Power equipment fault detection method based on MAE self-encoding model

Abstract

The invention provides a power equipment fault detection method based on an MAE self-encoding model, and relates to the technical field of power equipment fault detection, and the method comprises the steps: collecting an image of power equipment, and building an image data set; performing data preprocessing on the image data set, and marking the preprocessed image data set according to the image type; performing mask processing on the image data set, and inputting the image data set into an MAE auto-encoder model for self-supervised training; accessing a multi-layer perceptron into the trained MAE auto-encoder model, and training through a weighted loss function based on the marked image data set to obtain a power equipment fault detection model; inputting the to-be-detected image into the power equipment fault detection model to obtain a fault detection result of the power equipment; according to the method, the influence on the performance of the power equipment defect detection model caused by the problems of data annotation missing and class imbalance is solved, and the generalization ability of the power equipment defect detection model in real application is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment fault detection, and in particular to a power equipment fault detection method based on a MAE autoencoder model. Background Art

[0002] With the continuous expansion of the scale of power systems, defect detection and early warning technologies have become increasingly important. The stable operation of power systems is directly related to the normal operation of various fields of society, and the timely detection and location of faults can effectively reduce the downtime and losses of the system. Traditional power equipment fault detection methods mostly rely on manual experience and rule-based models, which often have great limitations in complex power networks. In recent years, deep learning technology has gradually become a research hotspot for power equipment fault detection with its powerful data processing and feature extraction capabilities. The application of deep learning in power equipment fault detection is mainly reflected in fault identification, fault diagnosis and prediction by analyzing the historical operation data and monitoring signals of power equipment. This type of method can effectively and automatically mine potential fault modes from large-scale data, and has higher accuracy and adaptability than traditional methods. Especially in terms of real-time and efficiency, deep learning models can predict possible faults in power equipment in advance and respond through real-time monitoring and automated learning, thereby improving the stability and safety of the system.

[0003] However, although deep learning technology has made some progress in power equipment fault detection, existing models still face many problems. One of the most notable problems is the small scale of labeled data, especially the large difference in the ratio between fault data and normal operation data. Since fault events are usually rare and the deployment environment of power equipment is extremely complex, the amount of data in the fault category in the training data is much lower than that in the normal data category. This data imbalance makes the existing deep learning model prone to misjudge the judgment results of the normal state as normal, resulting in poor sensitivity of fault detection and significant impact on the detection ability of the model. Summary of the invention

[0004] The purpose of the present invention is to provide a method for power equipment fault detection based on the MAE autoencoder model, which can solve the impact of missing data annotation problems and category imbalance problems on the performance of the power equipment defect detection model, and improve the generalization ability of the power equipment defect detection model in real applications.

[0005] The technical solution of the present invention is:

[0006] In a first aspect, the present application provides a method for detecting faults in power equipment based on a MAE autoencoder model, which comprises the following steps:

[0007] S1. Collect images of power equipment and establish an image dataset;

[0008] S2, preprocessing the image data set, and labeling the preprocessed image data set according to the image type;

[0009] S3, the image dataset is processed by mask and then input into the MAE autoencoder model for self-supervised training;

[0010] S4, a multi-layer perceptron is connected to the trained MAE autoencoder model, and a power equipment fault detection model is obtained by training the model through a weighted loss function based on the labeled image dataset;

[0011] S5. Input the image to be detected into the electric power equipment fault detection model to obtain the fault detection result of the electric power equipment.

[0012] Furthermore, in step S2, the above data preprocessing includes data denoising and data enhancement, and its expression is:

[0013] x denoised =Filter(x)

[0014] x aug =λ·x denoised +(1-λ)·Augment(x denoised )

[0015] In the formula, x denoised is the denoised image, Filter(·) is the median filter function, x is the original image in the dataset, and x aug is the image after data enhancement, λ represents the ratio of fusion with the original image, and Augment is the enhancement function.

[0016] Furthermore, in step S3, the mask processing includes: randomly masking a part of the input image to form a mask image, which is expressed as:

[0017] x masked =x aug ⊙M

[0018] In the formula, x masked is the mask image, x aug is the image after data enhancement, M is the mask matrix, and ⊙ represents element-by-element multiplication.

[0019] Furthermore, in step S3, the expression of the above MAE autoencoder model is:

[0020] x r =f d (f e (x masked ))

[0021] In the formula, x r is the image reconstructed by encoding and decoding the mask image through the autoencoder, f d (·) is the decoding reconstruction network, f e (·) is the encoding network, x masked is the mask image.

[0022] Furthermore, the self-supervised training of the above MAE autoencoder model includes:

[0023] The MAE autoencoder model is trained using mean square error as the loss function, and the weights of the MAE autoencoder model are optimized by the back propagation algorithm to minimize the loss function. At the same time, self-supervised training is achieved by maximizing the similarity between the output image of the MAE model and the image before masking. The expression of the loss function is:

[0024]

[0025] In the formula, L is the loss function, N is the number of samples, and x r is the image reconstructed by encoding and decoding the mask image through the autoencoder, x i is the original image.

