Implementation method of immersive visualization system for high-speed train bogie mechanical structure

Abstract

The invention relates to an implementation method of an immersive visualization system for a high-speed train bogie mechanical structure. The implementation method comprises the following steps: step 1, preprocessing a bullet train bogie mechanical structure model; step 2, connection matrix construction and mechanical connection detection oriented to a complex mechanical structure; and step 3, generating a mechanical part set based on complex topological graph segmentation, and the like. The method disclosed by the invention has the superior technical effects that through an efficient visualization auxiliary technology in a multi-scale and multi-instance scene, association among all mechanical parts in a railway system can be accurately established, a clear connection network is formed, a complex mechanical structure can be comprehensively and deeply presented, and the reliability of the railway system is improved. According to the method and the system, a plurality of interactive modes such as an overall mode, a gathering mode and a part mode are creatively provided, so that a user can check different levels of the mechanical structure, visual interaction with the mechanical structures of different levels is realized through natural grasping operation, and a more comprehensive and deeper opportunity for knowing the mechanical structure is provided for the user.

Description

Technical Field

[0001] The present invention relates to the field of immersive visualization and human-computer interaction in digital twin technology, and in particular to an immersive visualization navigation method and system for the mechanical structure of a high-speed train bogie. Background Art

[0002] With the development of China's railways, the mileage of railways has been continuously extended and the scope has gradually expanded, and railway operation and maintenance have ushered in more severe challenges. In this context, the introduction of remote operation and maintenance technology and digital twin technology is profoundly changing the operation and maintenance mode of the railway system. Traditional on-site operation and maintenance seems to be unable to cope with the scale and diversity of the railway system. Remote operation and maintenance technology closely connects various nodes of China's railways by building an efficient network communication system, realizing real-time monitoring and data aggregation of the huge railway network. Digital twin technology, supported by virtual models, can provide accurate simulation and real-time modeling for China's railway system to adapt to the growing scale of railways. Data collection and analysis have brought more efficient preventive maintenance support to the expanding railway system, effectively reducing maintenance costs and improving equipment availability. The intelligent fault diagnosis system combined with digital twin technology provides China's railway system with the ability to locate and repair faults in real time.

[0003] The high-speed train bogie is a key component on a high-speed train. It is located at the bottom of the train and is responsible for supporting and connecting the train body and wheels. Its main function is to enable the train to run smoothly on the railway track and turn when necessary. The bogie is usually composed of a series of mechanical parts, including the suspension system, steering device, axle, etc. The suspension system is used to reduce the vibration of the train when it passes through uneven tracks, while the steering device allows the train to turn at railway intersections, switches, etc.

[0004] Ensuring the normal operation of the bogie is crucial to the safe and stable operation of high-speed trains. First, the bogie directly affects the handling performance of the train, ensuring that the train can travel along the predetermined path and avoid accidents such as derailment. Secondly, through an effective suspension system and steering device, the bogie can reduce the bumps and vibrations during the train's travel and improve the riding comfort of passengers. In addition, a properly functioning bogie can also help reduce the lateral force of the train when it is traveling on a curve, improve the stability of the train, and thus reduce the wear and tear of the vehicle and energy consumption.

[0005] With the development of remote operation and maintenance technology for railway systems, how to enable technicians to participate in the operation and maintenance of bogies more effectively has become an important issue. First, the remote operation and maintenance system needs to provide a highly flexible visual interface to support technicians to monitor the operating status of the bogie in real time. Through digital twin technology, the system can generate an accurate virtual model to show the internal structure of the bogie and the operation of each component. In addition, the remote operation and maintenance system should integrate sensor data and real-time monitoring information to provide a comprehensive health status assessment, so that technicians can fully understand the performance and operation of the bogie.

[0006] In addition, solving the problem of technicians participating in the operation and maintenance of bogies in remote operation and maintenance requires comprehensive consideration of the implementation environment of remote operation and maintenance. The current challenge is that it is difficult for technicians to intuitively perceive the structure and operating status of the bogie, and there is a lack of convenience in field operations. Therefore, an innovative solution is urgently needed to improve the operating efficiency and operation and maintenance quality of technicians. By introducing immersive virtual reality technology, technicians can remotely enter the virtual environment of the bogie and deeply experience the operation therein, so as to more intuitively perceive the details of the structure and status. This immersive experience will provide technicians with richer perceptual information and create a more practical and efficient operation method for them to perform bogie operation and maintenance in a remote environment. Through interactive design, technicians can interact with the bogie in the virtual environment in a more natural and intuitive way, including gestures, touch and other methods. Such an immersive interactive method will provide technicians with a new operation and maintenance experience, enabling them to better participate in the remote operation and maintenance of the bogie and solve the current operation and maintenance challenges. Summary of the invention

[0007] In order to achieve the purpose of the present invention, the present invention provides a method for implementing an immersive visualization system for a high-speed train bogie mechanical structure.

[0008] The implementation method of the immersive visualization system comprises the following steps:

[0009] Step 1: Preprocessing of the mechanical structure model of the EMU bogie;

[0010] Step 2: Connection matrix construction and mechanical connection detection for complex mechanical structures;

[0011] Step 3: Generate a set of mechanical parts based on complex topological graph segmentation;

[0012] Step 4: Correction of model splitting direction based on geometric features;

[0013] Step 5: Generate a hierarchical structure based on the mechanical structure topology diagram;

[0014] Step 6: Build an immersive remote operation and maintenance environment.

[0015] Furthermore, in step 1, the mechanical structure model of the motor vehicle bogie is preprocessed, including:

[0016] Step 1.1 The EMU bogie model is composed of 3DFACE. These data structures geometrically represent a three-dimensional plane, which is a triangle or a quadrilateral. In CAD, defining a 3DFACE requires sequentially specifying the four vertices of the 3DFACE. The four vertices are usually arranged in a clockwise or counterclockwise order. After reading the CAD file, the 3DFACE structure is parsed into a set of points and edges. A 3DFACE corresponds to a set of points and a set of edges, which correspond to the connection relationship between the corresponding vertices of the 3DFACE.

