A three-dimensional modeling simulation system for port road yard construction

Abstract

The present application relates to the technical field of engineering construction simulation, and discloses a three-dimensional modeling simulation system for port road yard construction, comprising initialization, parameter mapping, grid reconstruction, dynamic rendering, pipe network linkage operation and historical state backtracking modules; the system constructs a voxel grid and a pipe network model; a pseudo-density increment is calculated based on a Gaussian attenuation function; spatial splitting or merging is performed according to the gradient vector module length; the displacement amount is calculated by vertical integration and the ground surface grid is updated; the pipe network following displacement is calculated, the pipe line fracture position is determined according to the discrete Gaussian curvature, and the closed end cap grid is generated; the lightweight data frame is extracted to reversely restore the state and parameters. The present application realizes the linkage calculation of foundation and pipe network deformation through the Gaussian attenuation and integration mechanism; the ghost vertex resource is used to generate the fracture end cap to avoid the reconstruction of the full-amount model; the lightweight data frame is used to replace the full-amount storage, thereby reducing the calculation and storage resource occupation of long-period simulation reverse backtracking.

Description

Technical Field

[0001] This invention relates to the field of engineering construction simulation technology, specifically a three-dimensional modeling and simulation system for port area road and storage yard construction. Background Technology

[0002] In the construction of new container road storage yards in coastal port areas, foundation reinforcement techniques such as dynamic compaction are used to improve the bearing capacity of the soil. The impact energy generated by the construction operation will cause three-dimensional non-uniform deformation of the surrounding foundation and exert compression and shearing effects on the existing underground heterogeneous pipeline network. In order to prevent damage to the underground pipeline network during the construction process, the engineers rely on a three-dimensional modeling simulation system to conduct construction simulation and status monitoring. Existing 3D simulation systems often rely on pre-set overall settlement models or boundary collision detection algorithms when dealing with the impact of foundation reinforcement construction. These methods struggle to dynamically calculate changes in soil density and the resulting non-uniform geometric deformation based on the propagation and attenuation of energy at the construction site. Furthermore, when calculating the impact of soil deformation on underground pipe networks, traditional methods disregard material physical properties and lack a linkage mechanism between foundation displacement differences and pipeline material stiffness, failing to accurately reflect the bending and stress state of the pipe network under the influence of surrounding soil deformation. When geological deformation exceeds the pipe material's tolerance limit, the pipe network structure will break. Currently, simulation systems typically re-topologically divide and reconstruct the entire 3D solid model of the pipe network after determining fracture when rendering the damaged state. However, the existing full-model reconstruction graphics processing logic requires clearing and reallocating vertex buffer resources in the graphics processor's memory when generating the damaged section, resulting in significant computational delays and stuttering in the 3D rendering. Foundation construction simulation is characterized by long operation cycles and large data volumes. After irreversible topological changes such as pipeline rupture occur, the simulation system needs to perform state backtracking to identify the specific construction batch that caused the deformation overload. The existing state backtracking mechanism mainly relies on snapshot saving of the full three-dimensional spatial mesh model at fixed time steps. Due to the large data structure of voxel mesh and irregular triangular mesh, continuous high-frequency full data writing will quickly exhaust storage space. When performing backtracking reading, the large data file will also cause data transmission congestion between main memory and external storage devices, making it difficult for the simulation system to achieve long-cycle, high-precision time-series state backtracking under limited hardware resources. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a three-dimensional modeling and simulation system for port area road and storage yard construction. This system solves the problems of existing foundation construction simulation systems, which struggle to accurately calculate the linkage between soil deformation and underground pipeline stress, and suffer from high graphics calculation latency and excessive data storage consumption during pipeline fracture rendering and long-cycle state backtracking.

[0004] To achieve the above objectives, the present invention provides the following technical solution: This invention provides a three-dimensional modeling and simulation system for port area road and storage yard construction, comprising: The initialization module is used to construct sparse octree voxel meshes and irregular triangular meshes, import the 3D vector model of underground heterogeneous pipe networks, and establish parameter stacks and topological event hash tables. The parameter mapping module is used to receive construction operation parameters, calculate the pseudo-compactness increment based on the three-dimensional Gaussian decay function, and update the scalar parameter pseudo-compactness. The mesh reconstruction module is used to compare the Euclidean magnitude of the 3D gradient vector with the preset splitting threshold and preset merging threshold, and perform spatial splitting or spatial merging. The dynamic rendering module is used to calculate the vertical geometric displacement based on the difference between the scalar parameter pseudo-density and the preset pseudo-density benchmark value, and update the grid vertex coordinates of the irregular triangular mesh. The pipeline linkage calculation module is used to calculate the following spatial displacement vector of the pipeline control points using the anisotropic stiffness tensor matrix, and to determine the fracture location of the same pipeline based on the discrete Gaussian curvature of the grid vertices on the pipeline surface, the discrete Gaussian curvature limit threshold, and the topological adjacency relationship. The historical state backtracking module is used to recover the spatial splitting state, spatial merging state, and scalar parameter pseudo-density of the sparse octree voxel mesh based on the historical state backtracking instructions, parameter stack, and topological event hash table.

