Epicyclic gearbox fault diagnosis method and system based on visual atlas feature fusion, medium and equipment

Abstract

The invention relates to the field of planetary gear box fault diagnosis, and discloses a planetary gear box fault diagnosis method, system, medium and equipment based on visual atlas feature fusion, and the method comprises the steps: collecting a normal state and a fault time domain signal during the operation of a planetary gear box through a pre-built planetary gear box working platform; carrying out Welch transformation on the time domain signal to obtain a power spectrum, and constructing a visibility graph map by adopting a visibility graph algorithm according to the power spectrum so as to extract centrality indexes of map nodes and fuse the centrality indexes into a feature matrix; the feature data is normalized and divided into a training set and a test set, a CNN-Inception model is trained based on the training set, the CNN-Inception model is checked by the test set, and the performance index is the accuracy rate of fault diagnosis; and after feature extraction is carried out on planetary gearbox signals collected in real time, the planetary gearbox signals are input into the trained CNN-Inception model for diagnosis. Effective feature information can be extracted, and the model diagnosis accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of planetary gearbox fault diagnosis, and in particular to a planetary gearbox fault diagnosis method, system, medium and equipment based on visual spectrum feature fusion. Background Art

[0002] Planetary gearboxes have the advantages of large transmission ratio and strong load-bearing capacity. They are widely used in the transmission systems of equipment in the fields of energy, aerospace, etc. At the same time, their maintenance and fault detection issues are also of great concern. Planetary gearboxes are composed of key components such as sun gear, planetary gear, ring gear and planet carrier. The planetary gear is in contact and meshing with both the sun gear and the ring gear. If the ring gear is fixed and the sun gear rotates, the planetary gear will rotate while also revolving around the sun gear with the planet carrier. Compared with traditional fixed-axis gearboxes, planetary gearboxes have more complex gear distribution and operating forms. At the same time, they are subject to variable environmental excitations. The frequency information of their vibration signals is complex, time-varying, and has obvious modulation characteristics. Therefore, the difficulty of signal analysis and fault diagnosis is increased.

[0003] Scholars have invested a lot of energy in developing effective planetary gearbox fault diagnosis technology. However, due to the complexity of the planetary gearbox vibration signal, how to effectively extract features and perform fault diagnosis based on the characteristics of the signal is still a hot research direction. Summary of the invention

[0004] In view of the above problems, the purpose of the present invention is to provide a planetary gearbox fault diagnosis method, system, medium and equipment based on visual spectrum feature fusion, which can extract effective feature information and improve the accuracy of model diagnosis.

[0005] To achieve the above-mentioned purpose, in a first aspect, the technical solution adopted by the present invention is: a planetary gearbox fault diagnosis method based on visible spectrum feature fusion, which comprises: collecting time domain signals of normal state and fault of the planetary gearbox during operation through a pre-built planetary gearbox working platform; performing Welch transform on the time domain signal to obtain a power spectrum, and constructing a visible graph spectrum according to the power spectrum using a visible graph algorithm to extract the centrality index of the graph nodes and fuse them into a feature matrix; normalizing the feature data and dividing it into a training set and a test set, training a CNN-Inception model based on the training set, and testing the CNN-Inception model with a test set, and the performance indicator is the accuracy of fault diagnosis; after feature extraction, the planetary gearbox signal collected in real time is input into the trained CNN-Inception model for diagnosis.

[0006] Furthermore, the time domain signals of normal state and fault include five states: normal state of planetary gearbox, planetary gear wear, planetary gear tooth breakage, planetary gear tooth surface wear and tooth breakage, bearing rolling element missing and planetary gear tooth breakage.

[0007] Furthermore, the time domain signal is subjected to Welch transform to obtain the power spectrum, including:

[0008] The time domain signal x(n) of length N is divided into L segments, each segment has M data, n = 0, 1, ..., N-1;

[0009] Add the window function w(n) to each data segment and find the periodogram of each segment;

[0010] All periodograms are approximated to be uncorrelated with each other and the power spectrum is obtained.

