Group convolution.
The present application also provides a fault diagnosis device for a motor current signal, which is characterized in that the fault diagnosis device for a motor current signal includes:
the acquisition module is used for acquiring a motor current signal;
the recognition module is used for inputting the motor current signal into a preset signal classification model, and performing fault recognition processing on the motor current signal based on the signal classification model to obtain fault information; the signal classification model comprises a feature extractor and a fault discriminator, wherein the feature extractor and the fault discriminator are trained based on a semantic alignment packet countermeasure method.
The present application also provides a fault diagnosis apparatus of a motor current signal, the fault diagnosis apparatus of a motor current signal including: a memory, a processor, and a program stored on the memory for implementing a fault diagnosis method of the motor current signal,
the memory is used for storing a program for realizing a fault diagnosis method of the motor current signal;
the processor is configured to execute a program for implementing a fault diagnosis method of the motor current signal to implement the steps of the fault diagnosis method of the motor current signal.
The present application also provides a storage medium having stored thereon a program for implementing a fault diagnosis method of a motor current signal, the program for implementing the fault diagnosis method of a motor current signal being executed by a processor to implement the steps of the fault diagnosis method of a motor current signal.
Compared with the prior art, because of the flexible link between the motor and the screw rod, the fault diagnosis method, the device, the equipment and the storage medium for the motor current signal provided by the application have the advantages that the fault characteristics contained in the motor current are extremely weak, so that the characteristic distance measurement is difficult to extract the fault characteristics compared with the self-adaptive method in the countermeasure field, and the accuracy of fault diagnosis is low; inputting the motor current signal into a preset signal classification model, and performing fault identification processing on the motor current signal based on the signal classification model to obtain fault information; the signal classification model comprises a feature extractor and a fault discriminator, wherein the feature extractor and the fault discriminator are trained based on a semantic alignment packet countermeasure method. In the method, the semantic alignment grouping countermeasure is carried out on the motor current signals through the feature extractor and the fault discriminator in the pre-trained signal classification model, so that samples from different fields but with the same category labels can be mapped nearby in the feature space, the feature extraction capacity of the model is improved, and the accuracy of fault diagnosis is improved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the application and together with the description, serve to explain the principles of the application. In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required to be used in the description of the embodiments or the prior art will be briefly described below, and it will be obvious to those skilled in the art that other drawings can be obtained from these drawings without inventive effort.
FIG. 1 is a schematic diagram of a device architecture of a hardware operating environment according to an embodiment of the present application;
FIG. 2 is a schematic flow chart of a first embodiment of a fault diagnosis method for motor current signals according to the present application;
FIG. 3 is a schematic block diagram of a fault diagnosis device for motor current signals according to the present application;
FIG. 4 is a schematic diagram of a semantic alignment grouping countermeasure setting in a first embodiment of a fault diagnosis method for motor current signals of the present application;
fig. 5 is a schematic structural diagram of a signal classification model of a first embodiment of a fault diagnosis method for motor current signals according to the present application;
FIG. 6 is a schematic diagram of a diagnosis flow of a second embodiment of a fault diagnosis method for motor current signals according to the present application;
Fig. 7 is a flowchart of a third embodiment of a fault diagnosis method for motor current signals according to the present application.
The realization, functional characteristics and advantages of the present application will be further described with reference to the embodiments, referring to the attached drawings.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.
As shown in fig. 1, fig. 1 is a schematic diagram of a terminal structure of a hardware running environment according to an embodiment of the present application.
The terminal in the embodiment of the application may be a PC, or may be a mobile terminal device with a display function, such as a smart phone, a tablet computer, an electronic book reader, an MP3 (Moving Picture Experts Group Audio Layer III, dynamic image expert compression standard audio layer 3) player, an MP4 (Moving Picture Experts Group Audio Layer IV, dynamic image expert compression standard audio layer 4) player, a portable computer, or the like.
As shown in fig. 1, the terminal may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, a communication bus 1002. Wherein the communication bus 1002 is used to enable connected communication between these components. The user interface 1003 may include a Display, an input unit such as a Keyboard (Keyboard), and the optional user interface 1003 may further include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1005 may be a high-speed RAM memory or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also optionally be a storage device separate from the processor 1001 described above.
Optionally, the terminal may also include a camera, an RF (Radio Frequency) circuit, a sensor, an audio circuit, a WiFi module, and so on. Among other sensors, such as light sensors, motion sensors, and other sensors. Specifically, the light sensor may include an ambient light sensor that may adjust the brightness of the display screen according to the brightness of ambient light, and a proximity sensor that may turn off the display screen and/or the backlight when the mobile terminal moves to the ear. As one of the motion sensors, the gravity acceleration sensor can detect the acceleration in all directions (generally three axes), and can detect the gravity and the direction when the mobile terminal is stationary, and the mobile terminal can be used for recognizing the gesture of the mobile terminal (such as horizontal and vertical screen switching, related games, magnetometer gesture calibration), vibration recognition related functions (such as pedometer and knocking), and the like; of course, the mobile terminal may also be configured with other sensors such as a gyroscope, a barometer, a hygrometer, a thermometer, an infrared sensor, and the like, which are not described herein.
It will be appreciated by those skilled in the art that the terminal structure shown in fig. 1 is not limiting of the terminal and may include more or fewer components than shown, or may combine certain components, or a different arrangement of components.
As shown in fig. 1, a memory 1005 as a computer storage medium may include an operation device, a network communication module, a user interface module, and a fault diagnosis program of a motor current signal.