[0005] The initialization module assigns a preset pseudo-density benchmark value to the voxel nodes of the sparse octree voxel mesh at the initial level, parses the three-dimensional vector model of the underground heterogeneous pipe network with the data structure stored using the boundary representation method, extracts the pipe network control points and the mesh vertices on the pipe surface, and establishes a spatial index mapping binding relationship. Pre-allocate ghosting vertex resources for pipeline surface mesh vertices within the graphics processor's video memory vertex buffer object, and set the initial graphics pipeline rendering state of the ghosting vertex resources to hibernation and hiding; configure anisotropic stiffness tensor matrices for each pipeline entity according to the engineering physical material properties of the 3D vector model.

[0006] The parameter mapping module receives the actual construction operation parameter set in the port area, extracts the coordinates of the construction action center, the energy level of the construction equipment, the construction process type and the effective influence space radius, converts the coordinates of the construction action center into three-dimensional spatial floating point coordinates in the graphics rendering pipeline, and determines the three-dimensional spatial scalar energy level and the influence range attenuation standard deviation. Before each construction batch begins, the pseudo-density increment of the voxel node in the current construction batch is initialized. The computation shader program of the graphics processor is called to calculate the pseudo-density increment of the voxel node within the effective spatial radius in parallel. The in-situ update of the scalar parameter pseudo-density is completed through parallel algebraic addition.

[0007] The mesh reconstruction module uses the central difference method to solve the first-order partial derivatives of the scalar parameter pseudo-density along the horizontal, vertical and vertical coordinate axes, and calculates the Euclidean modulus of the three-dimensional gradient vector. When the Euclidean modulus exceeds the preset splitting threshold, a dynamic splitting operation is performed on the corresponding voxel node to generate a child voxel node, and the scalar parameter pseudo-compactness and the pseudo-compactness increment of the current construction batch are allocated according to the trilinear interpolation algorithm. When the Euclidean modulus of all child voxel nodes generated from the same parent voxel node is lower than the preset merging threshold, spatial merging is performed. The arithmetic mean of the pseudo-density of the scalar parameters and the arithmetic mean of the pseudo-density increment of each child voxel node are assigned to the recovered upper-level voxel node, and the spatial splitting event and spatial merging event are written into the topology event hash table.

[0008] The dynamic rendering module extracts the scalar parameter pseudo density of voxel nodes in the leaf node state, calculates the difference between the scalar parameter pseudo density and the preset pseudo density benchmark value, and divides the spatial vertical column using the horizontal coordinate component and vertical coordinate component of the spatial center coordinate of the voxel node as the matching benchmark. The difference is integrated and accumulated algebraically along the vertical coordinate axis from bottom to top to obtain the columnar cumulative density increment. The columnar cumulative density increment is multiplied by the preset volume scaling factor to obtain the vertical geometric displacement and recorded as the spatial displacement. After spatial splitting or merging, the spatial index mapping binding relationship between the mesh vertices of the irregular triangular mesh and the voxel nodes in the uppermost leaf node state in space is re-established. The mesh vertex coordinates of the irregular triangular mesh are updated according to the spatial index mapping binding relationship, and the terrain geometry deformation rendering screen is output.

[0009] The pipeline linkage calculation module extracts the spatial displacement of voxel nodes including pipeline control points and the spatial displacement of multiple adjacent voxel nodes in the spatial range, and calculates the arithmetic mean of the spatial displacement of multiple adjacent voxel nodes and the spatial displacement difference vector between the spatial displacement of voxel nodes and the arithmetic mean. By processing the spatial displacement difference vector using the anisotropic stiffness tensor matrix, the following spatial displacement vector of the bending of the pipeline control point is obtained. The following spatial displacement vector of the bending is then algebraically added to the arithmetic mean to obtain the following spatial displacement vector of the pipeline control point that includes both translation and bending. Update the spatial center coordinates of the pipeline surface mesh vertices based on the spatial index mapping binding relationship.

[0010] The pipeline network linkage calculation module extracts triangular mesh primitives adjacent to the vertices of the pipeline surface mesh based on topological adjacency relationships, and calculates the discrete Gaussian curvature of the vertices of the pipeline surface mesh based on the principle of discrete differential geometry. When multiple pipeline surface mesh vertices with topological adjacency are connected sequentially to form a closed structure, and the discrete Gaussian curvature of each pipeline surface mesh vertex is greater than the discrete Gaussian curvature limit threshold of the corresponding pipeline entity, or less than the negative of the limit threshold, the multiple pipeline surface mesh vertices are determined as the same pipeline fracture location, and fracture status labeling data is added to the pipeline surface mesh vertices that have not been attached with fracture status labeling data.

[0011] The graphics processor's geometry shader program stops outputting continuous primitives of the pipeline surface mesh containing fracture status marker data, writes the updated spatial center coordinates of the pipeline surface mesh vertices at the same pipeline fracture location into the corresponding ghost vertex resource, and uses the ghost vertex resource to instantiate and generate closed end cap polygon meshes on both sides of the pipeline fracture location. The pipeline linkage calculation module extracts the spatial center coordinates of each closed end cap polygonal grid, combines the pipeline code identifier and the simulation system timestamp to generate a globally unique identifier for the graphic fracture event, asynchronously writes the graphic fracture event information into the graphics processor's video memory buffer, transmits it to the main memory space after the video memory barrier instruction is released, and writes the graphic fracture event information into the topology event hash table using the globally unique identifier as the hash key value.