[0011] Furthermore, a visible graph algorithm is used to construct a visible graph spectrum according to the power spectrum, including:

[0012] The horizontal visibility graph algorithm is adopted to treat each data point of the power spectrum signal as a node in the network. The data points are connected according to the visibility criterion to obtain a graph structure, which is a visibility graph.

[0013] Furthermore, the visibility criteria are:

[0014] If the midpoint A(t a ,x a ) and point B(t b ,x b ) can be connected, then point A and point B are adjacent or for any point C(t c ,x c ), where t a <t c <t b , satisfying the following formula:

[0015]

[0016] Furthermore, the centrality indicators of the graph include degree centrality, betweenness centrality and closeness centrality, and the three centrality indicators are connected in parallel to form a feature matrix.

[0017] Furthermore, the planetary gearbox signal collected in real time is extracted and input into the trained CNN-Inception model for diagnosis, including:

[0018] Any signal of any planetary gearbox is collected, and after passing through Welch transform, horizontal visible graph transform, visible graph index extraction and feature fusion in sequence, it is input into the trained CNN-Inception model for recognition. The CNN-Inception model is used to determine whether the planetary gearbox is faulty based on the output result; at the same time, the collected data is added to the database.

[0019] In the second aspect, the technical solution adopted by the present invention is: a planetary gearbox fault diagnosis system based on visible spectrum feature fusion, which includes: a preprocessing module, which collects time domain signals of normal state and fault of the planetary gearbox during operation through a pre-built planetary gearbox working platform; a data fusion module, which performs Welch transform on the time domain signal to obtain a power spectrum, and constructs a visible graph spectrum according to the power spectrum using a visible graph algorithm to extract the centrality index of the graph node and fuse it into a feature matrix; a model training and testing module, which normalizes the feature data and divides it into a training set and a test set, trains a CNN-Inception model based on the training set, and verifies the CNN-Inception model with the test set, and the performance indicator is the accuracy of fault diagnosis; a diagnosis module, which extracts the features of the planetary gearbox signal collected in real time and inputs it into the trained CNN-Inception model for diagnosis.

[0020] In a third aspect, the technical solution adopted by the present invention is: a computer-readable storage medium storing one or more programs, wherein the one or more programs include instructions, and when the instructions are executed by a computing device, the computing device executes any one of the above methods.

[0021] In a fourth aspect, the technical solution adopted by the present invention is: a computing device, comprising: one or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any of the above methods.

[0022] The present invention adopts the above technical solution, which has the following advantages:

[0023] 1. The present invention can extract effective information and shorten the signal length through Welch transformation.

[0024] 2. The visual graph algorithm of the present invention can quantify the signal dynamic characteristics from a spatial perspective and has certain advantages in processing nonlinear and non-stationary signals. By extracting and fusing the node features of the visual graph spectrum, the fault information hidden in the graph topology can be effectively extracted.

[0025] 3. The CNN-Inception model adopted in the present invention has the ability of multi-scale feature extraction, which can efficiently realize the classification of different operating states of the planetary gearbox. Compared with the traditional model, it has higher accuracy and efficient memory utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is a flow chart of a planetary gearbox fault diagnosis method based on visual spectrum feature fusion in an embodiment of the present invention;

[0027] Figure 2 It is a structural diagram of the CNN-Inception model in an embodiment of the present invention;

[0028] Figure 3 It is a schematic diagram of the structure of a computing device in one embodiment of the present invention. DETAILED DESCRIPTION

