CN115471453A - Power equipment fault diagnosis method based on image processing - Google Patents

Power equipment fault diagnosis method based on image processing Download PDF

Info

Publication number
CN115471453A
CN115471453A CN202210987204.3A CN202210987204A CN115471453A CN 115471453 A CN115471453 A CN 115471453A CN 202210987204 A CN202210987204 A CN 202210987204A CN 115471453 A CN115471453 A CN 115471453A
Authority
CN
China
Prior art keywords
power equipment
detection model
convolution
fault detection
image
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
Application number
CN202210987204.3A
Other languages
Chinese (zh)
Inventor
陈运蓬
赵锐
尚文
逯建林
张碧龙
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.)
Datong Power Supply Co of State Grid Shanxi Electric Power Co Ltd
Original Assignee
Datong Power Supply Co of State Grid Shanxi Electric Power Co Ltd
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 Datong Power Supply Co of State Grid Shanxi Electric Power Co Ltd filed Critical Datong Power Supply Co of State Grid Shanxi Electric Power Co Ltd
Priority to CN202210987204.3A priority Critical patent/CN115471453A/en
Publication of CN115471453A publication Critical patent/CN115471453A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/0002Inspection of images, e.g. flaw detection
    • G06T7/0004Industrial image inspection
    • G06T7/0008Industrial image inspection checking presence/absence
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/08Learning methods
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/10Image acquisition modality
    • G06T2207/10048Infrared image
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/20Special algorithmic details
    • G06T2207/20081Training; Learning
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/20Special algorithmic details
    • G06T2207/20084Artificial neural networks [ANN]
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30108Industrial image inspection
    • G06T2207/30164Workpiece; Machine component

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Quality & Reliability (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Image Analysis (AREA)

Abstract

The application discloses a power equipment fault diagnosis method based on image processing, which comprises the following steps: acquiring an infrared induction image of the power equipment to be detected, processing a gray value, and recording the processed gray image as a sampling image; training an initial fault detection model based on a convolutional neural network and sample data, and recording the converged initial fault detection model as a power equipment fault detection model when the initial fault detection model is judged to be converged, wherein the initial fault detection model comprises two convolutional links connected by a full-connection layer, any convolutional link consists of a plurality of convolutional layers which are sequentially connected, and the convolutional layers with odd numbers are connected through a first characteristic function; and carrying out fault identification on the sampled image according to the fault detection model of the power equipment, and carrying out fault diagnosis on the power equipment to be detected according to the fault identification result. Through the technical scheme in this application, help improving power equipment fault diagnosis's accuracy and efficiency.