In the terminal shown in fig. 1, the network interface 1004 is mainly used for connecting to a background server and performing data communication with the background server; the user interface 1003 is mainly used for connecting a client (user side) and performing data communication with the client; and the processor 1001 may be used to invoke a fault diagnosis program for the motor current signal stored in the memory 1005.
Referring to fig. 2, an embodiment of the present application provides a fault diagnosis method for a motor current signal, including:
step S100, obtaining a motor current signal;
step S200, inputting the motor current signal into a preset signal classification model, and performing fault identification processing on the motor current signal based on the signal classification model to obtain fault information;
the signal classification model comprises a feature extractor and a fault discriminator, wherein the feature extractor and the fault discriminator are trained based on a semantic alignment packet countermeasure method.
In this embodiment, the application scenario aimed at is:
as an example, the fault diagnosis of the motor current signal may be that the screw rod of the SCARA robot is easy to generate a fault such as a clamping stagnation, a steel ball shortage, etc. under the running condition of high rotation speed and long time, and the fault category needs to be accurately identified. The main method for engineering practice in the related art is migration learning. The diagnosis method of the transfer learning mainly comprises a characteristic distance measurement-based method, a field self-adaption method based on the countermeasure learning and the like. However, due to the flexible link between the motor and the screw rod, the fault characteristics contained in the motor current are extremely weak, so that the fault characteristics are difficult to extract by a self-adaptive method in the characteristic distance measurement and countermeasure field, and the fault diagnosis accuracy is low. Aiming at the scene, the fault diagnosis method of the motor current signal carries out semantic alignment grouping antagonism on the motor current signal through the feature extractor and the fault discriminator in the pre-trained signal classification model, so that samples from different fields but with the same category labels can be mapped nearby in the feature space, the feature extraction capacity of the model is improved, and the accuracy of fault diagnosis is improved.
As an example, the application scenario of the fault diagnosis of the motor current signal is not only the fault diagnosis processing of the screw rod of the opposite pressure SCARA robot described above, but also various motor fault diagnosis scenarios, which are not particularly limited herein.
The present embodiment aims at: and the accuracy of fault diagnosis is improved.
In the present embodiment, the failure diagnosis method of the motor current signal is applied to the failure diagnosis device of the motor current signal.
The method comprises the following specific steps:
step S100, obtaining a motor current signal;
in this embodiment, the motor current signal is a current signal output by a motor of the SCARA robot, and since the existing diagnostic methods all rely on vibration signals as diagnostic data, adding a vibration sensor to each SCARA robot in engineering application increases huge economic cost, so that the motor current signal is selected as diagnostic data, the scheme is convenient to implement, and the hardware cost is saved.
In this embodiment, the manner in which the device obtains the motor current signal may be to receive the motor current signal sent by the motor in real time; or the device can receive motor current signals transmitted by the motor at fixed time and judge whether the motor current signals are failure information or not.
Step S200, inputting the motor current signal into a preset signal classification model, and performing fault identification processing on the motor current signal based on the signal classification model to obtain fault information;
the signal classification model comprises a feature extractor and a fault discriminator, wherein the feature extractor and the fault discriminator are trained based on a semantic alignment packet countermeasure method.
In this embodiment, the signal classification model includes a feature extractor and a fault discriminator, where the feature extractor and the fault discriminator are trained based on a method of semantic alignment packet antagonism, where the semantic alignment packet antagonism is used to ensure that samples from different fields but with the same class of labels can be mapped closely in a feature space, find domain invariant features in different fields in the feature sample, and also implement semantic alignment between different labels, specifically, referring to fig. 4, a number of semantic alignment subgroups are set, each subgroup includes current data of different machines, where 0, 1, and 2 respectively represent different types of status labels, such as a stuck, missing steel ball, and a normal status label, and the number of subgroups can be set by themselves, where G1 and G2, G3, and G4 antagonism (adeversal) can be found not only the domain invariant features in different fields, but also implement semantic alignment between different labels, and G5 and G6, G7, and G8 are the same.
Before the step S100, the step of obtaining a motor current signal, the method comprises the following steps a100-a400:
step A100, acquiring a source domain current signal sample and a fault result label of the source domain current signal sample;
in this embodiment, the source domain (source domain) represents a different domain than the test sample, but has rich supervision information; the target domain (target domain) represents the domain in which the test sample is located, with no label or only a small number of labels. The source domain and the target domain often belong to the same task, but are distributed differently, the source domain current signal sample is a current signal sample obtained based on a laboratory, and a corresponding label exists, namely a fault result label of the source domain current signal sample.
Step A200, pre-training to obtain an initial feature extractor based on the source domain current signal sample;
in this embodiment, the device pre-trains to obtain an initial feature extractor based on the source domain current signal sample, where the initial feature extractor refers to a model with an extraction function of the current signal sample, and continuously converges in iterative training until a feature extractor meeting a precision condition is obtained.
Step A300, grouping the source domain current signal samples into a preset first number of semantic alignment subgroups to obtain grouped source domain current signal samples;
in this embodiment, the device groups the source domain current signal samples into a preset first number of semantic alignment subgroups, to obtain grouped source domain current signal samples, specifically, the device randomly samples from the source domain current signal samples and groups the source domain current signal samples into a preset first number of semantic alignment subgroups, for example, the device randomly samples from a source domain dataset and groups the source domain current signal samples into G1 to G8 groups.