[0029] In recent years, the development of graph signal processing has provided new ideas for fault diagnosis technology. Compared with the traditional time domain, frequency domain and time-frequency domain, the mechanical vibration signal is analyzed and processed in the vertex domain, graph spectrum domain and vertex-graph spectrum domain, which can effectively extract more fault information hidden in the graph topology structure. Therefore, the present invention proposes a planetary gearbox fault diagnosis method, system, medium and equipment based on visible graph spectrum feature fusion, which includes: pre-establishing a planetary gearbox working platform, collecting normal and faulty time domain signals of the planetary gearbox during operation; then performing Welch transformation on the planetary gearbox time domain signal to obtain a power spectrum, adopting a visible graph algorithm to construct a visible graph spectrum, calculating the centrality index of the graph spectrum node and merging it into a feature matrix. Secondly, the data is normalized and divided into a training set and a test set, the CNN-Inception model is trained based on the training set data, and the model is tested using the test set. The performance indicator is the accuracy of fault diagnosis. Finally, the collected planetary gearbox signal is feature extracted and input into the CNN-Inception model for diagnosis. The present invention analyzes the planetary gearbox signal, extracts the visual spectrum features, and effectively extracts the fault information hidden in the graph topology structure, showing the advantages of high accuracy and high diagnostic efficiency during diagnosis.

[0030] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0031] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0032] In one embodiment of the present invention, a planetary gearbox fault diagnosis method based on visual spectrum feature fusion is provided, which is implemented based on visual spectrum features and deep learning. In this embodiment, Figure 1 As shown, the method comprises the following steps:

[0033] 1) Through the pre-built planetary gearbox working platform, collect the time domain signals of the normal state and faults of the planetary gearbox during operation;

[0034] 2) Perform Welch transformation on the time domain signal to obtain the power spectrum, and construct a visible graph spectrum using the visible graph algorithm based on the power spectrum to extract the centrality index of the graph nodes and fuse them into a feature matrix;

[0035] 3) Normalize the feature data and divide it into a training set and a test set. Train the CNN-Inception model based on the training set and test the CNN-Inception model with the test set. The performance indicator is the accuracy of fault diagnosis.

[0036] 4) After feature extraction, the planetary gearbox signal collected in real time is input into the trained CNN-Inception model for diagnosis.

[0037] In the above step 1), the time domain signals of normal state and fault include five states: normal state of planetary gearbox, planetary gear wear, planetary gear broken teeth, planetary gear tooth surface wear and broken teeth, bearing rolling element missing and planetary gear broken teeth.

[0038] In the above step 2), the time domain signal is subjected to Welch transformation to obtain a power spectrum, which includes the following steps:

[0039] 2.1.1) The time domain signal x(n) of length N is divided into L segments, each segment has M data, n = 0, 1, ..., N-1;

[0040] Among them, the i-th data x i (n) is expressed as follows:

[0041] x i (n)=x(n+iM-M),0≤n≤M,1≤i≤L

[0042] 2.1.2) Add the window function w(n) to each data segment and find the periodogram of each segment;

[0043] Among them, the periodogram I of the i-th segment i (ω) is:

[0044]

[0045] In the formula, U becomes the normalization factor, as shown below:

[0046]

[0047] 2.1.3) All periodograms are approximated to be uncorrelated with each other, and the estimated power spectrum is obtained;

[0048] Among them, the power spectrum P xx (e jω )for:

[0049]

[0050] Where ω is the frequency variable, I i (ω) is the periodogram of the i-th segment of data.

[0051] In this embodiment, a window size of 1024 is adopted to convert the time domain signal into a power spectrum signal.

[0052] In the above step 2), a visible graph algorithm is used to construct a visible graph spectrum according to the power spectrum, specifically:

[0053] Using the horizontal visibility graph algorithm, each data point of the power spectrum signal is regarded as a node in the network. The data points are connected according to the visibility criterion to obtain a graph structure, which is the visibility graph.

[0054] In this embodiment, the visibility criterion is: if the midpoint A(t a ,x a ) and point B(t b ,x b ) can be connected, then point A and point B are adjacent or for any point C(t c ,x c ), where t a <t c <t b , satisfying the following formula:

[0055]

[0056] Among them, (t a ,x a ) is the coordinate value of point A, (t b ,x b ) is the coordinate value of point B, (t c ,x c ) is the coordinate value of point C.