Description

Power equipment fault diagnosis method based on image processing
Technical Field
The application relates to the technical field of power equipment fault diagnosis, in particular to a power equipment fault diagnosis method based on image processing.
Background
The power equipment has a complex structure and large general type and quantity, and generally has faults due to various reasons in the operation process, so that the fault diagnosis of the power equipment at regular and irregular intervals is required, and the safe and stable operation of the power equipment is ensured.
The manual inspection of the power equipment is an effective fault diagnosis method generally, and an inspector judges whether the power equipment has a fault or has a fault hidden trouble by observing the running condition and running sound of the equipment, and the inspection method generally needs the inspection worker to have richer working experience.
An important phenomenon representation of the power equipment fault is equipment heating, so that a fault diagnosis mode based on an infrared temperature measurement technology and an image processing technology is introduced into inspection of the power equipment to identify the power equipment with abnormal temperature, and the purpose of fault diagnosis of the power equipment is achieved.
In the prior art, the fault diagnosis mode based on the infrared temperature measurement technology and the image processing technology has the main problems that: in order to be carried by inspection personnel conveniently, the size and the weight of the equipment are usually small, so that the data processing performance of related equipment is limited, the fault diagnosis precision cannot be guaranteed, and the efficiency is low.
Therefore, how to ensure the fault diagnosis precision of the power equipment and improve the fault diagnosis efficiency of the power equipment on the premise of not increasing the size and the cost of the equipment becomes a technical problem which needs to be solved by technical research and development personnel in the field.
Disclosure of Invention
The purpose of this application lies in: how to improve the accuracy and efficiency of the fault diagnosis of the power equipment.
The technical scheme of the application is as follows: provided is a power equipment fault diagnosis method based on image processing, which comprises the following steps: step 1, acquiring an infrared induction image of power equipment to be detected, performing gray value processing on the infrared induction image, and recording the processed gray image as a sampling image; step 2, training an initial fault detection model based on a convolutional neural network and sample data, and when the initial fault detection model is judged to be converged, recording the converged initial fault detection model as a power equipment fault detection model, wherein the initial fault detection model comprises two convolutional links connected by a full-connection layer, any convolutional link consists of a plurality of convolutional layers which are sequentially connected, the convolutional layers with odd numbers are connected through a first characteristic function, and the calculation formula of the first characteristic function is as follows:
Figure BDA0003802298220000021
wherein A is an input matrix of the current convolutional layer, Z is a random matrix, Y is an output matrix output to the next convolutional layer, beta is a correction coefficient, and Q (Y) is a characteristic function;
and 3, carrying out fault identification on the sampled image according to the fault detection model of the power equipment, and carrying out fault diagnosis on the power equipment to be detected according to the fault identification result.
In any of the above technical solutions, further, the two convolutional links include a main convolutional link and an auxiliary convolutional link, and convolutional layers numbered even in the auxiliary convolutional link are connected to convolutional layers numbered even in the main convolutional link through a second feature function.
In any of the above technical solutions, further, the calculation formula of the second feature function is:
L D =γE x [(D(x1,x2)-1) 2 ]+W
wherein γ is a proportionality coefficient, E x [·]For the fluctuation calculation function, D (-) is the variation calculation function, x1 is the output matrix element of the current convolution layer in the main convolution link, x2 is the output matrix element of the current convolution layer in the auxiliary convolution link, W is the weight matrix, L is the weight matrix D Is the second characteristic function.
In any of the above technical solutions, further, a pooling layer is disposed between two adjacent convolution layers in the auxiliary convolution link.
In any one of the above technical solutions, further, in step 1, the method further includes: and preprocessing the infrared induction image according to the size of a preset image, wherein the preprocessing at least comprises cutting and amplifying.
In any one of the above technical solutions, further, in step 2, training the initial fault detection model based on the convolutional neural network and the sample data specifically includes: step 21, extracting the characteristics of the sample image based on an attention mechanism to generate a characteristic diagram; step 22, matrixing the characteristic diagram, multiplying a first characteristic matrix obtained by matrixing by a random initialization matrix, and recording a calculation result as a second characteristic matrix; and 23, training the initial fault detection model based on the second feature matrix, adjusting the random initialization matrix according to the training result when the training result is judged not to be converged, and repeatedly executing the step 22 until the initial fault detection model is converged.
The beneficial effect of this application is:
according to the technical scheme, two convolution links are arranged, and two adjacent odd-numbered convolution layers in any link are connected through a first characteristic function, so that the extraction characteristics of the convolution layers are enhanced, and the efficiency of fault diagnosis is improved. Meanwhile, the convolution layers with the even numbers in the auxiliary convolution link are connected with the convolution layers with the even numbers in the main convolution link through the second characteristic function, so that the data processing efficiency of the power equipment fault detection model can be improved, and the fault diagnosis efficiency is further improved.
Drawings
The advantages of the above and/or additional aspects of the present application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
fig. 1 is a schematic flow diagram of a power equipment fault diagnosis method based on image processing according to an embodiment of the present application;