Step A400, pre-training to obtain an initial fault discriminator based on the grouped source domain current signal samples;
in this embodiment, the device pre-trains to obtain an initial fault identifier based on the grouped source domain current signal samples, where the initial fault identifier is a model for fault determination of the current signal samples, and continuously converges in iterative training until a fault identifier meeting a precision condition is obtained, where the fault identifier is used for assisting in training the feature extractor.
And step A500, performing iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample and the fault result label of the source domain current signal sample to obtain a signal classification model meeting the precision condition.
In this embodiment, the apparatus performs iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal samples and fault result labels of the source domain current signal samples to obtain a signal classification model with accuracy conditions, where, referring to fig. 5, the application uses a lightweight convolutional network: depth convolution DW (Depthwise convolution), point-wise convolution PW (Pointwise convolution), and group convolution GC (Group convolution). The feature fusion block consists of PW, DW and an average pooling layer, wherein the average pooling layer is used for multiplexing upper layer features, so that the generalization performance of the model is improved. Channel stitching can avoid the use of convolution kernels to boost the number of channels, and thus can further reduce the number of model parameters. The channel shuffling is used for fusing information of two branches of the feature fusion block, so that feature extraction capacity of the model is improved.
Specifically, the step A500 includes the following steps A510-A560:
step A510, inputting source domain current signal samples in the semantic alignment group to the initial feature extractor, extracting first fault features of the source domain current signal samples based on the initial feature extractor, and determining prediction fault information based on the first fault features;
In this embodiment, the apparatus inputs source domain current signal samples in the semantic alignment group to the initial feature extractor, extracts a first fault feature of the source domain current signal samples based on the initial feature extractor, and determines predicted fault information based on the first fault feature.
Specifically, the step A510 includes the following steps A511-A513:
step A511, based on the initial feature extractor, performing feature extraction on the source domain current signal sample to obtain a third fault feature;
in this embodiment, the apparatus performs feature extraction on the source domain current signal sample based on the initial feature extractor to obtain a third fault feature, where a method of feature extraction includes, but is not limited to, a HOG feature extraction algorithm, an LBP feature extraction algorithm, and a Haar feature extraction algorithm.
Step A512, rolling and averaging pooling the third fault feature to obtain a fourth fault feature;
and step A513, performing channel splicing and channel shuffling on the fourth fault characteristic, and fusing characteristic information to obtain a first fault characteristic of the source domain current signal sample.
In this embodiment, referring to fig. 5, the apparatus performs rolling and averaging pooling on the third fault feature based on the lightweight diagnostic model of DW, PW and GC to obtain a fourth fault feature, and performs channel splicing and channel shuffling on the fourth fault feature, and fuses feature information to obtain the first fault feature of the source domain current signal sample.
Step A520, determining a fault classification loss based on the predicted fault information and the fault result label;
in this embodiment, the apparatus determines the failure classification loss based on the predicted failure information and the failure result label, and specifically, the apparatus calculates the failure classification loss using a mean square error based on the predicted failure information and the failure result label.
Step A530, inputting the first fault characteristic to the initial fault discriminator to obtain a group countermeasure loss;
in this embodiment, the apparatus inputs the first fault signature to the initial fault discriminator to obtain a group countermeasure loss.
Step a540, determining a target loss based on the fault classification loss and the group countermeasure loss;
in the present embodiment, the apparatus determines a target loss based on the failure classification loss and the group countermeasure loss, specifically, the apparatus adds the failure classification loss and the group countermeasure loss as the target loss.
Step A550, judging whether the target loss meets a loss standard indicated by a preset loss threshold range;
in this embodiment, the apparatus determines whether the target loss meets a loss criterion indicated by a preset loss threshold range, where the loss criterion indicated by the loss threshold range is set by itself, that is, is used to determine whether the target loss in the current training period is converged, and if the target loss meets the loss criterion indicated by the preset loss threshold range, it indicates that the target loss in the current training period is converged, meets a precision requirement, and completes training of the signal classification model; if the target loss does not meet the loss standard indicated by the preset loss threshold range, namely the target loss in the current training period is not converged, the accuracy requirement is not met, and the next period of iterative training is required.
And step A560, if the target loss does not meet the loss standard indicated by the preset loss threshold range, returning to input the source domain current signal sample in the semantic alignment group to the initial feature extractor, extracting a first fault feature of the source domain current signal sample based on the initial feature extractor to obtain predicted fault information, and stopping training until the target loss meets the loss standard indicated by the preset loss threshold range, thereby obtaining a signal classification model meeting the precision condition.
In this embodiment, if the target loss does not meet the loss standard indicated by the preset loss threshold range, it indicates that the target loss in the current training period is not converged, and does not meet the accuracy requirement, and the next period of iterative training is required, the apparatus returns to input the source domain current signal samples in the semantic alignment group to the initial feature extractor, extracts the first fault feature of the source domain current signal samples based on the initial feature extractor, and obtains the predicted fault information, until the target loss meets the loss standard indicated by the preset loss threshold range, and stops training, thereby obtaining the signal classification model meeting the accuracy condition.
Compared with the prior art that a flexible link exists between a motor and a screw rod, the fault diagnosis method for the motor current signal provided by the application has the advantages that the fault characteristics contained in the motor current are extremely weak, so that the characteristic distance measurement is difficult to extract the fault characteristics compared with a self-adaptive method in the countermeasure field, and the accuracy of fault diagnosis is low, and in the application, the motor current signal is acquired; inputting the motor current signal into a preset signal classification model, and performing fault identification processing on the motor current signal based on the signal classification model to obtain fault information; the signal classification model comprises a feature extractor and a fault discriminator, wherein the feature extractor and the fault discriminator are trained based on a semantic alignment packet countermeasure method. In the method, the semantic alignment grouping countermeasure is carried out on the motor current signals through the feature extractor and the fault discriminator in the pre-trained signal classification model, so that samples from different fields but with the same category labels can be mapped nearby in the feature space, the feature extraction capacity of the model is improved, and the accuracy of fault diagnosis is improved.