[0057] In the above step 2), the centrality indicators of the graph include degree centrality, betweenness centrality and closeness centrality, and the three centrality indicators are connected in parallel to form a feature matrix.

[0058] Among them, 2.2.1) Degree centrality is: a measure of the number of connections between a node in the graph, that is, the degree of the node. The degree DC of the node i The expression is:

[0059]

[0060] Among them, k i It represents the number of edges associated with node i in the graph. The larger the value of node degree centrality, the greater its ability to communicate directly with other nodes and the more important it is in the graph.

[0061] 2.2.2) Betweenness centrality is to measure the bridging role and control power of a node in a graph, that is, the importance of a node in the flow and dissemination of information in a graph. Betweenness centrality BC i The expression is:

[0062]

[0063] Among them, n i st represents the number of paths passing through node i in the shortest path between nodes s and t, g st Represents the number of all shortest paths. The higher the value of node betweenness centrality, the greater the influence of the node and the more important it is.

[0064] 2.2.3) Closeness centrality is: It measures the average distance between the i node and other nodes, that is, the closeness between the i node and other nodes. Closeness centrality CC i The expression is:

[0065]

[0066] Where: d ij It indicates the number of edges contained in the shortest path from node i to node j in the graph. The larger the value of the node's proximity to centrality, the more central the node is in the network and the more important it is.

[0067] The three features are fused in parallel, and the fusion formula is as follows:

[0068]

[0069] Among them, X Fusion is the feature matrix; X DC is the degree centrality vector, X BC is the betweenness centrality vector, X CC is the proximity centrality vector.

[0070] In the above step 3), the feature data is normalized and divided into a training set and a test set. The CNN-Inception model is trained based on the training set. The structure diagram of the CNN-Inception model is as follows: Figure 2 As shown. The model is made to understand the characteristics of different states in the visual indicators and complete the correct identification and classification of data. The accuracy of the model is tested based on the test set.

[0071] In this embodiment, Figure 2 As shown in the figure, the CNN-Inception model structure includes a 3×1 convolutional layer, a ReLu layer, the first 3×1 maximum pooling layer, an Inception module, a second 3×1 maximum pooling layer, a fully connected layer, and a Softmax layer.

[0072] Among them, the Inception module includes three parallel 1×1 convolutional layers and a third 3×1 maximum pooling layer in parallel with the three 1×1 convolutional layers. Each 1×1 convolutional layer is sequentially followed by a batch normalization layer. The batch normalization layer at the front end is followed by a ReLu layer and then connected to the depth splicing layer for output; the other two batch normalization layers pass through the 3×1 convolutional layer and the 5×1 convolutional layer respectively, and then pass through the batch normalization layer again and enter the ReLu layer, and then connect to the depth splicing layer for output; the third 3×1 maximum pooling layer is sequentially followed by a 1×1 convolutional layer, a batch normalization layer, and a ReLu layer, which are connected to the depth splicing layer after the ReLu layer and then output.

[0073] In the above step 4), the planetary gearbox signal collected in real time is extracted and input into the trained CNN-Inception model for diagnosis, specifically:

[0074] Any signal of any planetary gearbox is collected, and after passing through Welch transform, horizontal visible graph transform, visible graph index extraction and feature fusion in sequence, it is input into the trained CNN-Inception model for recognition. The CNN-Inception model is used to determine whether the planetary gearbox is faulty based on the output result. At the same time, the collected data is added to the database, and a large amount of data is accumulated in actual operation to improve the accuracy of the model.

[0075] In one embodiment of the present invention, a planetary gearbox fault diagnosis system based on visual spectrum feature fusion is provided, which includes:

[0076] The pre-processing module collects the time domain signals of the normal state and faults of the planetary gearbox during operation through the pre-built planetary gearbox working platform;

[0077] The data fusion module performs Welch transformation on the time domain signal to obtain the power spectrum, and uses the visible graph algorithm to construct a visible graph based on the power spectrum to extract the centrality index of the graph nodes and fuse them into a feature matrix;

[0078] Model training and testing module: normalizes the feature data and divides it into training set and test set. The CNN-Inception model is trained based on the training set and tested with the test set. The performance indicator is the accuracy of fault diagnosis.

