Gear fault diagnosis method based on multi-scale coarse-grained convolutional neural network

Abstract

The invention relates to a gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network. The method comprises the steps that original vibration data are acquired and preprocessed; constructing a multi-scale coarse-grained convolutional neural network fault diagnosis model based on the convolutional neural network, wherein the model comprises an improved multi-scale coarse-grained layer, a multi-scale feature learning layer and a classification layer; inputting the preprocessed original vibration data into the multi-scale coarse-grained convolutional neural network fault diagnosis model, and carrying out model training; and preprocessing to-be-diagnosed data, inputting the preprocessed to-be-diagnosed data into the trained multi-scale coarse-grained convolutional neural network fault diagnosis model, and outputting a gear fault diagnosis result. Through multi-scale cascade architecture design and introduction of an improved multi-scale coarse graining layer, signal features can be effectively learned on multiple time scales, multi-scale features are directly and automatically extracted from original vibration signals, fault classification is completed, the classification precision is improved, and the requirement for manual intervention is remarkably reduced.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence and mechanical engineering technology, and in particular to a gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network. Background Art

[0002] Gears are an indispensable key component in most mechanical transmission systems. In actual engineering applications, gears are easily damaged locally due to the influence of various factors such as fatigue, wear, overload, corrosion, etc., which may cause the mechanical system to shut down or even cause accidents. Therefore, in order to ensure the safety and stability of the mechanical system, improve the reliability of the equipment and reduce maintenance costs, the condition monitoring and fault diagnosis of gears have become a research field that has received increasing attention. Therefore, it is of great research value and practical significance to develop a new fault diagnosis method with efficient adaptive ability and high accuracy to accurately identify gear faults under different working conditions.

[0003] In recent years, data-driven fault diagnosis has become an important research direction in the field of intelligent fault diagnosis of machinery. Traditional data-driven fault diagnosis methods usually include two main stages: feature extraction and fault identification. There are a large number of research results on data-driven intelligent fault diagnosis methods, such as back propagation neural network, extreme learning machine, support vector machine, k-nearest neighbor, hidden Markov model and cluster analysis. However, most of these methods belong to shallow learning processes, and their fault classification is usually performed after manual feature extraction, which makes the process more complicated and cumbersome. In addition, data-driven diagnosis methods often require a lot of prior knowledge and expert experience in signal processing, which limits their adaptability and diagnostic capabilities. In general, these shallow learning methods have the following significant defects: (1) feature extraction and classification are performed independently and cannot be optimized simultaneously; (2) the extracted features are highly dependent on manual processing and expert experience; (3) most shallow learning methods are only applicable to specific diagnostic scenarios and lack versatility.

[0004] Therefore, in the related technology, there is an urgent need for a method that can realize automated feature extraction and fault diagnosis and improve the adaptability and robustness of the diagnosis method. Summary of the invention

[0005] Based on this, it is necessary to provide a gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network, which can realize automated feature extraction and fault diagnosis and improve the adaptability and robustness of the fault diagnosis method, in response to the above technical problems.

[0006] In a first aspect, the present application provides a gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network. The method comprises:

[0007] Get raw vibration data and pre-process it;

[0008] A multi-scale coarse-grained convolutional neural network fault diagnosis model is constructed based on a convolutional neural network. The model includes an improved multi-scale coarse-grained layer, a multi-scale feature learning layer, and a classification layer. The multi-scale feature learning layer is composed of a stack of multiple deep Boltzmann machine learning layers.

[0009] Inputting the preprocessed raw vibration data into the multi-scale coarse-grained convolutional neural network fault diagnosis model to perform model training;

[0010] After preprocessing, the data to be diagnosed is input into the trained multi-scale coarse-grained convolutional neural network fault diagnosis model to output the gear fault diagnosis results.

[0011] Optionally, in one embodiment of the present application, the preprocessing includes:

[0012] The original vibration data is segmented.

[0013] Optionally, in one embodiment of the present application, the calculation formula of the improved multi-scale coarse-grained layer is:

[0014]

[0015] in, is the jth element in the improved coarse-grained time series of length N-τ+1, τ is the scale factor, x(i) is the original time series, and N is the length of the original time series.

[0016] Optionally, in one embodiment of the present application, the output feature of the deep Boltzmann machine learning layer is expressed as:

[0017]

[0018] in, is the output feature of each j-th node in the hidden layer of the deep Boltzmann machine learning layer, σ(·) is the sigmoid activation function, is the connection weight between the ith visible unit and the jth hidden unit, is the jth element in the improved coarse-grained time series of length N-τ+1, is the bias term of the jth hidden unit.