Based on the first embodiment, the present application further provides another embodiment, and the fault diagnosis method of the motor current signal includes:
Referring to fig. 6, the step a500 of performing iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal samples and the fault result labels of the source domain current signal samples, to obtain a signal classification model with accuracy satisfying conditions further includes the following steps B100-B200:
step B100, obtaining a target domain current signal sample;
in this embodiment, the target domain current signal sample is a current signal sample obtained on the industrial site, and no corresponding tag, i.e. a current signal sample of an unknown tag, is present.
And step B200, performing iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample, the fault result label of the source domain current signal sample and the target domain current signal sample and combining a Maximum Mean Difference (MMD) distance measurement method to obtain a signal classification model meeting the precision condition.
In this embodiment, the device performs iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample, the fault result label of the source domain current signal sample, and the target domain current signal sample in combination with a maximum mean difference MMD distance measurement method, so as to obtain a signal classification model with precision conditions, that is, the signal classification model is a method of combining semantic alignment grouping opposition and the maximum mean difference MMD feature distance based on semantic alignment grouping, so that feature extraction capability of the model is improved, and generalization performance of the signal classification model is further improved.
Specifically, the step B200 includes the following steps B210-B280:
step B210, inputting source domain current signal samples in the semantic alignment group to the initial feature extractor, extracting first fault features of the source domain current signal samples based on the initial feature extractor, and determining prediction fault information based on the first fault features;
step B220, determining fault classification loss based on the predicted fault information and the fault result label;
step B230, inputting the first fault characteristic to the initial fault discriminator to obtain a group countermeasure loss;
in this embodiment, the steps B210-B230 are the same as the steps A510-A530 described above, and will not be described again.
Step B240, inputting the current signal sample of the target domain to the initial feature extractor to obtain a second fault feature of the target domain;
in this embodiment, the device inputs the target domain current signal sample to the initial feature extractor, and extracts the feature of the target domain current signal sample to obtain the second fault feature of the target domain.
Specifically, the step B240 includes the following steps B241-B243:
step B241, performing fast Fourier transform on the target domain current signal sample to obtain a transformed target domain current signal sample;
Step B242, normalizing the transformed target domain current signal sample to obtain a normalized target domain current signal sample;
and step B243, inputting the normalized target domain current signal sample to the initial feature extractor to obtain a second fault feature of the target domain.
In this embodiment, the device collects lead screw current data from an industrial site and performs fast fourier transform, then normalizes the data, and finally inputs the data into a feature extractor to obtain a fault feature of a target domain, where the fast fourier transform (fast Fourier transform), that is, a general term of a high-efficiency and fast computing method for computing Discrete Fourier Transform (DFT) by using a computer, is referred to as FFT for short.
Step B250, calculating the distance loss of the first fault feature and the second fault feature by adopting a Maximum Mean Difference (MMD) distance measurement function;
in this embodiment, the apparatus calculates the distance loss of the first fault feature and the second fault feature by using a maximum mean difference MMD distance metric function, where the maximum mean difference MMD distance metric function is used to measure the difference between the two distributions, and the moment with the largest difference between the two distributions should be used as a criterion for measuring the two distributions, so as to calculate the distance loss of the first fault feature and the second fault feature.
Step B260, calculating the sum of the fault classification loss, the group countermeasure loss and the distance loss to obtain a target loss;
in this embodiment, the apparatus calculates the sum of the fault classification loss, the group antagonism loss and the distance loss to obtain the target loss, i.e. the loss accumulation and uses it to optimize the feature extractor.
Step B270, judging whether the target loss meets a loss standard indicated by a preset loss threshold range;
and step B280, if the target loss does not meet the loss standard indicated by the preset loss threshold range, returning to input the source domain current signal samples in the semantic alignment group to the initial feature extractor, extracting first fault features of the source domain current signal samples based on the initial feature extractor to obtain predicted fault information, and stopping training until the target loss meets the loss standard indicated by the preset loss threshold range to obtain a signal classification model meeting accuracy conditions.
In this embodiment, the steps B270 and B280 refer to the steps A550-A560, and are not described herein.
In the embodiment, the method for combining the multi-source domain grouping countermeasure and the feature distance measurement improves the feature extraction capability of the model, and the problem that the current signal contains weak fault features can be solved. The fault diagnosis model designed by the method can complete real-time diagnosis from the end to the end of the wire rod, so that the output of unqualified products can be reduced, and the economic benefit is improved.
Based on the first embodiment and the second embodiment, the present application further provides another embodiment, and referring to fig. 7, the fault diagnosis method for the motor current signal includes:
and step 1, loading a source domain data set acquired by a laboratory.
And 2, pre-training a feature extractor by using the source domain data set.
Step 3, randomly sampling from the source domain data set and dividing the source domain data set into groups G1 to G8.
And 4, training a discriminator by using the data from the groups G1 to G8.
And 5, inputting the G1 to G8 group data into a feature extractor, extracting fault features of the source domain, and then solving fault classification loss of the source domain according to the real label.
And 6, inputting the fault characteristics extracted in the step 4 into a discriminator and solving the group countermeasures.