[0012] After a single simulation frame is rendered, the historical state backtracking module extracts the input feature parameter set, the simulation system timestamp, and the globally unique identifiers of the spatial splitting event, spatial merging event, and graphic breakage event that occurred in the current simulation frame. Based on the last-in-first-out rule, it encapsulates these into a lightweight data frame and pushes it onto the parameter stack. After receiving the historical state backtracking instruction carrying the target backtracking simulation system timestamp, a lightweight data frame with a simulation system timestamp greater than the target backtracking simulation system timestamp and not greater than the current simulation system timestamp is obtained through a loop pop operation. The lightweight data frame corresponding to the target backtracking simulation system timestamp is not popped from the stack.

[0013] The historical state backtracking module queries the topology event hash table based on the globally unique identifiers of the spatial splitting and spatial merging events recorded in the popped lightweight data frames, and processes the lightweight data frames one by one from the current simulation system timestamp back to the target simulation system timestamp. Perform spatial splitting on spatial merging events, and perform spatial merging on spatial splitting events, restoring the parameters and coordinates of the corresponding nodes; Before and after each lightweight data frame is processed, the pseudo-density increment of the voxel nodes in the leaf node state is initialized. The input feature parameter set is parsed, the Gaussian decay function in three-dimensional space is called to recalculate the pseudo-density increment and perform the inverse negation algebra operation, and the pseudo-density of the scalar parameter is restored by algebraic addition. Recalculate the spatial displacement, the grid vertex coordinates of the irregular triangular mesh, and the grid vertex coordinates of the pipeline surface; Based on the globally unique identifier of the graphical fracture event, query the topology event hash table, clear the fracture status marker data of the pipeline surface mesh vertices, stop generating closed end cap polygon meshes, and restore the continuous primitive output of the pipeline surface mesh; For vertices with unremoved markers, write the coordinates to the ghost vertex resource and generate a closed endcap polygon mesh.

[0014] This invention provides a three-dimensional modeling and simulation system for the construction of road and storage yards in port areas. It has the following beneficial effects: 1. This invention calculates the pseudo-compactness increment of voxel nodes using a three-dimensional Gaussian decay function, obtains the vertical geometric displacement by combining integral accumulation along the vertical coordinate axis, and processes the spatial displacement difference vector between adjacent voxel nodes using anisotropic stiffness tensor matrix. This transforms construction operation parameters into spatial deformation data of the foundation soil, enabling quantitative calculation of the spatial displacement of control points when foundation deformation is transmitted to the underground pipeline network.

[0015] 2. This invention determines the pipeline fracture location by calculating the discrete Gaussian curvature of the vertices of the pipeline surface mesh. At the fracture location, the geometry shader program of the graphics processor stops outputting continuous primitives and uses pre-allocated ghost vertex resources to instantiate and generate a closed end cap polygon mesh, thereby reducing the memory resource consumption and rendering computation delay caused by the full 3D model reconstruction when the pipeline structure is damaged.

[0016] 3. This invention encapsulates input feature parameters and topology event identifiers into lightweight data frames and pushes them onto the parameter stack. Upon receiving a backtracking instruction, the data frames are popped from the stack according to the last-in-first-out rule. The spatial merging and spatial splitting operations are performed in reverse, and the pseudo-density of scalar parameters is compensated by inverse negation algebraic operations. This replaces the full data storage mode for three-dimensional solid models, reducing the storage space occupation and data read / write overhead during long-term dimensional construction simulation state backtracking. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the timing interaction of the present invention; Figure 2 This is a schematic diagram of the method steps of the present invention; Figure 3 This is the voxel grid evolution topology diagram of the present invention; Figure 4 This is a topology diagram of the pipeline network fracture reconstruction according to the present invention; Figure 5 This is a schematic diagram illustrating the principle of vertical integration and displacement mapping in this invention. Detailed Implementation

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] See attached document Figure 1 A three-dimensional modeling and simulation system for port area road and storage yard construction includes: The initialization module is used to construct an initial level of sparse octree voxel mesh and an irregular triangular mesh as a surface visualization model in three-dimensional space. It parses the three-dimensional vector model of the underground heterogeneous pipe network using boundary representation to store data structure, establishes the spatial index mapping binding relationship between control points and pipe surface mesh vertices, allocates corresponding ghosting vertex resources for pipe surface mesh vertices in the graphics processor's video memory buffer, and establishes parameter stacks and topology event hash tables in main memory.

[0020] The parameter mapping module receives the coordinates of the construction action center, the construction process type, the energy level of the construction equipment, and the effective influence space radius. Before the start of each construction batch, it configures the pseudo-density increment of the current construction batch of the voxel nodes in the leaf node state of the sparse octree voxel mesh to zero, calls the computation shader program of the graphics processor, calculates the pseudo-density increment of the current construction batch within the corresponding action range based on the three-dimensional spatial Gaussian decay function, and updates the pseudo-density scalar field of the sparse octree voxel mesh in parallel.