[0019] Optionally, in one embodiment of the present application, the classification layer uses a softmax classifier to calculate the probability distribution of each fault mode.

[0020] In a second aspect, the present application also provides a gear fault diagnosis device based on a multi-scale coarse-grained convolutional neural network. The device comprises:

[0021] A data acquisition module, used to acquire raw vibration data and perform preprocessing;

[0022] A fault diagnosis model construction module is used to construct a multi-scale coarse-grained convolutional neural network fault diagnosis model based on a convolutional neural network. The model includes an improved multi-scale coarse-grained layer, a multi-scale feature learning layer, and a classification layer. The multi-scale feature learning layer is composed of a stack of multiple deep Boltzmann machine learning layers.

[0023] A fault diagnosis model training module, used for inputting the original vibration data after preprocessing into the multi-scale coarse-grained convolutional neural network fault diagnosis model to perform model training;

[0024] The fault diagnosis module is used to pre-process the data to be diagnosed, input it into the trained multi-scale coarse-grained convolutional neural network fault diagnosis model, and output the gear fault diagnosis results.

[0025] In a third aspect, the present application further provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor executes the steps of the methods described in the above embodiments.

[0026] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method described in each of the above embodiments are implemented.

[0027] The above-mentioned gear fault diagnosis method based on multi-scale coarse-grained convolutional neural network, first, obtains the original vibration data and performs preprocessing; then, constructs a multi-scale coarse-grained convolutional neural network fault diagnosis model based on the convolutional neural network, the model includes an improved multi-scale coarse-grained layer, a multi-scale feature learning layer, and a classification layer, and the multi-scale feature learning layer is composed of a stack of multiple deep Boltzmann machine learning layers; then, the original vibration data after preprocessing is input into the multi-scale coarse-grained convolutional neural network fault diagnosis model to perform model training; finally, after preprocessing the data to be diagnosed, it is input into the trained multi-scale coarse-grained convolutional neural network fault diagnosis model to output the gear fault diagnosis result. That is to say, through the multi-scale cascade architecture design and the introduction of the improved multi-scale coarse-grained layer, it is possible to effectively learn signal features at multiple time scales, automatically extract multi-scale features directly from the original vibration signal and complete fault classification, thereby improving the classification accuracy, significantly reducing the need for manual intervention, and accurately identifying fault modes under different working conditions, with good robustness and generalization ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1An application environment diagram of a gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network in one embodiment;

[0029] Figure 2 A schematic diagram of a process of a gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network in one embodiment;

[0030] Figure 3 A schematic diagram of data segmentation in one embodiment;

[0031] Figure 4 Schematic diagram of the structure of a multi-scale coarse-grained convolutional neural network fault diagnosis model in one embodiment;

[0032] Figure 5 It is a schematic diagram of analyzing the influence of the number of Boltzmann machine learning layers of different scales and depths on the accuracy of diagnosis results in one embodiment;

[0033] Figure 6 The figure is a schematic diagram of the overall process of performing fault diagnosis based on a multi-scale coarse-grained convolutional neural network fault diagnosis model in one embodiment;

[0034] Figure 7 is a structural block diagram of a gear fault diagnosis device based on a multi-scale coarse-grained convolutional neural network in one embodiment;

[0035] Figure 8 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0037] The gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal communicates with the server through the network. The data storage system can store the data that the server needs to process. The data storage system can be integrated on the server, or it can be placed on the cloud or other network servers. Among them, the terminal can be but not limited to various personal computers, laptops, smart phones, tablets, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server can be implemented as an independent server or a server cluster consisting of multiple servers.

[0038] In one embodiment, Figure 2 As shown in the figure, a gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network is provided. Figure 1 The server in the example is used to illustrate the following steps:

[0039] S201: Acquire raw vibration data and perform preprocessing.

[0040] In the embodiment of the present application, first, vibration data is obtained from a mechanical fault simulator through a data acquisition unit and monitoring software, and preprocessed to facilitate feature extraction, and a certain proportion of samples are selected as training sets and test sets. Generally speaking, the more training data there is, the better the classification effect, but the time required for training will also increase accordingly. In order to achieve a balance between classification accuracy and training efficiency, the data samples of each fault mode are allocated as training sets and test sets in a ratio of 1:1. Specifically, the gear vibration data is collected by an acceleration sensor installed on the gearbox with a sampling frequency of 5120Hz. The gearbox consists of a pinion and a gear, and covers six fault modes during operation, including normal operation, gear pitting fault (PF), gear fracture fault (FF), pinion wear fault (WF), gear pitting and pinion wear fault (PWF), and gear fracture and pinion wear fault (FWF).