[0017] Step 1.2 Considering that the CAD file records the detailed structure of the three-dimensional mechanical model and should truly and accurately reflect the relationship between all parts of the machine and the relationship within the parts, there are no common vertices between the various parts that constitute the machine, that is, there is a new part hierarchy between the overall mechanical structure presented in the CAD file and the 3DFACE structure that constitutes the CAD file. The elements of this hierarchy correspond to independent part structures, guiding the process of assembling the parts into the entire car body. In order to generate the part hierarchy, it is necessary to obtain the 3DFACE corresponding to each part. Since there are no common vertices between the various parts that constitute the machine, for any mechanical part, the vertices and edges should form a strongly connected graph: any vertex that constitutes the mechanical part can reach another vertex through the edge; and there is no connection between the corresponding drawings of different mechanical parts;

[0018] Step 1.3 After the entire system reads the 3D CAD drawing, the point set P and edge set L on the entire drawing are obtained. For each edge L i , can find the corresponding point P in the point set i1 and P i2 Based on this, it is determined that the two points belong to the same independent set, that is, the two points belong to the same part, and a new set Q is obtained, in which each element is a point set, and there is an edge connecting the points belonging to the set to determine that each element in Q is a subset of the vertex set of the independent part. In the system, a set P is established based on the edge set E, and it is continuously checked whether there are shared points between different sets in the set P. If there are common vertices, the two sets are merged, and the above process is repeated until there are no common vertices between different sets in the set P. At this time, each element in P is the point set of each independent part. For each element p in P, find the edge e whose all vertices are in p, and then combine the edges and vertices together to generate the CAD data partcad of each part, and then generate PartCAD.

[0019] Furthermore, in step 2, the connection matrix construction and mechanical connection detection for complex mechanical structures are as follows:

[0020] Step 2.1 In the EMU bogie mechanical system, a connection matrix is ​​used to record the connection information of the EMU bogie mechanical structure. The connection matrix is ​​used to describe the interference or collision relationship between different components in the EMU bogie mechanical system. The rows and columns in the matrix represent different components in the EMU bogie mechanical system, and each element indicates whether there is interference or collision between the corresponding components, usually represented by a binary value. Creating a connection matrix involves a detailed analysis of the geometry, motion path, and workspace of the EMU bogie mechanical system. This will lead to inconsistencies between the 3D drawings and the actual EMU bogie mechanical structure. The inaccuracy of the 3D model is caused by the uncertainty of the connection representation in the EMU bogie mechanical engineering drawings and the mismatch between the visual representation on the 3D drawings and the actual structure.

[0021] Step 2.2 Considering the difficulty of creating a connection matrix, a specific method is used to identify and extract the connections between components in the EMU bogie mechanical system. By analyzing the 3D models of two EMU bogie mechanical components, when they have overlapping volumes, it indicates that there is an interference fit or a transition fit. This method determines the connection relationship between the components by identifying the area where the volumes overlap. In order to deal with the clearance fit, the 3D model of the EMU bogie mechanical component is slightly expanded, and the clearance fit is converted into other fit methods. Based on the standard physics engine, the generated mesh collider is used to expand and detect the overlapping volume to extract the connection information. Overall, the connection detection algorithm realizes the visualization of the EMU bogie mechanical structure, and the connection information is updated to the connection matrix;

[0022] Step 2.3 In order to more comprehensively represent the mechanical structure of the EMU bogie, a topological diagram of the mechanical structure of the EMU bogie is generated. The topological diagram of the mechanical structure of the EMU bogie is represented by converting the connection matrix into a graph, in which nodes represent the mechanical components of the EMU bogie, edges represent the connections between these components, and the weights on the edges represent the distances between the mechanical components of the EMU bogie. This process provides users with an intuitive overall picture of the mechanical structure of the EMU bogie, facilitating the understanding and analysis of the EMU bogie structure.

[0023] Furthermore, in step 3, a set of mechanical parts segmented based on the complex topological graph is generated:

[0024] Step 3.1 Based on the mechanical structure topology graph, the mechanical structure is divided by graph partitioning to generate a mechanical set. During the graph partitioning process, an initial solution for graph partitioning is created by center expansion, and the initial solution is optimized based on the K-algorithm.

[0025] Step 3.2: To generate a mechanical set that is more consistent with the understanding of the mechanical structure, a central node is specified to generate an initial solution. Based on the central node, an initial cluster set is generated by central expansion. The initial cluster set is updated through an iterative process. In each iteration, for each node in the graph that does not belong to any initial cluster, the node is added to the initial cluster adjacent to it with the shortest distance. The distance is defined as the minimum weight of the edge connecting the node to the initial cluster. After processing all nodes, the segmentation of the initial solution of the EMU bogie mechanical structure topology graph is generated.

[0026] Step 3.3 Based on the initial solution, apply K-algorithm to optimize the segmentation result. K-algorithm is derived from the traditional K-means algorithm, but uses an incremental method to assign nodes to the nearest cluster instead of calculating the distance to the cluster average. In order to optimize the segmentation result, K-algorithm calculates the incremental value of moving the specified node j from the old cluster X to the new cluster Y. The incremental value is obtained by calculating the change of the cost function in X and Y. The cost function is defined as:

[0027] Δf(j, X, Y) = f(C′ X )+f(C′ Y )-f(C X )-f(C Y )

[0028] ......(1),

[0029] By choosing the inverse internal weight IIW as the loss function, then:

[0030]

[0031] Where: M is the sum of the edge weights in the weighted graph structure, W i is the sum of the internal weights in cluster i, and the value of k represents the number of clusters. In this cost function, a smaller result indicates a better segmentation. Based on this cost function, the inverse weighting provides a negative value in the derivative of the cost function, which can balance the number of nodes in different clusters. In addition, since the empty cluster has an infinite IIW value, the IIW cost function can effectively reduce the existence of empty clusters. By applying IIW to the incremental value, the incremental function of the K-algorithm is obtained, as shown in the following formula (2):

[0032]

[0033] Based on center expansion and K-algorithm, an optimized graph segmentation result can be obtained in the topological graph of the mechanical structure of the EMU bogie. This result corresponds to different mechanical sets in the mechanical structure, and constitutes the basis of the mechanical set hierarchy in the mechanical system hierarchy. By combining center expansion and K-algorithm, an optimized graph segmentation result is obtained. This result corresponds to different mechanical sets in the mechanical structure, and constitutes the basis of the mechanical set hierarchy in the mechanical system hierarchy.

[0034] Furthermore, in step 4, based on the hierarchical structure generation of the mechanical structure topology graph, the classification and management method based on the topological order of mechanical parts is aimed at solving the problem of a large number of mechanical parts in the mechanical parts clustering and the lack of clarity of their internal structures:

[0035] On the basis of the mechanical parts set generation, the mechanical components are comprehensively classified. Based on the mechanical parts set, by combining this with the overall connection matrix generated based on the mechanical structure, the individual connection matrix and the EMU bogie mechanical structure topology corresponding to each mechanical parts set are extracted. The generated structure diagram is a sub-graph of the overall mechanical structure diagram, representing the connection relationship within the mechanical set;

[0036] In the mechanical parts set generation part, each mechanical parts set has a central mechanical part, corresponding to the central node on the topological graph, which allows to calculate the topological distance from each node on the topology to the central node, map the nodes and their topological distances to the corresponding mechanical sets, and label each mechanical part in the mechanical set based on its topological distance to allow the mechanical parts in the mechanical set to be classified;

[0037] The mechanical classification based on topological distance reflects the hierarchical relationship between mechanical components within the mechanical structure. The mechanical components in the category with smaller topological distance are more closely related to the central mechanical component, indicating that they form a core structure of mechanical components clustered around the central component. Conversely, the mechanical components with larger topological distance are more distant from the central mechanical component, indicating that they are scattered at the periphery of the mechanical set and play a supporting and auxiliary connection role.