[0026] Furthermore, in step S4, the expression of the weighted loss function is:

[0027]

[0028] In the formula, L' is the weighted loss function, m is the total number of samples, and m i is the number of the i-th category, For category y i The weight of y i is the category annotated in the image dataset, The actual result output by the model.

[0029] In a second aspect, the present application provides an electronic device, including:

[0030] A memory for storing one or more programs;

[0031] processor;

[0032] When the one or more programs are executed by the processor, a method for detecting faults of electric power equipment based on a MAE autoencoder model as described in any one of the first aspects is implemented.

[0033] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for detecting faults in electric power equipment based on a MAE autoencoder model as described in any one of the first aspects above.

[0034] Compared with the prior art, the present invention has at least the following advantages or beneficial effects:

[0035] The present invention discloses a method for detecting faults of electric equipment based on a MAE autoencoder model. The MAE autoencoder model is adopted to learn and extract electric equipment image features with strong generalization ability from a large number of unlabeled on-site pictures of electric equipment. The MAE autoencoder model is trained by self-supervision, and a multi-layer perceptron is introduced into the encoding part of the MAE autoencoder model for classification. On this basis, the electric equipment fault detection model is trained based on a small amount of labeled data through a weighted loss function, thereby reducing the influence of data labeling missing and category imbalance problems on the performance of the electric equipment defect detection model, significantly improving the sensitivity and detection ability of the electric equipment defect detection model, and enhancing the generalization ability of the electric equipment defect detection model in real applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0037] Figure 1 A step diagram of a method for detecting faults of electric power equipment based on a MAE autoencoder model according to the present invention;

[0038] Figure 2 The figure is a schematic structural block diagram of an electronic device according to an embodiment of the present invention.

[0039] Icon: 101, memory; 102, processor; 103, communication interface. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0041] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0042] It should be noted that, in this article, the term "comprises" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprises..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0043] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0044] Example 1

[0045] See also Figure 1 , Figure 1 The figure shows a step diagram of a method for detecting faults of electric power equipment based on a MAE autoencoder model provided in an embodiment of the present application.

[0046] In a first aspect, the present application provides a method for detecting faults in power equipment based on a MAE autoencoder model, which comprises the following steps:

[0047] S1. Collect images of power equipment and establish an image dataset;

[0048] S2, preprocessing the image data set, and labeling the preprocessed image data set according to the image type;

[0049] S3, the image dataset is processed by mask and then input into the MAE autoencoder model for self-supervised training;

[0050] S4, a multi-layer perceptron is connected to the trained MAE autoencoder model, and a power equipment fault detection model is obtained by training the model through a weighted loss function based on the labeled image dataset;

[0051] S5. Input the image to be detected into the electric power equipment fault detection model to obtain the fault detection result of the electric power equipment.

[0052] It should be noted that when labeling according to image type, typical images are selected for labeling mainly through manual on-site collection. Typical images are mainly images that are most obviously distinguished from other images. After the MAE autoencoder model is trained, the encoder network in the model can obtain high-quality image feature representation, and then the weight of the encoder network is fixed. The output of the encoder network is input into a classification network composed of a 3-layer perceptron. The classification network is trained based on a small amount of labeled image data sets, and the weighted loss function is used to enhance the model's learning ability for minority categories, assigning higher weights to minority categories to ensure that the model can adapt to specific defect detection tasks.

[0053] As a preferred implementation, in step S2, data preprocessing includes data denoising and data enhancement, and the expression is:

[0054] x denoised =Filter(x)

[0055] x aug =λ·x denoised +(1-λ)·Augment(x denoised )

[0056] In the formula, x denoised is the denoised image, Filter(·) is the median filter function, x is the original image in the dataset, and x aug is the image after data enhancement, λ represents the ratio of fusion with the original image, and Aufment is the enhancement function.

[0057] Among them, the data preprocessing process also includes operations such as normalization.

[0058] As a preferred implementation, in step S3, the mask processing includes: randomly covering a portion of the input image to form a mask image, which is expressed as:

[0059] x masked =x aug ⊙M

[0060] In the formula, x masked is the mask image, x aug is the image after data enhancement, M is the mask matrix, and ⊙ represents element-by-element multiplication.

[0061] As a preferred implementation, in step S3, the expression of the MAE autoencoder model is:

[0062] x r =f d (f e (x masked ))

[0063] In the formula, x r is the image reconstructed by encoding and decoding the mask image through the autoencoder, f d (·) is the decoding reconstruction network, f e (·) is the encoding network, x masked is the mask image.

[0064] As a preferred implementation, the self-supervised training of the MAE autoencoder model includes:

[0065] The MAE autoencoder model is trained using mean square error as the loss function, and the weights of the MAE autoencoder model are optimized by the back propagation algorithm to minimize the loss function. At the same time, self-supervised training is achieved by maximizing the similarity between the output image of the MAE model and the image before masking. The expression of the loss function is:

[0066]

[0067] In the formula, L is the loss function, N is the number of samples, and x r is the image reconstructed by encoding and decoding the mask image through the autoencoder, x i is the original image.