[0041] In one embodiment of the present application, the preprocessing includes:

[0042] The original vibration data is segmented.

[0043] In one embodiment of the present application, Figure 3 As shown in the figure, the original vibration signal is segmented into multiple sub-signals of equal size using a sliding window with overlap. In the specific application, the original gear vibration signal is segmented by a sliding window with 1024 points overlap to obtain 100 sub-signals of equal size (each sub-signal contains 2048 data points), and each sub-signal is regarded as a sample. For each fault mode, 200 samples are obtained by data segmentation, of which 100 samples are randomly selected as training sets and the remaining 100 samples are used as test sets. Therefore, the entire gear data set contains 1200 samples, covering six fault modes, and the number of samples in the training set and the test set is 600.

[0044] S203: Construct a multi-scale coarse-grained convolutional neural network fault diagnosis model based on the convolutional neural network, wherein the model includes an improved multi-scale coarse-grained layer, a multi-scale feature learning layer, and a classification layer, wherein the multi-scale feature learning layer is composed of a stack of multiple deep Boltzmann machine learning layers.

[0045] In the embodiment of the present application, a multi-scale coarse-grained convolutional neural network fault diagnosis model is constructed with a convolutional neural network as the main body, such as Figure 4 As shown in the figure, the model mainly consists of three layers: (1) improved multi-scale coarse-graining layer, (2) multi-scale feature learning layer, and (3) classification layer. In the improved multi-scale coarse-graining layer, the improved multi-scale coarse-graining process is applied to the original signal to obtain coarse-grained time series at different scales. In the multi-scale feature learning layer, high-level and effective feature representations are automatically learned from the improved multi-scale coarse-grained time series in a parallel manner through multiple stacked deep Boltzmann machine (DBM) learning layers. In the classification layer, the goal is to solve the multi-category classification problem and automatically output the classification results through an appropriate classifier.

[0046] In one embodiment of the present application, the calculation formula of the improved multi-scale coarse-grained layer is:

[0047]

[0048] in, is the jth element in the improved coarse-grained time series of length N-τ+1, τ is the scale factor, x(i) is the original time series, and N is the length of the original time series.

[0049] In one embodiment of the present application, after introducing the improved multi-scale coarse-grained layer, feature information estimation can be performed more accurately and reliably at different scales. In fact, the improved multi-scale coarse-grained layer is essentially a compact moving average process with the characteristic of overlapping adjacent data. Compared with the traditional multi-scale conversion method, the improved multi-scale coarse-grained layer is not only easier to implement, but also has a shorter running time, thereby reducing the complexity of the model and improving the efficiency of feature learning. When the scale factor τ takes the values ​​of 1, 2 and 3, three sets of improved coarse-grained time series ( and ) in order to extract multi-scale feature information in different scale ranges.

[0050] In one embodiment of the present application, the output feature of the deep Boltzmann machine learning layer is expressed as:

[0051]

[0052] in, is the output feature of each j-th node in the hidden layer of the deep Boltzmann machine learning layer, σ(·) is the sigmoid activation function, is the connection weight between the ith visible unit and the jth hidden unit, is the jth element in the improved coarse-grained time series of length N-τ+1, is the bias term of the jth hidden unit.

[0053] In one embodiment of the present application, multiple stacked deep Boltzmann machine (DBM) learning layers are used to automatically learn high-level and effective feature representations from the improved multi-scale coarse-grained time series in parallel. Through this parallel approach, the deep convolutional network CNN can extract deep feature information of the signal at different scales, thereby improving the performance of fault classification. Taking the scale factor τ = 3 as an example, the improved coarse-grained time series is Then the output of the jth node in the hidden layer of DBM3 is defined as:

[0054]

[0055] Similarly, the output feature representation of the deep Boltzmann machine learning layer when the scale factor τ = 1 and 2 can be obtained and Through the cascade operation, we can finally get the following high-level feature representation:

[0056]

[0057] Obviously, compared with the traditional single-scale feature learning method, the above-mentioned high-level feature representation contains more comprehensive and richer fault features from multiple scales, thus enhancing the feature learning capability while also improving the ability to distinguish categories.