[0038] Based on the hierarchical structure generation strategy of the mechanical structure topology graph, the topological structure of the internal components of the mechanical set is presented separately when displaying complex mechanical structures, so as to be able to analyze the internal structure of the mechanical cluster. On the basis of the topological layer, the mechanical components are re-managed at the mechanical set level, and each mechanical set matches the corresponding mechanical component topological structure to avoid the complexity of mechanical connection analysis and mechanical structure visualization on the overall connection graph. The classification operation of the mechanical components within the cluster realizes the hierarchical organization of the internal components of the mechanical cluster and also enhances the hierarchical relationship between the mechanical cluster and the individual mechanical components, which provides guidance for the subsequent visualization and anatomical operations of the mechanical structure.

[0039] Furthermore, in step 5, the model splitting direction is corrected based on geometric features:

[0040] In order to realize the visualization of the internal structure of mechanical equipment, a set of visualization methods need to be designed. Based on the clustering of mechanical parts, disassembly animation is selected as the most intuitive visualization method, specifically:

[0041] In order to create a disassembly animation, it is necessary to solve the movement order of mechanical parts in the disassembly animation and the disassembly direction of each mechanical part. A disassembly strategy is designed to be executed from the highest topological layer to the lowest topological layer. The mechanical parts of the same topological layer are disassembled at the same time, which solves the problem of the movement order of mechanical parts. Mechanical parts of higher topological layers are usually the parts of the outermost mechanical structure. Starting to disassemble from these parts helps the observer to gradually understand the internal structure of the machine and perceive the assembly relationship between different layers of the machine. In areas composed of a large number of parts, disassembling parts of the same topological layer at the same time not only shortens the duration of the animation, but also enables the observer to focus on the interconnection of a group of mechanical parts, which helps to better understand specific mechanical areas.

[0042] In order to determine the disassembly direction of mechanical parts, a strategy based on part center and axial orientation is proposed. For a specified mechanical part, the hierarchical structure generated in the topological layer generation part is used to obtain all interconnected mechanical parts associated with the specified mechanical part. When considering the disassembly animation of the mechanical part, the inner parts are displayed by moving the outermost mechanical parts to define the indicated direction. For each mechanically connected part, the indicated direction represents the optimal direction in which the specified mechanical part should be disassembled. Due to the interconnection between mechanical parts and the hierarchical structure based on the topological layer, the outermost mechanical part is removed first. For any mechanical part located in the next layer, an interference vector is constructed, denoted as I. Based on the vector difference between its geometric center and the geometric center of the specified mechanical part, the interference vector represents the influence of the mechanical part on the disassembly of the specified mechanical part.

[0043] To ensure a smooth disassembly process and eliminate the influence of the disassembled parts, all the interference vectors i corresponding to all mechanical parts under one layer of the specified part are stacked and normalized. This process results in an optimized guidance direction D, which represents the best direction to move the mechanical part to reveal the mechanical structure;

[0044] By proposing a correction axis, denoted as A, to adjust the optimal guidance direction D, this correction aims to optimize the disassembly animation by resolving the mismatch between the optimal guidance direction D and the actual shape of the mechanical parts. Given that parts usually have complex shapes, there is no universal template available. In order to uniformly define the correction axis A for mechanical parts, the mechanical parts are firstly classified broadly. For those parts with obvious axial features (such as shafts and wheels), they are classified as having axial features. On the other hand, mechanical parts lacking obvious axial features (such as gearboxes and brake calipers) are classified as non-axial parts. Based on the classification of the oriented bounding box of the mechanical parts, the minimum bounding box is obtained for the 3D model of the complex mechanical parts, and then the generated cube is classified to facilitate the axial classification of the complex mechanical models.

[0045] Based on the axial classification, for mechanical parts with axial features, the axial direction corresponding to the oriented bounding box is extracted. This extracted axial direction acts as the correction axis A, allowing the mechanical part to move along its axial direction. This correction aligns the movement of the part with the actual assembly and disassembly direction of the mechanical structure. On the other hand, for non-axial parts, the correction axis A is defined as the vector DG of the maximum module. The vector DG is the projection between the projections of the optimal guiding direction D on the relative spatial coordinate axes. This ensures that when the optimal movement direction is uncertain, the non-axial part is guided to expand along the spatial axis so that the animation is aligned with the observer's perspective.

[0046] Based on the correction axis A, the projection vector DP of the optimal guidance direction DG on A is constructed to represent the correction movement direction of the mechanical component to improve the accuracy and consistency of the disassembly animation and ensure that the movement of the mechanical component is consistent with its actual assembly and disassembly direction or the observer's perspective. After generation, the disassembly direction is attached to the corresponding mechanical component of the mechanical component layer to enhance the disassembly animation in interactive navigation.

[0047] Furthermore, in step 6, an immersive remote operation and maintenance environment is constructed:

[0048] The user creates a hierarchical navigation through the UI interactive interface and configures relevant parameters. The user drags the mechanical model file to be displayed into the interface, and then attaches the model object in the corresponding script to replace the 3D model. This step is designed to simplify the user's operation of the display model and make it more intuitive.

[0049] In the UI interface created by the control split animation generation module, the user changes the number of central mechanical parts by controlling the parameters in the mechanical parts set generation script, and customizes the central mechanical parts by dragging the 3D model of the central mechanical parts to the specified position. This design enables the user to flexibly set the central mechanical parts used to guide the generation of the mechanical parts set. The implementation method of the immersive visualization system allows the user to sort the model according to the volume and number of faces, and delete parts with smaller volumes to improve the visualization effect. The user customizes the model simplification process by adjusting the threshold and the number of faces filter to make it more in line with the user's habits and needs.

[0050] After completing the configuration and construction of the immersive environment, the user interacts with the model at different levels to better view the details of the complex mechanical structure and demonstrate the presentation effects of the visualization configuration and user interaction modules. The user selects different interaction modes in the implementation method of the immersive visualization system to adapt to mechanical structures at different levels. The specific interaction modes include overall mode, collection mode and part mode. The overall mode allows the user to interact with the entire mechanical model to obtain a global view; the collection mode supports the user to explore and operate the mechanical part collection generated by the method of the immersive visualization system; the part mode provides operations for interacting with a single mechanical part. Based on different levels and interaction modes, the user uses grasping operations to interact with mechanical structures at different levels. This natural interaction method enables the user to intuitively experience the mechanical structure and enhances the user's sense of immersion. Through grasping interaction operations, the user simulates physical interaction with the actual machine, including assembly and inspection operations, so as to have a more comprehensive understanding of the mechanical structure.