[0021] The mesh reconstruction module is used to calculate the three-dimensional spatial gradient of the pseudo-density scalar field, compare it with the preset splitting threshold and the preset merging threshold, perform spatial splitting and trilinear interpolation operations on voxel nodes whose three-dimensional spatial gradient exceeds the preset splitting threshold, and perform spatial merging on eight adjacent child voxel nodes whose three-dimensional spatial gradient is lower than the preset merging threshold and have the same parent voxel node identity.

[0022] The dynamic rendering module is used to accumulate the difference between the pseudo-density of each spatial vertical column scalar parameter and the preset pseudo-density baseline value along the vertical coordinate axis, introduce a preset volume scaling factor to calculate the vertical geometric displacement of each voxel node and the mesh vertex of the irregular triangular mesh, update the mesh vertex coordinates of the irregular triangular mesh and output the terrain geometric deformation rendering screen.

[0023] The pipeline linkage calculation module is used to transfer the spatial displacement of voxel nodes to the pipeline control points of the three-dimensional vector model of the underground heterogeneous pipeline network through the anisotropic stiffness tensor matrix, calculate the discrete Gaussian curvature of the grid vertices on the pipeline surface, and determine the fracture location of the same pipeline based on the topological adjacency relationship. When the discrete Gaussian curvature of multiple pipeline surface mesh vertices at the same pipeline fracture location exceeds the discrete Gaussian curvature limit threshold of the corresponding pipeline entity, fracture state marker data is added. The graphics processor's geometry shader program is instructed to stop outputting continuous primitives of the pipeline surface mesh containing fracture state marker data. The updated spatial center coordinates of the pipeline surface mesh vertices are written into the corresponding ghosted vertex resources, and the ghosted vertex resources are used to generate two closed end cap polygon meshes corresponding to each pipeline fracture location.

[0024] The historical state backtracking module is used to receive historical state backtracking instructions, extract the input feature parameter set according to the parameter stack, restore the spatial splitting state and spatial merging state of the sparse octree voxel grid according to the topological event hash table, call the three-dimensional spatial Gaussian decay function used by the parameter mapping module to perform inverse inversion operation, restore the scalar parameter pseudo-compactness according to the pseudo-compactness increment of the current construction batch, and configure the corresponding pseudo-compactness increment of the current construction batch to zero. The historical state backtracking module recalculates the spatial displacement of each voxel node, the grid vertex coordinates of the irregular triangular mesh, and the grid vertex coordinates of the pipeline surface mesh based on the restored scalar parameter pseudo-density. It resets the fracture state marker data of the corresponding pipeline surface mesh vertex according to the graphics fracture event recorded in the popped lightweight data frame, retains the unreset fracture state marker data, and instructs the geometry shader program of the graphics processor to generate the corresponding two closed end cap polygon meshes or restore the continuous primitive output of the pipeline surface mesh.

[0025] See attached document Figure 2 A three-dimensional modeling and simulation method for port area road and storage yard construction includes the following steps: S1. Construct a sparse octree voxel mesh to carry scalar parameters and an irregular triangular mesh as a surface visualization model. Set the initial pseudo-density value of each voxel node. Import the 3D vector model of the underground heterogeneous pipe network. In the vertex buffer object, pre-allocate the hidden ghost vertex resources for the vertices of the pipe surface mesh. Complete the attribute mapping of the anisotropic stiffness tensor matrix. In the main memory space, establish a parameter stack and a topological event hash table for recording spatial splitting, spatial merging and graphic fracture events.

[0026] S2 sets a three-dimensional Gaussian decay function based on construction operation data. The calculation shader program processes the voxel nodes in the leaf node state of the sparse octree voxel mesh within the spatial radius in parallel, and updates the scalar parameter pseudo-density within the local spatial range.

[0027] S3 calculates the three-dimensional gradient vector and the Euclidean magnitude of the three-dimensional gradient vector of the updated pseudo-density scalar field. Based on the comparison results of the Euclidean magnitude of the three-dimensional gradient vector with the preset splitting threshold and preset merging threshold, the spatial resolution of the sparse octree voxel mesh is adjusted. Child voxel nodes are generated in the region where the pseudo-density of the scalar parameter changes abruptly, and spatial merging is performed in the region where the pseudo-density of the scalar parameter changes gradually.

[0028] S4. Perform columnar integration along the vertical coordinate axis on the difference between the scalar parameter pseudo-density and the preset pseudo-density benchmark value. Convert the columnar cumulative density increment into the vertical geometric displacement of each voxel node and the mesh vertex of the irregular triangular mesh, and update the mesh vertex coordinates of the irregular triangular mesh.