And 7, collecting screw current data from an industrial site, performing fast Fourier transform, normalizing the data, and finally inputting the data into a feature extractor to obtain fault features of a target domain.
And 8, solving the distance loss between the fault characteristics of the step 4 and the step 6 by using an MMD distance measurement function.
Step 9, accumulating the losses obtained in the steps 4, 5 and 7 and using the losses to optimize the feature extractor.
And step 10, collecting data from the industrial site again and inputting the data into the feature extractor to obtain a fault prediction result.
And step 11, ending.
In this embodiment, the motor current signal is selected as the diagnosis data to save the economic cost, the lightweight diagnosis network is designed to facilitate the direct deployment of the model on the mobile terminal, and the method of combining the multi-source domain grouping countermeasure and the maximum mean difference MMD (Max Mean discrepancy) feature distance is designed to improve the feature extraction capability of the model, thereby improving the generalization performance of the model.
The present application also provides a fault diagnosis device of a motor current signal, referring to fig. 3, the fault diagnosis device of a motor current signal includes:
an acquisition module 10 for acquiring a motor current signal;
the recognition module 20 is configured to input the motor current signal to a preset signal classification model, and perform fault recognition processing on the motor current signal based on the signal classification model to obtain fault information; the signal classification model comprises a feature extractor and a fault discriminator, wherein the feature extractor and the fault discriminator are trained based on a semantic alignment packet countermeasure method.
Optionally, the fault diagnosis device of the motor current signal further includes:
the sample acquisition module is used for acquiring a source domain current signal sample and a fault result label of the source domain current signal sample;
The first pre-training module is used for pre-training to obtain an initial feature extractor based on the source domain current signal sample;
the grouping module is used for grouping the source domain current signal samples into a preset first number of semantic alignment subgroups to obtain grouped source domain current signal samples;
the second pre-training module is used for pre-training to obtain an initial fault discriminator based on the grouped source domain current signal samples;
and the training module is used for carrying out iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample and the fault result label of the source domain current signal sample to obtain a signal classification model meeting the precision condition.
Optionally, the training module includes:
a feature extraction module for inputting source domain current signal samples in the semantic alignment group to the initial feature extractor, extracting first fault features of the source domain current signal samples based on the initial feature extractor, and determining predicted fault information based on the first fault features;
the classification loss determining module is used for determining fault classification loss based on the predicted fault information and the fault result label;
The group countermeasures loss determination module is used for inputting the first fault characteristic into the initial fault discriminator to obtain the group countermeasures loss;
a target loss determination module for determining a target loss based on the fault classification loss and the group countermeasure loss;
the judging module is used for judging whether the target loss meets a loss standard indicated by a preset loss threshold range;
and the iterative training module is used for returning to input the source domain current signal samples in the semantic alignment group to the initial feature extractor if the target loss does not meet the loss standard indicated by the preset loss threshold range, extracting the first fault feature of the source domain current signal samples based on the initial feature extractor, and obtaining the predicted fault information until the target loss meets the loss standard indicated by the preset loss threshold range, and stopping training to obtain the signal classification model meeting the precision condition.
Optionally, the training module further includes:
the target domain signal sample acquisition module is used for acquiring a target domain current signal sample;
and the training module is combined with a distance measurement method and is used for carrying out iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample, the fault result label of the source domain current signal sample and the target domain current signal sample and combined with a maximum mean difference MMD distance measurement method to obtain a signal classification model meeting the precision condition.
Optionally, the training module combined with the distance measurement method includes:
a feature extraction module for inputting source domain current signal samples in the semantic alignment group to the initial feature extractor, extracting first fault features of the source domain current signal samples based on the initial feature extractor, and determining predicted fault information based on the first fault features;
the classification loss determining module is used for determining fault classification loss based on the predicted fault information and the fault result label;
the group countermeasures loss determination module is used for inputting the first fault characteristic into the initial fault discriminator to obtain the group countermeasures loss;
the fault feature extraction module is used for inputting the current signal sample of the target domain to the initial feature extractor to obtain a second fault feature of the target domain;
the distance measurement calculation module is used for calculating the distance loss of the first fault feature and the second fault feature by adopting a Maximum Mean Difference (MMD) distance measurement function;
the loss calculation module is used for calculating the sum of the fault classification loss, the group antagonism loss and the distance loss to obtain target loss;
The judging module is used for judging whether the target loss meets a loss standard indicated by a preset loss threshold range;
and the iterative training module is used for returning to input the source domain current signal samples in the semantic alignment group to the initial feature extractor if the target loss does not meet the loss standard indicated by the preset loss threshold range, extracting the first fault feature of the source domain current signal samples based on the initial feature extractor, and obtaining the predicted fault information until the target loss meets the loss standard indicated by the preset loss threshold range, and stopping training to obtain the signal classification model meeting the precision condition.
Optionally, the fault feature extraction module includes:
the Fourier transform module is used for performing fast Fourier transform on the target domain current signal sample to obtain a transformed target domain current signal sample;
the normalization module is used for normalizing the transformed target domain current signal sample to obtain a normalized target domain current signal sample;
and the target domain fault feature determining module is used for inputting the normalized target domain current signal sample to the initial feature extractor to obtain a second fault feature of the target domain.
The specific implementation manner of the fault diagnosis device for the motor current signal is basically the same as that of each embodiment of the fault diagnosis method for the motor current signal, and is not repeated here.
Referring to fig. 1, fig. 1 is a schematic diagram of a terminal structure of a hardware running environment according to an embodiment of the present application.