[0051] Based on the topological hierarchy, users can view the mechanical structure in different interactive modes and increase their understanding of the mechanical functional areas and the order of mechanical assembly and disassembly through additional coloring and splitting animations;

[0052] In the implementation method of the immersive visualization system, the mechanical parts are colored in the immersive environment according to the hierarchical information of the mechanical structure. Such a design helps the user to more clearly identify the structures at different levels. The user can also adjust the transparency to reduce the occlusion between different levels when necessary, so that the structure at a specific level is more prominent.

[0053] Based on the generated disassembly direction, the implementation method of the immersive visualization system provides a disassembly animation function. The user chooses to perform a disassembly animation on the entire mechanical structure or a specific set of mechanical parts. The degree of disassembly is controlled by a slider. The user can dynamically display the disassembly process of the mechanical structure to help the user understand the connection relationship between the mechanical parts. This not only provides a visual effect, but also strengthens the user's operational perception of the mechanical structure.

[0054] The superior technical effects of the implementation method of the immersive visualization system of the present invention are as follows:

[0055] 1. The implementation method of the immersive visualization system described in the present invention can accurately establish the relationship between various mechanical components in the railway system through efficient visualization auxiliary technology in multi-scale and multi-instance scenarios, form a clear connection network, and enable complex mechanical structures to be presented comprehensively and in-depth.

[0056] 2. In terms of user interaction, the implementation method of the immersive visualization system described in the present invention creatively provides multiple interaction modes such as overall mode, collection mode and part mode, so that users can flexibly view different levels of mechanical structures, and achieve intuitive interaction with mechanical structures at different levels through natural grasping operations, providing users with the opportunity to understand the mechanical structure more comprehensively and deeply, and enhancing the user's perception and operation experience in the immersive environment.

[0057] 3. The implementation method of the immersive visualization system described in the present invention further enhances the visualization effect of the mechanical structure through innovative functions such as structural layer coloring and disassembly animation.

[0058] 4. The implementation method of the immersive visualization system described in the present invention helps users to clearly identify structures at different levels through structural hierarchy coloring, while the disassembly animation helps users to have a deeper understanding of the connection relationship between mechanical parts by dynamically displaying the disassembly process of the mechanical structure, thereby enhancing the user's perception of the operation of the mechanical structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a schematic diagram of model preprocessing in a specific embodiment of the present invention;

[0060] Figure 2 A schematic diagram of connection matrix construction and mechanical connection detection for complex mechanical structures in a specific embodiment of the present invention;

[0061] Figure 3 A schematic diagram of a collection of mechanical parts generated based on segmentation of a complex topological graph in a specific embodiment of the present invention;

[0062] Figure 4 It is a schematic diagram of the model splitting direction correction based on geometric features in a specific implementation manner of the present invention;

[0063] Figure 5 It is a schematic diagram of initializing an immersive environment in constructing an immersive remote operation and maintenance environment in a specific implementation manner of the present invention;

[0064] Figure 6A schematic diagram of an immersive complex model visualization strategy in the construction of an immersive remote operation and maintenance environment in a specific implementation manner of the present invention; DETAILED DESCRIPTION

[0065] In order to more clearly understand the above-mentioned objectives, features and advantages of the present invention, the specific implementation method of the immersive visualization system of the present invention is described in detail below in conjunction with the accompanying drawings.

[0066] The method for implementing the immersive visualization system described in the present invention comprises the following specific steps:

[0067] Step 1. Preprocessing of the mechanical structure model of the EMU bogie, including:

[0068] Step 1.1 The EMU bogie model is composed of 3DFACE. These data structures geometrically represent a three-dimensional plane, which is a triangle or a quadrilateral. In CAD, defining a 3DFACE requires sequentially specifying the four vertices of the 3DFACE. The four vertices are usually arranged in a clockwise or counterclockwise order. After reading the CAD file, the 3DFACE structure is parsed into a set of points and edges: a 3DFACE corresponds to a set of points and a set of edges, which correspond to the connection relationship between the corresponding vertices of the 3DFACE;

[0069] Step 1.2 Considering that the CAD file records the detailed structure of the three-dimensional mechanical model and should truly and accurately reflect the relationship between all parts of the machine and the relationship within the parts, there are no common vertices between the various parts that constitute the machine, otherwise the parts will "stick" to each other, making the mechanical structure recorded in the CAD file inaccurate. This is to say that there is a new part hierarchy between the overall mechanical structure presented in the CAD file and the 3DFACE structure that constitutes the CAD file. The elements of this hierarchy correspond to independent part structures, guiding the process of assembling the parts into the entire car body. In order to generate the part hierarchy, it is necessary to obtain the 3DFACE corresponding to each part. Since there are no common vertices between the various parts that constitute the machine, for any mechanical part, the vertices and edges should form a strongly connected graph: any vertex that constitutes a mechanical part can reach another vertex through an edge; and there is no connection between the corresponding drawings of different mechanical parts;

[0070] Step 1.3 After the entire system reads the 3D CAD drawing, the point set P and edge set L on the entire drawing are obtained. For each edge L i , can find the corresponding point P in the point set i1 and P i2Based on this, it is determined that the two points belong to the same independent set, that is, the two points belong to the same part, and a new set Q is obtained, in which each element is a point set, and there is an edge connecting the points belonging to the set, so as to determine that each element in Q is a subset of the vertex set of the independent part. In the system, a set P is established based on the edge set E, and it is continuously checked whether there are shared points between different sets in the set P. If there are common vertices, the two sets are merged, and the above process is repeated until there are no common vertices between different sets in the set P. At this time, each element in P is the point set of each independent part. For each element p in P, find the edge e whose all vertices are in p, then combine the edges and vertices together to generate the CAD data partcad of each part, and then generate PartCAD; Figure 1 Demonstrate all strategies for preprocessing the mechanical structure model of EMU bogies;

[0071] Step 2. Connection matrix construction and mechanical connection detection for complex mechanical structures:

[0072] Step 2.1 In the EMU bogie mechanical system, a connection matrix is ​​used to record the connection information of the EMU bogie mechanical structure. The connection matrix is ​​used to describe the interference or collision relationship between different components in the EMU bogie mechanical system. The rows and columns in the matrix represent different components in the EMU bogie mechanical system, and each element indicates whether there is interference or collision between the corresponding components, usually represented by a binary value. Creating a connection matrix involves a detailed analysis of the geometry, motion path, and workspace of the EMU bogie mechanical system. This will lead to inconsistencies between the 3D drawings and the actual EMU bogie mechanical structure. The inaccuracy of the 3D model is caused by the uncertainty of the connection representation in the EMU bogie mechanical engineering drawings and the mismatch between the visual representation on the 3D drawings and the actual structure.