[0079] The diagnosis module extracts the features of the planetary gearbox signal collected in real time and inputs it into the trained CNN-Inception model for diagnosis.

[0080] In the above embodiment, the time domain signals of normal state and fault include five states: normal state of the planetary gearbox, planetary gear wear, planetary gear broken teeth, planetary gear tooth surface wear and broken teeth, bearing rolling element missing and planetary gear broken teeth.

[0081] In the above embodiment, the time domain signal is subjected to Welch transformation to obtain the power spectrum, including:

[0082] The time domain signal x(n) of length N is divided into L segments, each segment has M data, n = 0, 1, ..., N-1;

[0083] Add the window function w(n) to each data segment and find the periodogram of each segment;

[0084] All periodograms are approximated to be uncorrelated with each other and the power spectrum is obtained.

[0085] In the above embodiment, the visible graph algorithm is used to construct the visible graph spectrum according to the power spectrum, including:

[0086] The horizontal visibility graph algorithm is adopted to treat each data point of the power spectrum signal as a node in the network. The data points are connected according to the visibility criterion to obtain a graph structure, which is a visibility graph.

[0087] In the above embodiment, the visibility criterion is:

[0088] If the midpoint A(t a ,x a ) and point B(t b ,x b ) can be connected, then point A and point B are adjacent or for any point C(t c ,x c ), where t a <t c <t b , satisfying the following formula:

[0089]

[0090] In the above embodiment, the centrality indexes of the graph include degree centrality, betweenness centrality and closeness centrality, and the three centrality indexes are connected in parallel to form a feature matrix.

[0091] In the above embodiment, the planetary gearbox signal collected in real time is subjected to feature extraction and then input into the trained CNN-Inception model for diagnosis, specifically:

[0092] Any signal of any planetary gearbox is collected, and after passing through Welch transform, horizontal visible graph transform, visible graph index extraction and feature fusion in sequence, it is input into the trained CNN-Inception model for recognition. The CNN-Inception model is used to determine whether the planetary gearbox is faulty based on the output result; at the same time, the collected data is added to the database.

[0093] The system provided in this embodiment is used to execute the above-mentioned method embodiments. Please refer to the above-mentioned embodiments for specific processes and detailed contents, which will not be repeated here.

[0094] like Figure 3 As shown, it is a schematic diagram of the structure of a computing device provided in an embodiment of the present invention. The computing device can be a terminal, which can include: a processor, a communication interface, a memory, a display screen and an input device. Among them, the processor, the communication interface and the memory complete mutual communication through a communication bus. The processor is used to provide computing and control capabilities. The memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. When the computer program is executed by the processor, the method in each of the above embodiments is implemented; 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 is used to communicate with an external terminal in a wired or wireless manner. The wireless manner can be realized through WIFI, a management network, NFC (near field communication) or other technologies. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the housing of the computing device, or an external keyboard, touchpad or mouse. The processor can call the logical instructions in the memory.

[0095] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for 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 method described in each embodiment of the present invention. 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.

[0096] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the scheme of the present application, and does not constitute a limitation on the computing device to which the scheme of the present application is applied. The specific computing device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0097] In one embodiment of the present invention, a computer program product is provided, wherein the computer program product includes a computer program stored on a non-transitory computer-readable storage medium, wherein the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the methods provided by the above-mentioned method embodiments.

[0098] In one embodiment of the present invention, a non-transitory computer-readable storage medium is provided, wherein the non-transitory computer-readable storage medium stores server instructions, and the computer instructions enable a computer to execute the methods provided in the above embodiments.

[0099] The above embodiment provides a computer-readable storage medium, whose implementation principle and technical effect are similar to those of the above method embodiment, and will not be repeated here.