FIG. 2 is a schematic diagram of an initial fault detection model according to one embodiment of the present application.
Detailed Description
In order that the above objects, features and advantages of the present application can be more clearly understood, the present application will be described in further detail with reference to the accompanying drawings and detailed description. It should be noted that the embodiments and features of the embodiments of the present application may be combined with each other without conflict.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application, however, the present application may be practiced in other ways than those described herein, and therefore the scope of the present application is not limited by the specific embodiments disclosed below.
As shown in fig. 1, the present embodiment provides an image processing-based power equipment fault diagnosis method, including:
step 1, acquiring an infrared induction image of power equipment to be detected, performing gray value processing on the infrared induction image, and recording the processed gray image as a sampling image;
specifically, the portable infrared equipment of personnel's accessible of patrolling and examining acquires the infrared induction image of waiting to detect power equipment, transmits modes such as wired, bluetooth, WIFI that will acquire to portable data processing equipment in, is handled infrared induction image by this data processing equipment. Before fault diagnosis is carried out, a temperature fitting mode is adopted to carry out gray value processing on the infrared induction image, fitting is carried out to obtain a corresponding gray image, and specific fitting process is not repeated.
Further, in step 1, the method further comprises: and preprocessing the infrared induction image according to the size of a preset image, wherein the preprocessing at least comprises cutting and amplifying.
Specifically, in the process of acquiring the infrared induction image of the power equipment to be detected, the acquired infrared induction image is inconsistent due to different factors such as height, distance and angle, so that the acquired infrared induction image needs to be zoomed, translated, rotated and the like, the image including the power equipment, which is identified by using an image processing technology, is locally cut so as to be cut into images with the same size, the cut images are labeled, such as information of equipment types and positions, and then the labeled images are subjected to gray value processing.
And 2, training the initial fault detection model based on the convolutional neural network and sample data, and recording the converged initial fault detection model as the power equipment fault detection model when the initial fault detection model is judged to be converged.
In this embodiment, an implementation manner of an initial fault detection model is shown, as shown in fig. 2, the initial fault detection model includes two convolutional links connected by a full connection layer FC, where any one of the convolutional links is composed of a plurality of sequentially connected convolutional layers Conv, and convolutional layers numbered as odd numbers are connected by a first characteristic function, and a calculation formula of the first characteristic function is:
Figure BDA0003802298220000041
wherein A is the input matrix of the current convolutional layer, A T Is the transpose of the input matrix A, Z is the random matrix, Y is the output matrix to the next convolution layer, β is the correction factor, and Q (Y) is the eigenfunction.
Specifically, the number of convolution layers may be set to 6 or 8, taking 6 layers as an example, the convolution layers of 1 st, 3 rd and 5 th layers are connected by a first characteristic function, that is, an output of the 1 st convolution layer is connected to an input of the 3 rd convolution layer, and an output of the 3 rd convolution layer is connected to an input of the 5 th convolution layer, where a calculation formula of the first characteristic function is:
Figure BDA0003802298220000042
where A is the input matrix of the current convolutional layer, A T Is the transpose of the input matrix A, Z is the random matrix, Y is the output matrix to the next convolution layer, β is the correction coefficient, and Q (Y) is the eigen function.
In this embodiment, the sample data includes a training set and a test set, and the types of data in the two data sets are the same, and include corresponding grayscale images and corresponding fault types, such as faults, hidden dangers, and non-faults, where the faults and hidden dangers may be further divided according to actual requirements, such as first-level hidden dangers, second-level hidden dangers, and third-level hidden dangers.
And inputting the training set into the initial fault detection model, matching the prediction result of the model with the corresponding fault type, further adjusting the model parameters in the initial fault detection model, testing the initial fault detection model after the model parameters are adjusted by using the test set, and repeating the process until the initial fault detection model is converged to obtain the equipment fault detection model.
In the initial fault detection model, the output of the 1 st and 3 rd convolutional layers is calculated by using a first characteristic function, and when the characteristic function Q (Y) is calculated to be minimum, the output matrix Y corresponding to the matrix A is input to enhance the characteristics of the sample data extracted from the 1 st and 3 rd convolutional layers, and the extracted characteristics (output matrix Y) are input to the 3 rd and 5 th convolutional layers, so that the image characteristics representing the temperature in the gray level image are enhanced.
Further, the two convolution links include a main convolution link and an auxiliary convolution link, convolution layers with even numbers in the auxiliary convolution link are connected with convolution layers with even numbers in the main convolution link through a second characteristic function, wherein a calculation formula of the second characteristic function is as follows:
L D =γE x [(D(x1,x2)-1) 2 ]+W
wherein γ is a proportionality coefficient, E x [·]For the fluctuation calculation function, D (-) is the variation calculation function, x1 is the output matrix element of the current convolution layer in the main convolution link, x2 is the output matrix element of the current convolution layer in the auxiliary convolution link, W is the weight matrix, L is the weight matrix D Is the second characteristic function.