[0029] S5, using the anisotropic stiffness tensor matrix to calculate the following spatial displacement vector of the control points of the underground heterogeneous pipe network in the three-dimensional vector model, calculates the discrete Gaussian curvature of the grid vertices on the pipe surface through the vertex shader program of the graphics processor, and determines the fracture location of the same pipe based on the topological adjacency relationship. When the discrete Gaussian curvature of multiple pipeline surface mesh vertices at the same pipeline fracture location exceeds the discrete Gaussian curvature limit threshold of the corresponding pipeline entity, fracture status marker data is added, triggering the graphics processor's geometry shader program to stop outputting continuous primitives of the pipeline surface mesh containing fracture status marker data, and generating two closed end cap polygon meshes corresponding to each pipeline fracture location. Graphics breakage event information is asynchronously written to the graphics processor's memory buffer and then transferred to the topology event hash table in main memory after the memory barrier instruction is released.

[0030] S6 records the input feature parameter set, simulation system timestamp, and globally unique identifiers of spatial splitting, spatial merging, and graphic fracture events for a single operation in the parameter stack; after receiving the historical state backtracking instruction, it obtains the input feature parameter set through a stack popping operation, restores the spatial splitting and spatial merging states of the sparse octree voxel mesh according to the topological event hash table, performs the same three-dimensional spatial Gaussian decay function inverse operation as the parameter mapping module, restores the scalar parameter pseudo-compactness according to the pseudo-compactness increment of the current construction batch, and configures the pseudo-compactness increment of the current construction batch to zero; Based on the restored scalar parameter pseudo-density, the spatial displacement of each voxel node, the grid vertex coordinates of the irregular triangular mesh, and the grid vertex coordinates of the pipeline surface mesh are recalculated. The corresponding fracture state marker data is reset according to the graphic fracture event recorded in the popped lightweight data frame. The unreset fracture state marker data is retained, and the graphics processor is instructed to generate the corresponding two closed end cap polygon meshes or re-output the pipeline surface mesh continuous primitives of the three-dimensional vector model of the underground heterogeneous pipe network.

[0031] See attached document Figure 3 The S1 provided by the present invention may include the following sub-steps in a specific implementation: S11, In the initial stage of system operation, the initialization module constructs an initial level of sparse octree voxel mesh based on the three-dimensional spatial boundary bounding box of the port construction operation area, and reserves the physical space of the graphics processor's video memory for generating child voxel nodes in dynamic splitting operation. The initialization module sets a preset pseudo-density benchmark value based on the initial foundation void ratio provided by the engineering geological survey report, and defines multiple voxel nodes contained in the sparse octree voxel mesh as a set of voxel nodes. Each voxel node in the set of voxel nodes has a spatial center coordinate consisting of horizontal coordinate components, vertical coordinate components, and vertical coordinate components. The initialization module assigns a scalar parameter pseudo-compaction degree to the voxel nodes in the voxel node set to characterize the compaction state of the engineering foundation soil. During the initialization phase, the voxel nodes of the sparse octree voxel mesh at the initial level are assigned the same preset pseudo-compaction degree benchmark value. An increase in the value of the scalar parameter pseudo-compaction degree indicates an increase in the compaction degree of the engineering foundation soil. When the initial foundation void ratio increases, the preset pseudo-compaction degree benchmark value decreases. The formula is expressed as: ; In the formula, The initial level sparse octree voxel mesh is in the leaf voxel node state. The initial pseudo-density of individual element nodes, This is the preset global environment background density constant.

[0032] After the initialization module completes the construction of the sparse octree voxel mesh, it constructs an irregular triangular mesh at the top spatial boundary of the sparse octree voxel mesh as a surface visualization model, and records the mesh vertex coordinates of the irregular triangular mesh during the initialization phase. The initialization module establishes a spatial index mapping binding relationship between the mesh vertices of the irregular triangular mesh and the voxel nodes of the uppermost sparse octree voxel mesh in space. After the sparse octree voxel mesh completes spatial splitting or merging, the initialization module re-establishes the spatial index mapping binding relationship based on the horizontal and vertical coordinate components of the grid vertices of the irregular triangular mesh. The process of constructing the irregular triangular mesh based on the spatial discrete coordinate points adopts the Delaunay triangulation algorithm.

[0033] S12, the initialization module imports the three-dimensional vector model of the underground heterogeneous pipe network using the boundary representation method to store the data structure. The boundary representation method storage data structure includes multiple control points that control the spatial geometry of the three-dimensional vector model of the underground heterogeneous pipe network, as well as the vertices of the pipeline surface mesh of the three-dimensional vector model of the underground heterogeneous pipe network. The initialization module parses the boundary representation method storage data structure, extracts multiple control points and pipeline surface mesh vertices, and records the spatial center coordinates of the pipeline surface mesh vertices during the initialization phase. Multiple control points together constitute a control point set. The control points in the control point set serve as the pipeline control points of the three-dimensional vector model of the underground heterogeneous pipeline network. The initialization module calculates the spatial center coordinates of each pipeline surface grid vertex and the spatial distance between multiple pipeline control points, and establishes a spatial index mapping binding relationship between each pipeline surface grid vertex and the pipeline control point with the smallest spatial distance. The initialization module allocates physical memory space for pipeline control points and pipeline surface mesh vertices within the graphics processor's memory vertex buffer object, and records the corresponding spatial index for each pipeline surface mesh vertex. The spatial index is used to determine the corresponding pipeline surface mesh vertex within the graphics processor's memory vertex buffer object. At the same time, the initialization module writes the topological adjacency relationship in the boundary representation storage data structure into the graphics processor's memory buffer. The initialization module reserves ghosting vertex resources in parallel for pipeline surface mesh vertices within the graphics processor's video memory vertex buffer object. The initial spatial center coordinates stored in each ghosting vertex resource coincide with the spatial center coordinates of the corresponding pipeline surface mesh vertex. Each pipeline surface mesh vertex corresponds to one ghosting vertex resource, and multiple pipeline surface mesh vertices at the same pipeline break point correspond to multiple ghosting vertex resources. The initial graphics pipeline rendering state of the ghosting vertex resources is set to hibernation and hidden, and they do not participate in the geometric primitive assembly and pixel output calculation of the current simulation frame.