[0073] Step 2.2 Considering the difficulty of creating a connection matrix, a method is used to identify and extract the connections between components in the EMU bogie mechanical system. By analyzing the 3D models of two EMU bogie mechanical components, when they have overlapping volumes, it indicates that there is an interference fit or a transition fit. This method determines the connection relationship between the components by identifying the area where the volumes overlap. In order to deal with clearance fit, the 3D model of the EMU bogie mechanical component is slightly expanded, and the clearance fit is converted into other fit methods. Based on the standard physics engine, the generated mesh collider is used to expand and detect overlapping volumes to extract the connection information. Overall, the connection detection algorithm realizes the visualization of the EMU bogie mechanical structure, and updates the connection information to the connection matrix. Figure 2 Demonstrate strategies for extracting connections between EMU bogie components;

[0074] Step 2.3 In order to more comprehensively represent the mechanical structure of the EMU bogie, a topological diagram of the mechanical structure of the EMU bogie is generated. The topological diagram of the mechanical structure of the EMU bogie is represented by converting the connection matrix into a graph, in which nodes represent the mechanical components of the EMU bogie, edges represent the connections between these components, and the weights on the edges represent the distances between the mechanical components of the EMU bogie. This process provides users with an intuitive overview of the mechanical structure of the EMU bogie, facilitating the understanding and analysis of the EMU bogie structure.

[0075] Step 3. Generation of mechanical parts set based on complex topological graph segmentation:

[0076] Step 3.1 Based on the mechanical structure topology graph, the mechanical structure is divided by graph partitioning to generate a mechanical set. During the graph partitioning process, an initial solution for graph partitioning is created by center expansion, and the initial solution is optimized based on the K-algorithm.

[0077] Step 3.2: To generate a mechanical set that is more consistent with the understanding of the mechanical structure, a central node is specified to generate an initial solution. Based on the central node, an initial cluster set is generated by central expansion. The initial cluster set is updated through an iterative process. In each iteration, for each node in the graph that does not belong to any initial cluster, the node is added to the initial cluster adjacent to it with the shortest distance. The distance is defined as the minimum weight of the edge connecting the node to the initial cluster. After processing all nodes, the segmentation of the initial solution of the EMU bogie mechanical structure topology graph is generated.

[0078] Step 3.3 Based on the initial solution, apply K-algorithm to optimize the segmentation result. K-algorithm is derived from the traditional K-means algorithm, but uses an incremental method to assign nodes to the nearest cluster instead of calculating the distance to the cluster average. In order to optimize the segmentation result, K-algorithm calculates the incremental value of moving the specified node j from the old cluster X to the new cluster Y. The incremental value is obtained by calculating the change of the cost function in X and Y. The cost function is defined as:

[0079] Δf(j, X, Y) = f(C′ X )+f(C′ Y )-f(C X )-f(C Y )

[0080] ......(1),

[0081] By choosing the inverse internal weight IIW as the loss function, then:

[0082]

[0083] In the above formula, M is the sum of the edge weights in the weighted graph structure, and W i is the sum of the internal weights in cluster i, and the value of k represents the number of clusters. In this cost function, a smaller result indicates a better segmentation. Based on this cost function, the inverse weighting provides a negative value in the derivative of the cost function, which can balance the number of nodes in different clusters. In addition, since the empty cluster has an infinite IIW value, the IIW cost function can effectively reduce the existence of empty clusters. By applying IIW to the incremental value, the incremental function of the K-algorithm is obtained, as shown in the following formula (2):

[0084]

[0085] Based on center expansion and K-algorithm, an optimized graph segmentation result can be obtained in the topological graph of the mechanical structure of the EMU bogie. This result corresponds to different mechanical sets in the mechanical structure and constitutes the basis of the mechanical set hierarchy in the mechanical system hierarchy. By integrating center expansion and K-algorithm, an optimized graph segmentation result is obtained. This result corresponds to different mechanical sets in the mechanical structure and constitutes the basis of the mechanical set hierarchy in the mechanical system hierarchy. Figure 3 Demonstrate the strategy of realizing EMU component aggregation based on graph structure;

[0086] Step 4. Hierarchical structure generation based on mechanical structure topology graph and classification and management method based on mechanical parts topology order are aimed at solving the problem of a large number of mechanical parts in mechanical parts clustering and the lack of clarity of their internal structures:

[0087] On the basis of the mechanical parts set generation, the mechanical components are comprehensively classified. Based on the mechanical parts set, by combining this with the overall connection matrix generated based on the mechanical structure, the individual connection matrix and the EMU bogie mechanical structure topology corresponding to each mechanical parts set are extracted. The generated structure diagram is a sub-graph of the overall mechanical structure diagram, representing the connection relationship within the mechanical set;

[0088] In the mechanical parts set generation part, each mechanical parts set has a central mechanical part, corresponding to the central node on the topological graph, which allows to calculate the topological distance from each node on the topology to the central node, map the nodes and their topological distances to the corresponding mechanical sets, and label each mechanical part in the mechanical set based on its topological distance to allow the mechanical parts in the mechanical set to be classified;

[0089] The mechanical classification based on topological distance reflects the hierarchical relationship between mechanical components within the mechanical structure. The mechanical components in the category with smaller topological distance are more closely related to the central mechanical component, indicating that they form a core structure of mechanical components clustered around the central component. Conversely, the mechanical components with larger topological distance are more distant from the central mechanical component, indicating that they are scattered at the periphery of the mechanical set and play a supporting and auxiliary connection role.

[0090] Based on the hierarchical structure generation strategy of the mechanical structure topology map, the topological structure of the internal components of the mechanical set is presented separately to analyze the internal structure of the mechanical cluster. On the basis of the topological layer, the mechanical components are re-managed at the mechanical set level. Each mechanical set matches the corresponding mechanical component topological structure to avoid the complexity of mechanical connection analysis and mechanical structure visualization on the overall connection map. The classification operation of the mechanical components within the cluster realizes the hierarchical organization of the internal components of the mechanical cluster and enhances the hierarchical relationship between the mechanical cluster and the individual mechanical components, which provides guidance for the subsequent visualization and anatomical operation of the mechanical structure.

[0091] Step 5. Model splitting direction correction based on geometric features:

[0092] In order to realize the visualization of the internal structure of mechanical equipment, a set of visualization methods need to be designed. Based on the clustering of mechanical parts, disassembly animation is selected as the most intuitive visualization method:

[0093] In order to create a disassembly animation, it is necessary to solve the movement order of mechanical parts in the disassembly animation and the disassembly direction of each mechanical part. A disassembly strategy is designed to be executed from the highest topological layer to the lowest topological layer. The mechanical parts of the same topological layer are disassembled at the same time, which solves the problem of the movement order of mechanical parts. Mechanical parts of higher topological layers are usually the parts of the outermost mechanical structure. Starting to disassemble from these parts helps the observer to gradually understand the internal structure of the machine and perceive the assembly relationship between different layers of the machine. In areas composed of a large number of parts, disassembling parts of the same topological layer at the same time shortens the duration of the animation. The observer can focus on the interconnection of a group of mechanical parts, which helps to better understand a specific mechanical area.