Specifically, in this embodiment, a change calculation function D (-) is introduced, matrix elements output by the current convolutional layer in the auxiliary convolutional link and the main convolutional link are calculated, and the change of the auxiliary convolutional link relative to the main convolutional link is calculated, where the value range is [ -1,1], that is, the more obvious the difference between the two is, the larger the value of D (-) is. And then the calculated result is input to the corresponding convolution layer in the main convolution link as the input quantity of the second characteristic function so as to improve the overall data processing efficiency of the convolution layer.
It should be noted that the scaling coefficient γ and the weight matrix W are adjustment values, and when it is determined that the initial fault detection model is not converged, the scaling coefficient γ and the weight matrix W may be adjusted, and the adjusted initial fault detection model may be retrained until the initial fault detection model can be converged.
Preferably, a pooling layer pool is disposed between two adjacent convolution layers in the auxiliary convolution link.
It should be noted that, after the pooling layer operation, in order to make the output results of each convolution layer in the main convolution link and the auxiliary convolution link have the same dimension, the corresponding output results of the convolution layer in the auxiliary convolution link need to be subjected to dimension expansion processing based on the original data, that is, the inverse process of the pooling process. For pooling, the data in a certain area is reduced from 2 × 2 dimensions to a single data, and for expanding dimensions, the data is expanded to 2 × 2 dimensions based on the single data.
Further, in step 2, training the initial fault detection model based on the convolutional neural network and the sample data specifically includes:
step 21, based on an attention mechanism, performing feature extraction on the sample image to generate a feature map;
specifically, in this embodiment, based on an attention mechanism in the neural network, such as a hierarchical attention model, the features of the grayscale image in the sample image are extracted to obtain feature information representing the temperature in the image, and a corresponding feature map is generated, which is not described in detail again in the specific process.
And step 22, performing matrixing on the characteristic diagram, multiplying a first characteristic matrix obtained by matrixing by a random initialization matrix, and recording a calculation result as a second characteristic matrix, wherein the random initialization matrix is mainly used as an adjusting matrix of model convergence. And if the model cannot be converged by adjusting the random initialization matrix, adjusting a second characteristic function in the initial fault detection model, and training again until the model is converged.
And 23, training the initial fault detection model based on the second feature matrix, adjusting the random initialization matrix according to the training result when the training result is judged not to be converged, and repeatedly executing the step 22 until the initial fault detection model is converged.
Specifically, 80% of the sample data is used as a training set, and 20% is used as a test set. The method comprises the steps of firstly training an initial fault detection model by using a training set, testing by using a test set when the training result is judged to be converged, and taking the model as a power equipment fault detection model to perform fault identification on a sampling image if the accuracy of the test result meets the requirement. And if the test result does not meet the requirement, repeating the training and testing processes until the accuracy of the test result meets the requirement.
In this embodiment, whether the initial fault detection model is converged is mainly determined by a convergence function, and the implementation manner of the convergence function is not limited in this embodiment, and the following calculation formula of the convergence function may be adopted:
E z [log(1-D(X,G(z))]→0
in the formula, E z [·]For the convergence function, G (z) is the training result and X is the labeling result in the sample data. With the change computation function D (·), the transformation between the training result G (z) and the marking result (fault type) in the sample data, i.e. the difference between the two, is computed. Then, when the convergence function E is determined z [·]And when the initial fault detection model approaches 0, the convergence of the initial fault detection model is proved.
And 3, carrying out fault identification on the sampled image according to the fault detection model of the power equipment, carrying out fault diagnosis on the power equipment to be detected according to the fault identification result, and outputting a corresponding fault type.
The technical scheme of the present application is described in detail above with reference to the accompanying drawings, and the present application proposes an image processing-based power equipment fault diagnosis method, which includes: step 1, acquiring an infrared induction image of electric equipment to be detected, performing gray value processing on the infrared induction image, and recording the processed gray image as a sampling image; step 2, training an initial fault detection model based on a convolutional neural network and sample data, and recording the converged initial fault detection model as a power equipment fault detection model when the initial fault detection model is judged to be converged, wherein the initial fault detection model comprises two convolutional links connected by a full-connection layer, any convolutional link consists of a plurality of sequentially connected convolutional layers, and the convolutional layers with odd numbers are connected through a first characteristic function; and 3, carrying out fault identification on the sampled image according to the fault detection model of the power equipment, and carrying out fault diagnosis on the power equipment to be detected according to the fault identification result. Through the technical scheme in this application, help improving power equipment fault diagnosis's accuracy and efficiency.
The steps in the present application may be sequentially adjusted, combined, and subtracted according to actual requirements.
The units in the device can be merged, divided and deleted according to actual requirements.
Although the present application has been disclosed in detail with reference to the accompanying drawings, it is to be understood that such description is merely illustrative and not restrictive of the application of the present application. The scope of the present application is defined by the appended claims and may include various modifications, adaptations, and equivalents of the subject invention without departing from the scope and spirit of the present application.