[0034] S13, the initialization module configures anisotropic stiffness tensor matrices for each pipeline entity based on the engineering physical material properties of the three-dimensional vector model of the underground heterogeneous pipeline network. The anisotropic stiffness tensor matrix is ​​generated by mapping the pipeline elastic modulus and bending stiffness values ​​according to a preset ratio relationship, with a value range of zero to one. The anisotropic stiffness tensor matrices of rigid material pipelines and flexible material pipelines correspond to different eigenvalue distributions. When the elastic modulus and bending stiffness of the pipeline increase, the eigenvalues ​​of the anisotropic stiffness tensor matrix decrease in the direction of a specific coordinate axis. Consequently, the bending spatial displacement vector of the control points of the underground heterogeneous pipeline network in the three-dimensional vector model decreases. Conversely, when the elastic modulus and bending stiffness of the pipeline decrease, the eigenvalues ​​of the anisotropic stiffness tensor matrix increase in the direction of a specific coordinate axis. Consequently, the bending spatial displacement vector of the control points of the underground heterogeneous pipeline network in the three-dimensional vector model increases. The anisotropic stiffness tensor matrix is ​​used to restrict the relative degrees of freedom of the pipeline control points in the direction of a specific coordinate axis. The initialization module updates the spatial center coordinates of the grid vertices on the surface of adjacent pipelines according to the fracture elongation parameters calibrated by the pipe material corresponding to each pipeline entity, and calculates the spatial spacing in the initialization stage and the spatial spacing after coordinate update. When the spatial spacing after coordinate update is greater than the spatial spacing in the initialization stage, the initialization module calculates the difference between the two spatial spacings and calculates the proportional relationship between the difference and the spatial spacing in the initialization stage. The initialization module calculates the discrete Gaussian curvature corresponding to different proportional relationships and establishes the correspondence between the fracture elongation parameter and the discrete Gaussian curvature. When the proportional relationship reaches the fracture elongation parameter, the discrete Gaussian curvature value greater than zero remains unchanged. When the discrete Gaussian curvature is less than zero, the discrete Gaussian curvature is subtracted from zero to obtain a value greater than zero. The value greater than zero is set as the discrete Gaussian curvature limit threshold of the corresponding pipeline entity. The discrete Gaussian curvature limit threshold of the corresponding pipeline entity is used as the trigger reference value for the shader program of the graphics processor to determine whether geometric twisting fracture has occurred on the pipeline surface.

[0035] S14, the initialization module allocates a data recording area in the main memory space of the 3D modeling simulation system for the construction of the port area road yard, and establishes a parameter stack that pushes the input feature parameter set of a single construction operation into the simulation system according to the timestamp order. The parameter stack adopts the last-in-first-out data structure specification. At the same time, the initialization module establishes a topological event hash table that adopts the lock-free query structure specification, and uses the globally unique identifiers of spatial splitting events, spatial merging events and graphic break events as index keys. The spatial splitting event uses the type of the spatial splitting event, the simulation system timestamp, the identity of the parent voxel node, the spatial center coordinates of the parent voxel node, the scalar parameter pseudo-compactness, and the pseudo-compactness increment of the current construction batch as storage values. The spatial merging event uses the type of spatial merging event, the simulation system timestamp, the identity of the parent voxel node, the spatial center coordinates of the eight child voxel nodes, the scalar parameter pseudo-compactness, and the pseudo-compactness increment of the current construction batch as storage values. The graphical fracture event uses the following stored values: the type of graphical fracture event, the simulation system timestamp, the pipeline code identifier, the spatial center coordinates of the two closed end cap polygon meshes corresponding to each pipeline fracture location, the spatial center coordinates of multiple pipeline surface mesh vertices at the same pipeline fracture location, the spatial index corresponding to each pipeline surface mesh vertex, and the fracture state marker data.