[0094] In order to determine the disassembly direction of mechanical parts, a strategy based on part center and axial orientation is proposed. For a specified mechanical part, the hierarchical structure generated in the topological layer generation part is used to obtain all interconnected mechanical parts associated with the specified mechanical part. When considering the disassembly animation of the mechanical part, the inner parts are displayed by moving the outermost mechanical parts to define the indicated direction. For each mechanically connected part, the indicated direction represents the optimal direction in which the specified mechanical part should be disassembled. Due to the interconnection between mechanical parts and the hierarchical structure based on the topological layer, the outermost mechanical part is usually removed first. For any mechanical part located in the next layer, an interference vector is constructed, denoted as I. Based on the vector difference between its geometric center and the geometric center of the specified mechanical part, the interference vector represents the influence of the mechanical part on the disassembly of the specified mechanical part.

[0095] To ensure a smooth disassembly process and eliminate the influence of the disassembled parts, all the interference vectors i corresponding to all mechanical parts under one layer of the specified part are stacked and normalized. This process results in an optimized guidance direction D, which represents the best direction to move the mechanical part to reveal the mechanical structure;

[0096] By proposing a correction axis, denoted as A, to adjust the optimal guidance direction D, this correction aims to optimize the disassembly animation by resolving the mismatch between the optimal guidance direction D and the actual shape of the mechanical parts. Given that parts usually have complex shapes, there is no universal template available. In order to uniformly define the correction axis A for mechanical parts, the mechanical parts are firstly classified broadly. For those parts with obvious axial features (such as shafts and wheels), they are classified as having axial features. On the other hand, mechanical parts lacking obvious axial features (such as gearboxes and brake calipers) are classified as non-axial parts. Based on the classification of the oriented bounding box of the mechanical parts, the minimum bounding box is obtained for the 3D model of the complex mechanical parts, and then the generated cube is classified to facilitate the axial classification of the complex mechanical models.

[0097] Based on the axial classification, for mechanical parts with axial features, the axial direction corresponding to the oriented bounding box is extracted. This extracted axial direction acts as the correction axis A, allowing the mechanical part to move along its axial direction. This correction aligns the movement of the part with the actual assembly and disassembly direction of the mechanical structure. On the other hand, for non-axial parts, the correction axis A is defined as the vector DG of the maximum module. The vector DG is the projection between the projections of the optimal guiding direction D on the relative spatial coordinate axes. This ensures that when the optimal movement direction is uncertain, the non-axial part is guided to expand along the spatial axis so that the animation is aligned with the observer's perspective.

[0098] Based on the correction axis A, the projection vector DP of the optimal guidance direction DG on A is constructed to represent the correction movement direction of the mechanical component to improve the accuracy and consistency of the disassembly animation and ensure that the movement of the mechanical component is consistent with its actual assembly and disassembly direction or the observer's perspective. After generation, the disassembly direction is attached to the corresponding mechanical component of the mechanical component layer to enhance the disassembly animation in the interactive navigation. Figure 4 Demonstrated a strategy for splitting the animation generation module;

[0099] Step 6. Build an immersive remote operation and maintenance environment:

[0100] The user creates a hierarchical navigation through the UI interactive interface and configures relevant parameters. The user drags the mechanical model file to be displayed into the interface, and then attaches the model object in the corresponding script to replace the 3D model. This step is designed to simplify the user's operation of the display model and make it more intuitive.

[0101] In the interface of the control split animation generation module, the user changes the number of central mechanical parts by controlling the parameters in the mechanical parts set generation script, and customizes the central mechanical parts by dragging the 3D model of the central mechanical parts to the specified position. This design enables the user to flexibly set the central mechanical parts used to guide the generation of mechanical parts sets. The implementation method of the immersive visualization system allows the user to sort the model according to the volume and number of faces, and delete parts with smaller volumes to improve the visualization effect. The user customizes the model simplification process by adjusting the threshold and the number of faces filter to make it more in line with the user's habits and needs. Figure 5 Demonstrate the method of controlling the generation of split animation through interface operations.

[0102] After configuring and building the immersive environment, users interact with the model at different levels to better view the details of the complex mechanical structure. Figure 6 The presentation effect of the visualization configuration and user interaction module is displayed. The user selects different interaction modes in the implementation method of the immersive visualization system to adapt to mechanical structures at different levels. The specific interaction modes include overall mode, collection mode and part mode. The overall mode allows the user to interact with the entire mechanical model to obtain a global view; the collection mode supports the user to explore and operate the mechanical part collection generated by the method of the immersive visualization system; the part mode provides operations for interacting with a single mechanical part. Based on different levels and interaction modes, the user uses grasping operations to interact with mechanical structures at different levels. This natural interaction method enables the user to intuitively experience the mechanical structure and enhances the user's sense of immersion. Through grasping interaction operations, the user simulates physical interaction with the actual machine, including assembly and inspection operations, so as to have a more comprehensive understanding of the mechanical structure.

[0103] Based on the topological hierarchy, users can view the mechanical structure in different interactive modes and increase their understanding of the mechanical functional areas and the order of mechanical assembly and disassembly through additional coloring and splitting animations. Figure 6 It also shows how rendering and animation actually look;

[0104] In the implementation method of the immersive visualization system, the mechanical parts are colored in the immersive environment according to the hierarchical information of the mechanical structure. Such a design helps the user to more clearly identify the structures at different levels. The user can also adjust the transparency to reduce the occlusion between different levels when necessary, so that the structure at a specific level is more prominent.

[0105] Based on the generated disassembly direction, the implementation method of the immersive visualization system provides a disassembly animation function. The user chooses to perform a disassembly animation on the entire mechanical structure or a specific set of mechanical parts. The degree of disassembly is controlled by a slider. The user can dynamically display the disassembly process of the mechanical structure to help the user understand the connection relationship between the mechanical parts. This not only provides a visual effect, but also strengthens the user's operational perception of the mechanical structure.

[0106] The present invention is not limited by the above embodiments. The above embodiments and descriptions are only for explaining the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention may be subject to various changes and improvements, which are within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the appended claims.

Claims

1. A method for implementing an immersive visualization system for the mechanical structure of a high-speed train bogie, characterized in that: The following steps are involved: Step 1: Preprocessing of the mechanical structure model of the EMU bogie; Step 2: Connection matrix construction and mechanical connection detection for complex mechanical structures; Step 3: Generate a set of mechanical parts based on complex topological graph segmentation; Step 4: Correction of model splitting direction based on geometric features; Step 5: Generate a hierarchical structure based on the mechanical structure topology diagram; Step 6: Build an immersive remote operation and maintenance environment.