Claims (6)

1. An image processing-based power equipment fault diagnosis method is characterized by comprising the following steps:
step 1, acquiring an infrared induction image of power equipment to be detected, performing gray value processing on the infrared induction image, and recording the processed gray image as a sampling image;
step 2, training an initial fault detection model based on a convolutional neural network and sample data, recording the converged initial fault detection model as the power equipment fault detection model when the initial fault detection model is judged to be converged,
the initial fault detection model comprises two convolution links connected by a full-connection layer, wherein any convolution link is composed of a plurality of convolution layers connected in sequence, the convolution layers with odd numbers are connected by a first characteristic function, and the calculation formula of the first characteristic function is as follows:
Figure FDA0003802298210000011
wherein A is an input matrix of a current convolution layer, and Y is an output matrix output to a next convolution layer;
and 3, carrying out fault identification on the sampling image according to the power equipment fault detection model, and carrying out fault diagnosis on the power equipment to be detected according to the fault identification result.
2. The image-processing-based power equipment fault diagnosis method as claimed in claim 1, wherein the two convolution links include a main convolution link and an auxiliary convolution link, and the convolution layer with the even number in the auxiliary convolution link is connected with the convolution layer with the even number in the main convolution link through a second characteristic function.
3. The image processing-based power equipment fault diagnosis method according to claim 2, wherein the calculation formula of the second characteristic function is:
L D =γE x [(D(x1,x2)-1) 2 ]+W
wherein γ is a proportionality coefficient, E x [·]And D (-) is a fluctuation calculation function, x1 is an output matrix element of the current convolutional layer in the main convolutional link, x2 is an output matrix element of the current convolutional layer in the auxiliary convolutional link, and W is a weight matrix.
4. The image processing-based power equipment fault diagnosis method according to claim 2 or 3, wherein a pooling layer is provided between two adjacent convolution layers in the auxiliary convolution link.
5. The method for diagnosing the fault of the power equipment based on the image processing as claimed in claim 1, wherein the step 1 further comprises:
and preprocessing the infrared induction image according to the size of a preset image, wherein the preprocessing at least comprises cutting and amplifying.
6. The method according to claim 1, wherein in the step 2, the training of the initial fault detection model based on the convolutional neural network and the sample data specifically includes:
step 21, extracting features of the sample image based on an attention mechanism to generate a feature map;
step 22, performing matrixing on the characteristic diagram, multiplying a first characteristic matrix obtained by matrixing by a random initialization matrix, and recording a calculation result as a second characteristic matrix;
and step 23, training the initial fault detection model based on the second feature matrix, when the training result is judged not to be converged, adjusting the random initialization matrix according to the training result, and repeatedly executing the step 22 until the initial fault detection model is converged.
CN202210987204.3A 2022-08-17 2022-08-17 Power equipment fault diagnosis method based on image processing Pending CN115471453A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210987204.3A CN115471453A (en) 2022-08-17 2022-08-17 Power equipment fault diagnosis method based on image processing