As shown in fig. 1, the terminal may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, a communication bus 1002. Wherein the communication bus 1002 is used to enable connected communication between these components. The user interface 1003 may include a Display, an input unit such as a Keyboard (Keyboard), and the optional user interface 1003 may further include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1005 may be a high-speed RAM memory or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also optionally be a storage device separate from the processor 1001 described above.
Optionally, the fault diagnosis device of the motor current signal may further include a rectangular user interface, a network interface, a camera, an RF (Radio Frequency) circuit, a sensor, an audio circuit, a WiFi module, and the like. The rectangular user interface may include a Display screen (Display), an input sub-module such as a Keyboard (Keyboard), and the optional rectangular user interface may also include a standard wired interface, a wireless interface. The network interface may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface).
It will be appreciated by those skilled in the art that the configuration of the fault diagnosis device for motor current signals shown in fig. 1 does not constitute a limitation of the fault diagnosis device for motor current signals, and may include more or less components than those illustrated, or may combine certain components, or may be a different arrangement of components.
As shown in fig. 1, a memory 1005, which is a storage medium, may include an operating system, a network communication module, and a fault diagnosis program of a motor current signal. The operating system is a program that manages and controls the hardware and software resources of the fault diagnosis device for the motor current signal, supporting the operation of the fault diagnosis program for the motor current signal, as well as other software and/or programs. The network communication module is used to enable communication between components within the memory 1005 and other hardware and software in the fault diagnosis system for motor current signals.
In the motor current signal fault diagnosis apparatus shown in fig. 1, a processor 1001 is configured to execute a motor current signal fault diagnosis program stored in a memory 1005, and implement the steps of the motor current signal fault diagnosis method described in any one of the above.
The specific implementation manner of the fault diagnosis device for the motor current signal is basically the same as that of each embodiment of the fault diagnosis method for the motor current signal, and is not repeated here.
The present application also provides a storage medium having stored thereon a program that implements a failure diagnosis method of a motor current signal, the program that implements the failure diagnosis method of a motor current signal being executed by a processor to implement the failure diagnosis method of a motor current signal as follows:
acquiring a motor current signal;
inputting the motor current signal into a preset signal classification model, and performing fault identification processing on the motor current signal based on the signal classification model to obtain fault information;
the signal classification model comprises a feature extractor and a fault discriminator, wherein the feature extractor and the fault discriminator are trained based on a semantic alignment packet countermeasure method.
Optionally, before the step of acquiring the motor current signal, the method includes:
acquiring a source domain current signal sample and a fault result label of the source domain current signal sample;
based on the source domain current signal sample, pre-training to obtain an initial feature extractor;
grouping the source domain current signal samples into a preset first number of semantic alignment subgroups to obtain grouped source domain current signal samples;
Pre-training to obtain an initial fault discriminator based on the grouped source domain current signal samples;
and performing iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample and a fault result label of the source domain current signal sample to obtain a signal classification model meeting the precision condition.
Optionally, the step of performing iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample and the fault result label of the source domain current signal sample to obtain a signal classification model meeting the accuracy condition includes:
inputting source domain current signal samples in the semantic alignment group to the initial feature extractor, extracting first fault features of the source domain current signal samples based on the initial feature extractor, and determining predicted fault information based on the first fault features;
determining a fault classification loss based on the predicted fault information and the fault result tag;
inputting the first fault characteristic to the initial fault discriminator to obtain group countermeasure loss;
determining a target loss based on the fault classification loss and the group countermeasure loss;
Judging whether the target loss meets a loss standard indicated by a preset loss threshold range or not;
and if the target loss does not meet the loss standard indicated by the preset loss threshold range, returning to input the source domain current signal sample in the semantic alignment group to the initial feature extractor, extracting a first fault feature of the source domain current signal sample based on the initial feature extractor to obtain predicted fault information, and stopping training until the target loss meets the loss standard indicated by the preset loss threshold range to obtain a signal classification model meeting accuracy conditions.
Optionally, the step of performing iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample and the fault result label of the source domain current signal sample to obtain a signal classification model meeting the accuracy condition includes:
obtaining a target domain current signal sample;
and performing iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample, the fault result label of the source domain current signal sample and the target domain current signal sample and combining a Maximum Mean Difference (MMD) distance measurement method to obtain a signal classification model meeting the precision condition.
Optionally, the step of iteratively training the initial feature extractor and the initial fault discriminator to obtain a signal classification model with a precision condition met based on the source domain current signal sample, the fault result label of the source domain current signal sample, and the target domain current signal sample in combination with a maximum mean difference MMD distance measurement method includes:
inputting source domain current signal samples in the semantic alignment group to the initial feature extractor, extracting first fault features of the source domain current signal samples based on the initial feature extractor, and determining predicted fault information based on the first fault features;
determining a fault classification loss based on the predicted fault information and the fault result tag;
inputting the first fault characteristic to the initial fault discriminator to obtain group countermeasure loss;
inputting the target domain current signal sample to the initial feature extractor to obtain a second fault feature of the target domain;
calculating the distance loss of the first fault feature and the second fault feature by adopting a Maximum Mean Difference (MMD) distance measurement function;
Calculating the sum of the fault classification loss, the group antagonism loss and the distance loss to obtain a target loss;
judging whether the target loss meets a loss standard indicated by a preset loss threshold range or not;
and if the target loss does not meet the loss standard indicated by the preset loss threshold range, returning to input the source domain current signal sample in the semantic alignment group to the initial feature extractor, extracting a first fault feature of the source domain current signal sample based on the initial feature extractor to obtain predicted fault information, and stopping training until the target loss meets the loss standard indicated by the preset loss threshold range to obtain a signal classification model meeting accuracy conditions.