[0058] In one embodiment of the present application, the classification layer uses a softmax classifier to calculate the probability distribution of each fault mode.

[0059] In one embodiment of the present application, the goal of the classification layer is to solve the multi-class classification problem and automatically output the classification results through an appropriate classifier. Specifically, a Softmax classifier is used to calculate the probability distribution of each fault mode. Assuming that the input sample x contains K types of health states, the probability corresponding to the jth class can be calculated by the following formula:

[0060]

[0061] where θ represents the parameters learned from the convolutional neural network CNN, and After the above process, the classification accuracy of different fault modes can be finally obtained.

[0062] S205: Inputting the original vibration data after preprocessing into the multi-scale coarse-grained convolutional neural network fault diagnosis model to perform model training.

[0063] In the embodiment of the present application, the original vibration data after preprocessing, i.e., the data of the training set and the test set, is input into the constructed multi-scale coarse-grained convolutional neural network fault diagnosis model, and the end-to-end fault diagnosis architecture based on the model is trained. Optionally, the K-nearest neighbor (KNN) algorithm can be used to optimize the model parameters.

[0064] S207: After preprocessing the data to be diagnosed, the data is input into a trained multi-scale coarse-grained convolutional neural network fault diagnosis model to output the gear fault diagnosis result.

[0065] In an embodiment of the present application, the data to be diagnosed is input into a trained multi-scale coarse-grained convolutional neural network fault diagnosis model after preprocessing such as data segmentation. The model automatically learns multi-scale features and outputs diagnostic results of different fault modes.

[0066] In the process of multi-scale coarse-graining and multi-scale feature learning, appropriate scale factors and DBM numbers must be preset, such as Figure 5 As shown in the figure, it is a schematic diagram of the analysis of the influence of the number of Boltzmann machine learning layers of different scales and depths on the accuracy of diagnostic results. According to the prior knowledge of multi-scale analysis, it can be concluded that the more scale factors or DBMs, the higher the training time and computational complexity. Therefore, under the premise of maintaining the universality of the model, the scale factors and the number of DBMs are set to 3 in the embodiment of the present application to achieve a balance between accuracy and efficiency. Figure 6 As shown, it is a schematic diagram of the overall process of fault diagnosis based on the corresponding multi-scale coarse-grained convolutional neural network fault diagnosis model.

[0067] In the above-mentioned gear fault diagnosis method based on multi-scale coarse-grained convolutional neural network, first, the original vibration data is obtained and preprocessed; then, a multi-scale coarse-grained convolutional neural network fault diagnosis model is constructed based on the convolutional neural network, and the model includes an improved multi-scale coarse-grained layer, a multi-scale feature learning layer, and a classification layer, and the multi-scale feature learning layer is composed of a stack of multiple deep Boltzmann machine learning layers; then, the original vibration data after preprocessing is input into the multi-scale coarse-grained convolutional neural network fault diagnosis model for model training; finally, after preprocessing the data to be diagnosed, it is input into the trained multi-scale coarse-grained convolutional neural network fault diagnosis model to output the gear fault diagnosis result. That is to say, through the multi-scale cascade architecture design and the introduction of the improved multi-scale coarse-grained layer, the signal features can be effectively learned on multiple time scales, the multi-scale features can be automatically extracted directly from the original vibration signal and the fault classification can be completed, the classification accuracy is improved, the need for manual intervention is significantly reduced, and the fault mode can be accurately identified under different working conditions, and it has good robustness and generalization ability.

[0068] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0069] Based on the same inventive concept, the embodiment of the present application also provides a gear fault diagnosis device based on a multi-scale coarse-grained convolutional neural network for implementing the above-mentioned gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in one or more embodiments of a gear fault diagnosis device based on a multi-scale coarse-grained convolutional neural network provided below can refer to the above limitations on a gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network, which will not be repeated here.

[0070] In one embodiment, Figure 7 As shown, a gear fault diagnosis device 700 based on a multi-scale coarse-grained convolutional neural network is provided, comprising: a data acquisition module 701, a fault diagnosis model construction module 703, a fault diagnosis model training module 705 and a fault diagnosis module 707, wherein:

[0071] The data acquisition module 701 is used to acquire original vibration data and perform preprocessing.