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210987204.3A CN115471453A (en) 2022-08-17 2022-08-17 Power equipment fault diagnosis method based on image processing

Publications (1)

Publication Number Publication Date
CN115471453A true CN115471453A (en) 2022-12-13

Family

ID=84365927

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210987204.3A Pending CN115471453A (en) 2022-08-17 2022-08-17 Power equipment fault diagnosis method based on image processing

Country Status (1)

Country Link
CN (1) CN115471453A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116010896A (en) * 2023-02-03 2023-04-25 南京南瑞继保电气有限公司 Wind driven generator fault diagnosis method based on countermeasure training and transducer

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116010896A (en) * 2023-02-03 2023-04-25 南京南瑞继保电气有限公司 Wind driven generator fault diagnosis method based on countermeasure training and transducer

Similar Documents

Publication Publication Date Title
WO2022077901A1 (en) Bearing failure mode diagnosis method using small sample data sets, and system
CN109164343B (en) Transformer fault diagnosis method based on characteristic information quantization and weighted KNN
CN111340238A (en) Fault diagnosis method, device, equipment and storage medium of industrial system
CN107505133A (en) The probability intelligent diagnosing method of rolling bearing fault based on adaptive M RVM
CN112034310A (en) Partial discharge defect diagnosis method and system for combined electrical appliance
CN112200244A (en) Intelligent detection method for anomaly of aerospace engine based on hierarchical countermeasure training
CN109858352A (en) A kind of method for diagnosing faults based on compressed sensing and the multiple dimensioned network of improvement
CN111860775B (en) Ship fault real-time diagnosis method based on CNN and RNN fusion
CN112465124B (en) Twin depth space-time neural network model acquisition/fault diagnosis method and device
CN111833310B (en) Surface defect classification method based on neural network architecture search
CN113065223B (en) Multi-level probability correction method for digital twin model of tower mast cluster
CN115471453A (en) Power equipment fault diagnosis method based on image processing
CN114022715A (en) Method and device for detecting lead sealing defect of high-voltage cable terminal
CN111598854B (en) Segmentation method for small defects of complex textures based on rich robust convolution feature model
CN114357372A (en) Aircraft fault diagnosis model generation method based on multi-sensor data driving
CN117076935B (en) Digital twin-assisted mechanical fault data lightweight generation method and system
CN117315365A (en) Camshaft surface damage detecting system based on visual analysis
CN112802011A (en) Fan blade defect detection method based on VGG-BLS
CN116662800A (en) Rolling bearing fault diagnosis method based on self-adaptive attention mechanism
CN111428772A (en) Photovoltaic system depth anomaly detection method based on k-nearest neighbor adaptive voting
CN113537010B (en) Fifteen-phase asynchronous motor rolling bearing fault diagnosis method based on single-channel diagram data enhancement and migration training residual error network
CN115659271A (en) Sensor abnormality detection method, model training method, system, device, and medium
CN115311238A (en) Photovoltaic module defect damage identification method based on image analysis
CN111753876B (en) Product quality detection method based on deep neural network
CN111967507A (en) Discrete cosine transform and U-Net based time sequence anomaly detection 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