2. According to the method for implementing the immersive visualization system for the mechanical structure of a high-speed train bogie according to claim 1, the preprocessing of the mechanical structure model of the motor vehicle bogie in step 1 comprises: Step 1.1 The EMU bogie model is composed of 3DFACE. These data structures geometrically represent a three-dimensional plane, which is a triangle or a quadrilateral. In CAD, defining a 3DFACE requires sequentially specifying the four vertices of the 3DFACE. The four vertices are usually arranged in a clockwise or counterclockwise order. After reading the CAD file, the 3DFACE structure is parsed into a set of points and edges. A 3DFACE corresponds to a set of points and a set of edges, which correspond to the connection relationship between the corresponding vertices of the 3DFACE. Step 1.2 Considering that the CAD file records the detailed structure of the three-dimensional mechanical model and should truly and accurately reflect the relationship between all parts of the machine and the relationship within the parts, there are no common vertices between the various parts that constitute the machine, that is, there is a new part hierarchy between the overall mechanical structure presented in the CAD file and the 3DFACE structure that constitutes the CAD file. The elements of this hierarchy correspond to independent part structures, guiding the process of assembling the parts into the entire car body. In order to generate the part hierarchy, it is necessary to obtain the 3DFACE corresponding to each part. Since there are no common vertices between the various parts that constitute the machine, for any mechanical part, the vertices and edges should form a strongly connected graph, and any vertex that constitutes the mechanical part can reach another vertex through the edge; and there is no connection between the corresponding drawings of different mechanical parts; Step 1.3 After the entire system reads the 3D CAD drawing, the point set P and edge set L on the entire drawing are obtained. For each edge L i , can find the corresponding point P in the point set i1 and P i2 Based on this, it is determined that the two points belong to the same independent set, that is, the two points belong to the same part, and a new set Q is obtained, in which each element is a point set, and there is an edge connecting the points belonging to the set to determine that each element in Q is a subset of the vertex set of the independent part. In the system, a set P is established based on the edge set E, and it is continuously checked whether there are shared points between different sets in the set P. If there are common vertices, the two sets are merged, and the above process is repeated until there are no common vertices between different sets in the set P. At this time, each element in P is the point set of each independent part. For each element p in P, find the edge e whose all vertices are in p, and then combine the edges and vertices together to generate the CAD data partcad of each part, and then generate PartCAD.

3. According to the method for implementing the immersive visualization system for the mechanical structure of a high-speed train bogie as claimed in claim 1, the connection matrix construction and mechanical connection detection for complex mechanical structures as described in step 2 include: Step 2.1 In the EMU bogie mechanical system, a connection matrix is ​​used to record the connection information of the EMU bogie mechanical structure. The connection matrix is ​​used to describe the interference or collision relationship between different components in the EMU bogie mechanical system. The rows and columns in the matrix represent different components in the EMU bogie mechanical system, and each element indicates whether there is interference or collision between the corresponding components, usually represented by a binary value. Creating a connection matrix involves a detailed analysis of the geometry, motion path, and workspace of the EMU bogie mechanical system. This will lead to inconsistencies between the 3D drawings and the actual EMU bogie mechanical structure. The inaccuracy of the 3D model is caused by the uncertainty of the connection representation in the EMU bogie mechanical engineering drawings and the mismatch between the visual representation on the 3D drawings and the actual structure. Step 2.2 Considering the difficulty of creating a connection matrix, a method is used to identify and extract the connections between components in the EMU bogie mechanical system. By analyzing the 3D models of two EMU bogie mechanical components, when they have overlapping volumes, it indicates that there is an interference fit or a transition fit. This method determines the connection relationship between the components by identifying the area where the volumes overlap. In order to deal with the clearance fit, the 3D model of the EMU bogie mechanical component is slightly expanded, and the clearance fit is converted into other fit methods. Based on the standard physics engine, the generated mesh collider is used to expand and detect the overlapping volume to extract the connection information. Overall, the connection detection algorithm realizes the visualization of the EMU bogie mechanical structure, and the connection information is updated to the connection matrix. Step 2.3 In order to more comprehensively represent the mechanical structure of the EMU bogie, a topological diagram of the mechanical structure of the EMU bogie is generated. The topological diagram of the mechanical structure of the EMU bogie is represented by converting the connection matrix into a graph, in which nodes represent the mechanical components of the EMU bogie, edges represent the connections between these components, and the weights on the edges represent the distances between the mechanical components of the EMU bogie. This process provides users with an intuitive overall picture of the mechanical structure of the EMU bogie, facilitating the understanding and analysis of the EMU bogie structure.

4. According to the method for implementing an immersive visualization system for a high-speed train bogie mechanical structure as claimed in claim 1, the generation of a set of mechanical parts based on complex topological graph segmentation in step 3 comprises: Step 3.1 Based on the mechanical structure topology graph, the mechanical structure is divided by graph partitioning to generate a mechanical set. During the graph partitioning process, an initial solution for graph partitioning is created by center expansion, and the initial solution is optimized based on the K-algorithm. Step 3.2: To generate a mechanical set that is more consistent with the understanding of the mechanical structure, a central node is specified to generate an initial solution. Based on the central node, an initial cluster set is generated by central expansion. The initial cluster set is updated through an iterative process. In each iteration, for each node in the graph that does not belong to any initial cluster, the node is added to the initial cluster adjacent to it with the shortest distance. The distance is defined as the minimum weight of the edge connecting the node to the initial cluster. After processing all nodes, the segmentation of the initial solution of the EMU bogie mechanical structure topology graph is generated. Step 3.3 Based on the initial solution, K-algorithm is applied to optimize the segmentation result. In order to optimize the segmentation result, K-algorithm calculates the incremental value of moving the specified node j from the old cluster X to the new cluster Y. The incremental value is obtained by calculating the change of the cost function in X and Y. The cost function is defined as: Δf(j,X,Y)=f(C′ X )+f(C′ Y )-f(C X )-f(C Y )......(1), By choosing the inverse internal weight IIW as the loss function, then: Where: M is the sum of the edge weights in the weighted graph structure, W i is the sum of the internal weights in cluster i, and the value of k represents the number of clusters. In this cost function, a smaller result indicates a better segmentation. Based on this cost function, the inverse weighting provides a negative value in the derivative of the cost function, which can balance the number of nodes in different clusters. In addition, since the empty cluster has an infinite IIW value, the IIW cost function can effectively reduce the existence of empty clusters. By applying IIW to the incremental value, the incremental function of the K-algorithm is obtained, as shown in the following formula (2): Based on center expansion and K-algorithm, an optimized graph segmentation result can be obtained in the topological graph of the mechanical structure of the EMU bogie. This result corresponds to different mechanical sets in the mechanical structure, and constitutes the basis of the mechanical set hierarchy in the mechanical system hierarchy. By combining center expansion and K-algorithm, an optimized graph segmentation result is obtained. This result corresponds to different mechanical sets in the mechanical structure, and constitutes the basis of the mechanical set hierarchy in the mechanical system hierarchy.