Optionally, the step of inputting the target domain current signal sample to the initial feature extractor to obtain a second fault feature of the target domain includes:
performing fast Fourier transform on the target domain current signal sample to obtain a transformed target domain current signal sample;
normalizing the transformed target domain current signal sample to obtain a normalized target domain current signal sample;
And inputting the normalized target domain current signal sample to the initial feature extractor to obtain a second fault feature of the target domain.
Optionally, the signal classification model employs a lightweight convolutional network, wherein the lightweight convolutional network is constructed from a depth Depthwise convolution, a point-by-point Pointwise convolution, and a Group convolution.
The specific implementation manner of the storage medium is basically the same as the above embodiments of the fault diagnosis method of the motor current signal, and will not be repeated here.
The present application also provides a computer program product comprising a computer program which, when executed by a processor, implements the steps of the above-described fault diagnosis method of a motor current signal.
The specific implementation manner of the computer program product of the present application is basically the same as the above embodiments of the fault diagnosis method for the motor current signal, and will not be repeated here.
It should be noted that, in this document, 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.
The foregoing embodiment numbers of the present application are merely for describing, and do not represent advantages or disadvantages of the embodiments.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) as described above, including several instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method described in the embodiments of the present application.
The foregoing description is only of the preferred embodiments of the present application, and is not intended to limit the scope of the claims, and all equivalent structures or equivalent processes using the descriptions and drawings of the present application, or direct or indirect application in other related technical fields are included in the scope of the claims of the present application.
Claims (10)
1. A fault diagnosis method of a motor current signal, characterized in that the fault diagnosis method of a motor current signal comprises:
acquiring a motor current signal;
inputting the motor current signal into a preset signal classification model, and performing fault identification processing on the motor current signal based on the signal classification model to obtain fault information;
the signal classification model comprises a feature extractor and a fault discriminator, wherein the feature extractor and the fault discriminator are trained based on a semantic alignment packet countermeasure method.
2. The fault diagnosis method of motor current signal according to claim 1, wherein, before the step of acquiring motor current signal, the method comprises:
acquiring a source domain current signal sample and a fault result label of the source domain current signal sample;
based on the source domain current signal sample, pre-training to obtain an initial feature extractor;
grouping the source domain current signal samples into a preset first number of semantic alignment subgroups to obtain grouped source domain current signal samples;
pre-training to obtain an initial fault discriminator based on the grouped source domain current signal samples;
And performing iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample and a fault result label of the source domain current signal sample to obtain a signal classification model meeting the precision condition.
3. The method for fault diagnosis of motor current signals according to claim 2, wherein the step of iteratively training the initial feature extractor and the initial fault discriminator based on the fault result labels of the source domain current signal samples and the source domain current signal samples to obtain a signal classification model satisfying a precision condition comprises:
inputting source domain current signal samples in the semantic alignment group to the initial feature extractor, extracting first fault features of the source domain current signal samples based on the initial feature extractor, and determining predicted fault information based on the first fault features;
determining a fault classification loss based on the predicted fault information and the fault result tag;
inputting the first fault characteristic to the initial fault discriminator to obtain group countermeasure loss;
determining a target loss based on the fault classification loss and the group countermeasure loss;
Judging whether the target loss meets a loss standard indicated by a preset loss threshold range or not;
and if the target loss does not meet the loss standard indicated by the preset loss threshold range, returning to input the source domain current signal sample in the semantic alignment group to the initial feature extractor, extracting a first fault feature of the source domain current signal sample based on the initial feature extractor to obtain predicted fault information, and stopping training until the target loss meets the loss standard indicated by the preset loss threshold range to obtain a signal classification model meeting accuracy conditions.
4. The method for fault diagnosis of motor current signals according to claim 2, wherein the step of iteratively training the initial feature extractor and the initial fault discriminator based on the fault result labels of the source domain current signal samples and the source domain current signal samples to obtain a signal classification model satisfying a precision condition comprises:
obtaining a target domain current signal sample;
and performing iterative training on the initial feature extractor and the initial fault discriminator based on the source domain current signal sample, the fault result label of the source domain current signal sample and the target domain current signal sample and combining a Maximum Mean Difference (MMD) distance measurement method to obtain a signal classification model meeting the precision condition.
5. The motor current signal fault diagnosis method according to claim 4, wherein the step of iteratively training the initial feature extractor and the initial fault discriminator based on the source domain current signal sample, the fault result label of the source domain current signal sample, and the target domain current signal sample in combination with a maximum mean difference MMD distance metric method to obtain a signal classification model satisfying a precision condition comprises:
inputting source domain current signal samples in the semantic alignment group to the initial feature extractor, extracting first fault features of the source domain current signal samples based on the initial feature extractor, and determining predicted fault information based on the first fault features;
determining a fault classification loss based on the predicted fault information and the fault result tag;
inputting the first fault characteristic to the initial fault discriminator to obtain group countermeasure loss;
inputting the target domain current signal sample to the initial feature extractor to obtain a second fault feature of the target domain;
calculating the distance loss of the first fault feature and the second fault feature by adopting a Maximum Mean Difference (MMD) distance measurement function;
Calculating the sum of the fault classification loss, the group antagonism loss and the distance loss to obtain a target loss;
judging whether the target loss meets a loss standard indicated by a preset loss threshold range or not;
and if the target loss does not meet the loss standard indicated by the preset loss threshold range, returning to input the source domain current signal sample in the semantic alignment group to the initial feature extractor, extracting a first fault feature of the source domain current signal sample based on the initial feature extractor to obtain predicted fault information, and stopping training until the target loss meets the loss standard indicated by the preset loss threshold range to obtain a signal classification model meeting accuracy conditions.