[0100] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0101] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0102] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0103] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A planetary gearbox fault diagnosis method based on visual spectrum feature fusion, characterized in that: include: Through the pre-built planetary gearbox working platform, the time domain signals of normal state and fault of the planetary gearbox during operation are collected; The time domain signal is transformed by Welch to obtain the power spectrum. The visible graph algorithm is used to construct the visible graph graph based on the power spectrum to extract the centrality index of the graph nodes and fuse them into a feature matrix. The feature data is normalized and divided into a training set and a test set. The CNN-Inception model is trained based on the training set and tested with the test set. The performance indicator is the accuracy of fault diagnosis. The planetary gearbox signal collected in real time is extracted through feature extraction and then input into the trained CNN-Inception model for diagnosis.

2. The planetary gearbox fault diagnosis method based on visual spectrum feature fusion as claimed in claim 1, characterized in that: The time domain signals of normal state and fault include five states: normal state of planetary gearbox, planetary gear wear, planetary gear broken teeth, planetary gear tooth surface wear and broken teeth, bearing rolling element missing and planetary gear broken teeth.

3. The planetary gearbox fault diagnosis method based on visual spectrum feature fusion as claimed in claim 1, characterized in that: The time domain signal is transformed by Welch to obtain the power spectrum, including: The time domain signal x(n) of length N is divided into L segments, each segment has M data, n = 0, 1, ..., N-1; Add the window function w(n) to each data segment and find the periodogram of each segment; All periodograms are approximated to be uncorrelated with each other and the power spectrum is obtained.

4. The planetary gearbox fault diagnosis method based on visual spectrum feature fusion as claimed in claim 1, characterized in that: The visible graph algorithm is used to construct a visible graph spectrum based on the power spectrum, including: The horizontal visibility graph algorithm is adopted to treat each data point of the power spectrum signal as a node in the network. The data points are connected according to the visibility criterion to obtain a graph structure, which is a visibility graph.

5. The planetary gearbox fault diagnosis method based on visual spectrum feature fusion as claimed in claim 4, characterized in that: The visibility guidelines are: If the midpoint A(t a ,x a ) and point B(t b ,x b ) can be connected, then point A and point B are adjacent or for any point C(t c ,x c ), where t a <t c <t b , satisfying the following formula:

6. The planetary gearbox fault diagnosis method based on visual spectrum feature fusion as claimed in claim 1, characterized in that: The centrality indicators of a graph include degree centrality, betweenness centrality, and closeness centrality. The three centrality indicators are connected in parallel to form a feature matrix.

7. The planetary gearbox fault diagnosis method based on visual spectrum feature fusion as claimed in claim 1, characterized in that: After feature extraction, the planetary gearbox signal collected in real time is input into the trained CNN-Inception model for diagnosis, including: Any signal of any planetary gearbox is collected, and after passing through Welch transform, horizontal visible graph transform, visible graph index extraction and feature fusion in sequence, it is input into the trained CNN-Inception model for recognition. The CNN-Inception model is used to determine whether the planetary gearbox is faulty based on the output result; at the same time, the collected data is added to the database.

8. A planetary gearbox fault diagnosis system based on visual spectrum feature fusion, characterized in that: include: The pre-processing module collects the time domain signals of the normal state and faults of the planetary gearbox during operation through the pre-built planetary gearbox working platform; The data fusion module performs Welch transformation on the time domain signal to obtain the power spectrum, and uses the visible graph algorithm to construct a visible graph based on the power spectrum to extract the centrality index of the graph nodes and fuse them into a feature matrix; Model training and testing module: normalizes the feature data and divides it into training set and test set. The CNN-Inception model is trained based on the training set and tested with the test set. The performance indicator is the accuracy of fault diagnosis. The diagnosis module extracts the features of the planetary gearbox signal collected in real time and inputs it into the trained CNN-Inception model for diagnosis.

9. A computer-readable storage medium storing one or more programs, characterized in that: The one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any one of the methods of claims 1 to 7.

10. A computing device, characterized in that: include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any one of the methods described in claims 1 to 7.

Images