5. According to the method for implementing the immersive visualization system for the mechanical structure of a high-speed train bogie according to claim 1, the hierarchical structure generation based on the mechanical structure topology diagram in step 4: On the basis of the mechanical parts set generation, the mechanical components are comprehensively classified. Based on the mechanical parts set, by combining this with the overall connection matrix generated based on the mechanical structure, the individual connection matrix and the EMU bogie mechanical structure topology corresponding to each mechanical parts set are extracted. The generated structure diagram is a sub-graph of the overall mechanical structure diagram, representing the connection relationship within the mechanical set; In the mechanical parts set generation part, each mechanical parts set has a central mechanical part, corresponding to the central node on the topological graph, which allows to calculate the topological distance from each node on the topology to the central node, map the nodes and their topological distances to the corresponding mechanical sets, and label each mechanical part in the mechanical set based on its topological distance to allow the mechanical parts in the mechanical set to be classified; Based on the hierarchical structure generation strategy of the mechanical structure topology graph, the topological structure of the internal components of the mechanical set is presented separately when displaying complex mechanical structures, so as to be able to analyze the internal structure of the mechanical cluster. On the basis of the topological layer, the mechanical components are re-managed at the mechanical set level, and each mechanical set matches the corresponding mechanical component topological structure to avoid the complexity of mechanical connection analysis and mechanical structure visualization on the overall connection graph. The classification operation of the mechanical components within the cluster realizes the hierarchical organization of the internal components of the mechanical cluster and also enhances the hierarchical relationship between the mechanical cluster and the individual mechanical components, which provides guidance for the subsequent visualization and anatomical operations of the mechanical structure.

6. According to the method for implementing the immersive visualization system for the mechanical structure of a high-speed train bogie as claimed in claim 1, the model splitting direction correction based on geometric features in step 5 selects the disassembly animation as the most intuitive visualization method, specifically: In order to determine the disassembly direction of mechanical parts, a strategy based on part center and axial orientation is proposed. For a specified mechanical part, the hierarchical structure generated in the topological layer generation part is used to obtain all interconnected mechanical parts associated with the specified mechanical part. When considering the disassembly animation of the mechanical part, the inner parts are displayed by moving the outermost mechanical parts to define the indicated direction. For each mechanically connected part, the indicated direction represents the optimal direction in which the specified mechanical part should be disassembled. Due to the interconnection between mechanical parts and the hierarchical structure based on the topological layer, the outermost mechanical part is removed first. For any mechanical part located in the next layer, an interference vector is constructed, denoted as I. Based on the vector difference between its geometric center and the geometric center of the specified mechanical part, the interference vector represents the influence of the mechanical part on the disassembly of the specified mechanical part. To ensure a smooth disassembly process and eliminate the influence of the disassembled parts, all the interference vectors i corresponding to all mechanical parts under one layer of the specified part are stacked and normalized. This process results in an optimized guidance direction D, which represents the best direction to move the mechanical part to reveal the mechanical structure; By proposing a correction axis, denoted as A, to adjust the optimal guiding direction D, this correction aims to optimize the disassembly animation by resolving the mismatch between the optimal guiding direction D and the actual shape of the mechanical parts. Given that parts usually have complex shapes, there is no universal template available. In order to uniformly define the correction axis A for mechanical parts, mechanical parts are extensively classified. For those parts with obvious axial features, they are classified as having axial features, and for mechanical parts lacking obvious axial features, they are classified as non-axial parts. Based on the classification of the oriented bounding box of the mechanical parts, the minimum bounding box is obtained for the 3D model of the complex mechanical parts, and then the generated cube is classified to facilitate the axial classification of the complex mechanical models. Based on axial classification, for mechanical parts with axial features, the axial direction corresponding to the oriented bounding box is extracted. This extracted axial direction acts as the correction axis A, allowing the mechanical part to move along its axial direction. This correction aligns the movement of the part with the actual assembly and disassembly direction of the mechanical structure. On the other hand, for non-axial parts, the correction axis A is defined as the vector DG of the maximum module. The vector DG is the projection between the projections of the optimal guiding direction D on the relative spatial coordinate axes. When the optimal movement direction is uncertain, the non-axial part is guided to expand along the spatial axis to align the animation with the observer's perspective. Based on the correction axis A, the projection vector DP of the optimal guidance direction DG on A is constructed to represent the correction movement direction of the mechanical component to improve the accuracy and consistency of the disassembly animation and ensure that the movement of the mechanical component is consistent with its actual assembly and disassembly direction or the observer's perspective. After generation, the disassembly direction is attached to the corresponding mechanical component of the mechanical component layer to enhance the disassembly animation in interactive navigation.

7. According to the method for implementing the immersive visualization system for the mechanical structure of a high-speed train bogie as claimed in claim 1, the step 6 of constructing an immersive remote operation and maintenance environment comprises: The user creates a hierarchical navigation through the UI interactive interface and configures relevant parameters. The user drags the mechanical model file to be displayed into the interface, and then attaches the model object in the corresponding script to replace the 3D model. In the UI interface created by the control split animation generation module, the user changes the number of central mechanical parts by controlling the parameters in the mechanical parts set generation script, and customizes the central mechanical parts by dragging the 3D model of the central mechanical parts to the specified position, so that the user can flexibly set the central mechanical parts used to guide the generation of mechanical parts sets, and allows the user to sort the model according to the volume and number of faces, and delete the parts with smaller volume to improve the visualization effect. The user customizes the model simplification process by adjusting the threshold and the number of faces filter; After completing the configuration and construction of the immersive environment, the user interacts with the model at different levels to better view the details of the complex mechanical structure and demonstrate the presentation effects of the visualization configuration and user interaction modules. The user selects different interaction modes to adapt to mechanical structures at different levels. The specific interaction modes include overall mode, collection mode and part mode. The overall mode allows the user to interact with the entire mechanical model to obtain a global view; the collection mode supports the user to explore and operate the mechanical parts collection generated by the method of the immersive visualization system; the part mode provides operations for interacting with a single mechanical part. Based on different levels and interaction modes, the user uses grasping operations to interact with mechanical structures at different levels. This interaction method enables the user to intuitively experience the mechanical structure and enhance the user's sense of immersion. Through grasping interaction operations, the user simulates physical interaction with the actual machine, including assembly and inspection operations.

Images