[0036] In specific implementations, S2 provided by the present invention may include the following sub-steps: S21, the parameter mapping module receives the actual construction operation parameter set of the port area, and extracts the coordinates of the construction action center, the energy level of the construction equipment, the construction process type and the effective influence space radius from the actual construction operation parameter set of the port area; the parameter mapping module constructs a coordinate system transformation matrix based on the translation and scaling ratio between the global positioning system reference point of the construction site and the origin of the virtual three-dimensional space, and converts the coordinates of the construction action center into three-dimensional space floating point coordinates in the graphics rendering pipeline through the coordinate system transformation matrix; The parameter mapping module generates an equipment efficiency lookup table based on historical construction site equipment rated power and measured soil influence depth data. It then uses the equipment efficiency lookup table to convert the construction equipment energy level and construction process type into the three-dimensional spatial scalar energy level used by the three-dimensional spatial Gaussian decay function. The parameter mapping module determines the influence range attenuation standard deviation based on the positive proportional mapping relationship between the effective influence space radius and the influence range attenuation standard deviation.

[0037] S22, Before each construction batch begins, the parameter mapping module configures the pseudo-density increment of the voxel nodes in the leaf node state of the sparse octree voxel mesh to zero for the current construction batch. The parameter mapping module constructs a spatial bounding sphere with the three-dimensional space floating point coordinates as the center and the effective influence space radius as the radius, extracts the voxel nodes in the leaf node state of the sparse octree voxel mesh inside the spatial bounding sphere, and calls the graphics processor's computation shader program. The parameter mapping module divides the thread group of the computation shader program according to the number of extracted voxel nodes, and assigns an independent computation thread to each extracted voxel node. The three-dimensional Gaussian decay function characterizes the nonlinear decreasing propagation relationship of construction energy from the coordinates of the construction action center to the surrounding area. Based on the three-dimensional Gaussian decay function, the independent computation thread of the computation shader program calculates in parallel the pseudo-compactness increment of each voxel node inside the spatially enclosing sphere for the current construction batch. The formula is expressed as: ; In the formula, The initial level sparse octree voxel mesh is in the leaf voxel node state. The pseudo-compactness increment of the current construction batch for individual element nodes; The maximum central influence energy amplitude of the sedimentation-induced source For the natural constant An exponential function with base 0; The initial level sparse octree voxel mesh is in the leaf voxel node state. The horizontal coordinate component of the spatial center coordinates of an individual element node, which consists of the horizontal coordinate component, the vertical coordinate component, and the vertical coordinate component. The initial level sparse octree voxel mesh is in the leaf voxel node state. The longitudinal coordinate component of the spatial center coordinates of an individual element node, which consists of the horizontal coordinate component, the vertical coordinate component, and the vertical coordinate component. The initial level sparse octree voxel mesh is in the leaf voxel node state. The vertical coordinate component of the spatial center coordinates of an individual element node, which consists of the horizontal coordinate component, the vertical coordinate component, and the vertical coordinate component. The lateral coordinate components of the spatial center coordinates of the settlement induction source; The longitudinal coordinate component represents the spatial center coordinates of the settlement induction source. The vertical coordinate components of the spatial center coordinates of the settlement induction source; The standard deviation of the attenuation range of the influence of the sedimentation-induced source.

[0038] S23, after the graphics processor's computation shader program completes the parallel calculation of the pseudo-density increment of the current construction batch, the parameter mapping module triggers the memory barrier instruction; the memory barrier instruction synchronizes the execution progress of each independent computing thread and intercepts the memory write operations that have not yet been completed. After the memory barrier instruction is released, the parameter mapping module adds the pseudo-compactness increment of the current construction batch of each voxel node to the original scalar parameter pseudo-compactness of the corresponding voxel node, and completes the in-situ update of the scalar parameter pseudo-compactness within the local space range through parallel algebraic addition operation. The formula is expressed as: ; In the formula, For the updated number The scalar parameter pseudo-density of the voxel nodes of a sparse octree voxel mesh of an initial level. For the previous version The scalar parameter pseudo-density of the voxel nodes of a sparse octree voxel mesh of an initial level. For the first The pseudo-density increment of the voxel nodes of the sparse octree voxel mesh at the initial level for the current construction batch.

[0039] In specific implementations, S3 provided by the present invention may include the following sub-steps: S31, the mesh reconstruction module traverses the voxel nodes in the leaf node state of the sparse octree voxel mesh and obtains the scalar parameter pseudo-density after in-situ update. The three-dimensional gradient vector reflects the spatial rate of change of the scalar parameter pseudo-density in each orthogonal direction, and the Euclidean magnitude of the three-dimensional gradient vector characterizes the intensity of change of the local scalar parameter pseudo-density. The mesh reconstruction module employs the central difference method, using the spatial distance between adjacent voxel nodes as the difference step size. It calculates the first-order partial derivatives of the scalar parameter pseudo-density along the horizontal, vertical, and triangular coordinate axes, and constructs a three-dimensional gradient vector. When adjacent voxel nodes are at different levels, the mesh reconstruction module calculates the scalar parameter pseudo-density at the corresponding spatial position using a trilinear interpolation algorithm, and uses the actual spatial distance between the corresponding spatial positions as the difference step size. Voxel nodes at the boundary of the three-dimensional spatial bounding box calculate the first-order partial derivatives of the scalar parameter pseudo-density of adjacent voxel nodes inside the three-dimensional spatial bounding box. The mesh reconstruction module calculates the Euclidean modulus of the three-dimensional gradient vector and uses the calculation result as a quantitative indicator of the degree of change in the spatial distribution of the pseudo-density of the local scalar parameter. The formula is expressed as: ; In the formula, For the first The Euclidean magnitude of the 3D gradient vector of the voxel nodes of the sparse octree voxel mesh of an initial level. For the first The first-order partial derivative of the pseudo-density of the scalar parameter of the voxel node of the initial level sparse octree voxel mesh in the direction of the horizontal coordinate axis. For the first The first-order partial derivative of the pseudo-density of the scalar parameter of the voxel node of the initial level sparse octree voxel mesh in the direction of the longitudinal coordinate axis. For the first The first-order partial derivative of the pseudo-density scalar parameter of the voxel node in the vertical coordinate axis direction of the initial level sparse octree voxel mesh.