[0072] The fault diagnosis model construction module 703 is used to construct a multi-scale coarse-grained convolutional neural network fault diagnosis model based on the convolutional neural network. The model includes an improved multi-scale coarse-grained layer, a multi-scale feature learning layer, and a classification layer. The multi-scale feature learning layer is composed of a stack of multiple deep Boltzmann machine learning layers.

[0073] The fault diagnosis model training module 705 is used to input the original vibration data after preprocessing into the multi-scale coarse-grained convolutional neural network fault diagnosis model to perform model training.

[0074] The fault diagnosis module 707 is used to pre-process the data to be diagnosed and then input it into the trained multi-scale coarse-grained convolutional neural network fault diagnosis model to output the gear fault diagnosis result.

[0075] In one embodiment of the present application, the data acquisition module is further used to:

[0076] The original vibration data is segmented.

[0077] In one embodiment of the present application, the calculation formula of the improved multi-scale coarse-grained layer is:

[0078]

[0079] in, is the jth element in the improved coarse-grained time series of length N-τ+1, τ is the scale factor, x(i) is the original time series, and N is the length of the original time series.

[0080] In one embodiment of the present application, the output feature of the deep Boltzmann machine learning layer is expressed as:

[0081]

[0082] in, is the output feature of each j-th node in the hidden layer of the deep Boltzmann machine learning layer, σ(·) is the sigmoid activation function, is the connection weight between the ith visible unit and the jth hidden unit, is the jth element in the improved coarse-grained time series of length N-τ+1, is the bias term of the jth hidden unit.

[0083] In one embodiment of the present application, the classification layer uses a softmax classifier to calculate the probability distribution of each fault mode.

[0084] Each module in the above-mentioned gear fault diagnosis device based on multi-scale coarse-grained convolutional neural network can be implemented in whole or in part by software, hardware and their combination. Each of the above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.

[0085] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 8As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0086] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0087] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0088] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0089] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0090] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0091] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0092] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0093] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network, characterized in that: The method comprises: Get raw vibration data and pre-process it; A multi-scale coarse-grained convolutional neural network fault diagnosis model is constructed based on a convolutional neural network. The model includes an improved multi-scale coarse-grained layer, a multi-scale feature learning layer, and a classification layer. The multi-scale feature learning layer is composed of a stack of multiple deep Boltzmann machine learning layers. Inputting the preprocessed raw vibration data into the multi-scale coarse-grained convolutional neural network fault diagnosis model to perform model training; After preprocessing, the data to be diagnosed is input into the trained multi-scale coarse-grained convolutional neural network fault diagnosis model to output the gear fault diagnosis results.

2. A gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network according to claim 1, characterized in that: The pre-processing comprises: The original vibration data is segmented.

3. The gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network according to claim 1 is characterized in that: The calculation formula of the improved multi-scale coarse-grained layer is: in, is the jth element in the improved coarse-grained time series of length N-τ+1, τ is the scale factor, x(i) is the original time series, and N is the length of the original time series.

4. The gear fault diagnosis method based on a multi-scale coarse-grained convolutional neural network according to claim 1 is characterized in that: The output feature of the deep Boltzmann machine learning layer is expressed as: in, is the output feature of each j-th node in the hidden layer of the deep Boltzmann machine learning layer, σ(·) is the sigmoid activation function, is the connection weight between the ith visible unit and the jth hidden unit, is the jth element in the improved coarse-grained time series of length N-τ+1, is the bias term of the jth hidden unit.

5. The gear fault diagnosis method based on multi-scale coarse-grained convolutional neural network according to claim 1 is characterized in that: The classification layer uses a softmax classifier to calculate the probability distribution of each fault mode.

6. A gear fault diagnosis device based on a multi-scale coarse-grained convolutional neural network, characterized in that: The device comprises: A data acquisition module, used to acquire raw vibration data and perform preprocessing; A fault diagnosis model construction module is used to construct a multi-scale coarse-grained convolutional neural network fault diagnosis model based on a convolutional neural network. The model includes an improved multi-scale coarse-grained layer, a multi-scale feature learning layer, and a classification layer. The multi-scale feature learning layer is composed of a stack of multiple deep Boltzmann machine learning layers. A fault diagnosis model training module, used for inputting the original vibration data after preprocessing into the multi-scale coarse-grained convolutional neural network fault diagnosis model to perform model training; The fault diagnosis module is used to pre-process the data to be diagnosed, input it into the trained multi-scale coarse-grained convolutional neural network fault diagnosis model, and output the gear fault diagnosis results.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are 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, the steps of the method according to any one of claims 1 to 5 are implemented.

Images