[0068] As a preferred implementation, in step S4, the expression of the weighted loss function is:

[0069]

[0070] In the formula, L' is the weighted loss function, m is the total number of samples, and m i is the number of the i-th category, For category y i The weight of y i is the category annotated in the image dataset, The actual result output by the model.

[0071] Example 2

[0072] See also Figure 2 , Figure 2 A schematic structural block diagram of an electronic device provided in an embodiment of the present application.

[0073] An electronic device includes a memory 101, a processor 102 and a communication interface 103, wherein the memory 101, the processor 102 and the communication interface 103 are electrically connected to each other directly or indirectly to realize data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines. The memory 101 can be used to store software programs and modules, and the processor 102 executes various functional applications and data processing by executing the software programs and modules stored in the memory 101. The communication interface 103 can be used to communicate signaling or data with other node devices.

[0074] Among them, the memory 101 can be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), etc.

[0075] The processor 102 may be an integrated circuit chip with signal processing capability. The processor 102 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0076] It is understood that the structure shown in the figure is only for illustration, and a method for detecting faults in power equipment based on a MAE autoencoder model may also include more or fewer components than those shown in the figure, or have a different configuration than that shown in the figure. Each component shown in the figure may be implemented by hardware, software, or a combination thereof.

[0077] In the embodiments provided in the present application, it should be understood that the disclosed method can also be implemented in other ways. The embodiments described above are merely schematic, for example, the flowchart or block diagram in the accompanying drawings shows the possible implementation architecture, functions and operations of the method and computer program product according to multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or the flowchart, and the combination of the boxes in the block diagram and / or the flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0078] In addition, the functional modules in the various embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

[0079] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0080] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0081] It will be apparent to those skilled in the art that the present application is not limited to the details of the exemplary embodiments described above, and that the present application can be implemented in other specific forms without departing from the spirit or essential features of the present application. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the present application is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims be included in the present application. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for detecting faults in power equipment based on a MAE autoencoder model, characterized in that: The following steps are involved: S1. Collect images of power equipment and establish an image dataset; S2, preprocessing the image data set, and labeling the preprocessed image data set according to the image type; S3, the image dataset is processed by mask and then input into the MAE autoencoder model for self-supervised training; S4, a multi-layer perceptron is connected to the trained MAE autoencoder model, and a power equipment fault detection model is obtained by training the model through a weighted loss function based on the labeled image dataset; S5. Input the image to be detected into the electric power equipment fault detection model to obtain the fault detection result of the electric power equipment.

2. A method for detecting faults in electric power equipment based on a MAE autoencoder model as claimed in claim 1, characterized in that: In step S2, the data preprocessing includes data denoising and data enhancement, and its expression is: x denoised =Filter(x) x aug =λ·x denoised +(1-λ)·Augment(x denoised ) In the formula, x denoised is the denoised image, Filter(·) is the median filter function, x is the original image in the dataset, and x aug is the image after data enhancement, λ represents the ratio of fusion with the original image, and Augment is the enhancement function.

3. A method for detecting faults in electric power equipment based on a MAE autoencoder model as claimed in claim 1, characterized in that: In step S3, the mask processing includes: randomly covering a part of the area on the input image to form a mask image, which is expressed as: x masked =x aug ⊙M In the formula, x masked is the mask image, x aug is the image after data enhancement, M is the mask matrix, and ⊙ represents element-by-element multiplication.

4. A method for detecting faults in electric power equipment based on a MAE autoencoder model as claimed in claim 3, characterized in that: In step S3, the expression of the MAE autoencoder model is: x r =f d (f e (x masked )) In the formula, x r is the image reconstructed by encoding and decoding the mask image through the autoencoder, f d (·) is the decoding reconstruction network, f e (·) is the encoding network, x masked is the mask image.

5. A method for detecting faults in electric power equipment based on a MAE autoencoder model as claimed in claim 4, characterized in that: The self-supervised training of the MAE autoencoder model includes: The MAE autoencoder model is trained using mean square error as the loss function, and the weights of the MAE autoencoder model are optimized by the back propagation algorithm to minimize the loss function. At the same time, self-supervised training is achieved by maximizing the similarity between the output image of the MAE model and the image before masking. The expression of the loss function is: In the formula, L is the loss function, N is the number of samples, and x r is the image reconstructed by encoding and decoding the mask image through the autoencoder, x i is the original image.

6. A method for detecting faults in electric power equipment based on a MAE autoencoder model as claimed in claim 1, characterized in that: In step S4, the expression of the weighted loss function is: In the formula, L' is the weighted loss function, m is the total number of samples, and m i is the number of the i-th category, For category y i The weight of y i is the category annotated in the image dataset, The actual result output by the model.

7. An electronic device, characterized in that: include: A memory for storing one or more programs; processor; When the one or more programs are executed by the processor, a method for detecting faults of electric power equipment based on a MAE autoencoding model as described in any one of claims 1 to 6 is implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, a method for detecting faults of electric power equipment based on a MAE autoencoder model as described in any one of claims 1 to 6 is implemented.

Images