6. The motor current signal fault diagnosis method according to claim 5, wherein the step of inputting the target domain current signal sample to the initial feature extractor to obtain a second fault feature of a target domain comprises:
performing fast Fourier transform on the target domain current signal sample to obtain a transformed target domain current signal sample;
normalizing the transformed target domain current signal sample to obtain a normalized target domain current signal sample;
And inputting the normalized target domain current signal sample to the initial feature extractor to obtain a second fault feature of the target domain.
7. The fault diagnosis method of motor current signals according to claim 1, wherein the signal classification model employs a lightweight convolution network, wherein the lightweight convolution network is constructed of a depth Depthwise convolution, a point-wise Pointwise convolution, and a Group-wise convolution.
8. A fault diagnosis device of a motor current signal, characterized in that the fault diagnosis device of a motor current signal comprises:
the acquisition module is used for acquiring a motor current signal;
the recognition module is used for inputting the motor current signal into a preset signal classification model, and performing fault recognition processing on the motor current signal based on the signal classification model to obtain fault information; the signal classification model comprises a feature extractor and a fault discriminator, wherein the feature extractor and the fault discriminator are trained based on a semantic alignment packet countermeasure method.
9. A fault diagnosis apparatus of a motor current signal, characterized in that the fault diagnosis apparatus of a motor current signal comprises: a memory, a processor, and a program stored on the memory for implementing a fault diagnosis method of the motor current signal,
The memory is used for storing a program for realizing a fault diagnosis method of the motor current signal;
the processor is configured to execute a program for implementing the motor current signal fault diagnosis method to implement the motor current signal fault diagnosis method steps of any one of claims 1 to 7.
10. A storage medium, characterized in that the storage medium has stored thereon a program realizing the failure diagnosis method of the motor current signal, the program realizing the failure diagnosis method of the motor current signal being executed by a processor to realize the steps of the failure diagnosis method of the motor current signal according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310504843.4A CN116561655A (en) | 2023-05-06 | 2023-05-06 | Fault diagnosis method, device, equipment and storage medium for motor current signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310504843.4A CN116561655A (en) | 2023-05-06 | 2023-05-06 | Fault diagnosis method, device, equipment and storage medium for motor current signal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116561655A true CN116561655A (en) | 2023-08-08 |
Family
ID=87497617
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310504843.4A Pending CN116561655A (en) | 2023-05-06 | 2023-05-06 | Fault diagnosis method, device, equipment and storage medium for motor current signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116561655A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117949823A (en) * | 2024-03-27 | 2024-04-30 | 南京信息工程大学 | Motor fault diagnosis method and device based on improved transfer learning model |
-
2023
- 2023-05-06 CN CN202310504843.4A patent/CN116561655A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117949823A (en) * | 2024-03-27 | 2024-04-30 | 南京信息工程大学 | Motor fault diagnosis method and device based on improved transfer learning model |
| CN117949823B (en) * | 2024-03-27 | 2024-05-31 | 南京信息工程大学 | Motor fault diagnosis method and device based on improved transfer learning model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107253196B (en) | Mechanical arm collision detection method, device, equipment and storage medium | |
| CN110737798B (en) | Indoor inspection method and related product | |
| CN116561655A (en) | Fault diagnosis method, device, equipment and storage medium for motor current signal | |
| CN110704661A (en) | Image classification method and device | |
| CN114595124B (en) | Time sequence abnormity detection model evaluation method, related device and storage medium | |
| KR102476679B1 (en) | Apparatus and method for object detection | |
| US20220155338A1 (en) | Method for testing sensing effect, moving apparatus, electronic device, storage medium, and system for testing sensing effect | |
| KR101228336B1 (en) | Personalization Service Providing Method by Using Mobile Terminal User's Activity Pattern and Mobile Terminal therefor | |
| CN111382270A (en) | Intention recognition method, device and equipment based on text classifier and storage medium | |
| CN112001948A (en) | Target tracking processing method and device | |
| CN108595013B (en) | Holding recognition method and device, storage medium and electronic equipment | |
| CN109726726B (en) | Event detection method and device in video | |
| CN115841575A (en) | Key point detection method, device, electronic apparatus, storage medium, and program product | |
| CN114612531A (en) | Image processing method and device, electronic equipment and storage medium | |
| CN112859136B (en) | Positioning method and related device | |
| CN112400147A (en) | Algorithm configuration method, equipment and system and movable platform | |
| CN110909804B (en) | Method, device, server and storage medium for detecting abnormal data of base station | |
| CN107423515B (en) | Mechanical arm friction identification method, device, equipment and storage medium | |
| WO2024012367A1 (en) | Visual-target tracking method and apparatus, and device and storage medium | |
| CN115474108B (en) | Event monitoring system and method based on edge calculation | |
| CN115629930A (en) | Fault detection method, device and equipment based on DSP system and storage medium | |
| CN115937970A (en) | Hand key point identification method, device, equipment and storage medium | |
| CN112632222B (en) | Terminal equipment and method for determining data belonging field | |
| CN114169258A (en) | Flow measuring method, device, equipment and storage medium for open channel | |
| CN116450384A (en) | Information processing method and related device |
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 |