[0040] S32, the mesh reconstruction module calculates the preset splitting threshold based on the maximum allowable uneven settlement deformation slope and the spatial resolution size of the current sparse octree voxel mesh through a proportional mapping relationship, and extracts voxel nodes whose Euclidean modulus of the three-dimensional gradient vector exceeds the preset splitting threshold. Before performing the dynamic splitting operation, the mesh reconstruction module generates a globally unique identifier for the spatial splitting event based on the type of the spatial splitting event, the simulation system timestamp, and the identity of the parent voxel node. The module then writes the globally unique identifier of the spatial splitting event, the type of the spatial splitting event, the simulation system timestamp, the identity of the parent voxel node, the spatial center coordinates of the parent voxel node, the scalar parameter pseudo-compactness, and the pseudo-compactness increment of the current construction batch into the topology event hash table. The mesh reconstruction module performs a dynamic splitting operation on the extracted voxel nodes to generate eight child voxel nodes with smaller spatial dimensions, and records the identity of the corresponding parent voxel node for each of the eight child voxel nodes. Based on the trilinear interpolation algorithm, the mesh reconstruction module calculates and assigns the scalar parameter pseudo-density and the pseudo-density increment of the current construction batch to each of the eight child voxel nodes, combining the pseudo-density of the parent voxel node and the adjacent voxel nodes and the pseudo-density increment of the current construction batch. The mesh reconstruction module calculates the arithmetic mean of the pseudo-density of the scalar parameters of the eight child voxel nodes and the arithmetic mean of the pseudo-density increment of the current construction batch, respectively. Then, it calculates the difference between the pseudo-density of the scalar parameters of the parent voxel node and its corresponding arithmetic mean, as well as the difference between the pseudo-density increment of the current construction batch of the parent voxel node and its corresponding arithmetic mean. The mesh reconstruction module performs algebraic addition operations on the corresponding differences with the pseudo-density of the scalar parameters of the eight child voxel nodes and the pseudo-density increment of the current construction batch, respectively, so that the arithmetic mean of the pseudo-density of the scalar parameters of the eight child voxel nodes is equal to the pseudo-density of the scalar parameters of the parent voxel node, and the arithmetic mean of the pseudo-density increment of the current construction batch of the eight child voxel nodes is equal to the pseudo-density increment of the current construction batch of the parent voxel node.

[0041] S33, the mesh reconstruction module sets a preset merging threshold based on the allowable error range under the natural consolidation and stability state of the soil, and makes the preset merging threshold less than the preset splitting threshold. The mesh reconstruction module periodically scans eight adjacent child voxel nodes with the same parent voxel node identity. When the Euclidean modulus of the three-dimensional gradient vectors of the eight adjacent child voxel nodes is lower than the preset merging threshold, the mesh reconstruction module performs spatial merging. The mesh reconstruction module calculates the arithmetic mean of the pseudo-compaction of the scalar parameters of eight adjacent sub-voxel nodes and the arithmetic mean of the pseudo-compaction increment of the current construction batch, and assigns the two arithmetic means to the upper-level voxel nodes restored after spatial merging. Before releasing the physical memory space occupied by the eight child voxel nodes, the mesh reconstruction module generates a globally unique identifier for the spatial merging event based on the type of the spatial merging event, the simulation system timestamp, and the identity of the parent voxel node. The module then writes the globally unique identifier of the spatial merging event, the type of the spatial merging event, the simulation system timestamp, the identity of the parent voxel node, the spatial center coordinates of the eight child voxel nodes, the scalar parameter pseudo-compactness, and the pseudo-compactness increment of the current construction batch into the topology event hash table. After completing the data writing, the mesh reconstruction module releases the physical memory space occupied by the eight child voxel nodes.

[0042] See attached document Figure 4 In specific implementations, S4 provided by the present invention may include the following sub-steps: S41, the dynamic rendering module traverses the voxel nodes in the leaf node state of the sparse octree voxel mesh, extracts the scalar parameter pseudo density of each voxel node, and calculates the difference between the scalar parameter pseudo density and the preset pseudo density benchmark value. The dynamic rendering module uses the horizontal and vertical coordinate components of the spatial center coordinates of each voxel node as a matching reference to divide the spatial vertical column, and extracts the voxel nodes whose spatial range in the horizontal and vertical coordinate directions contains the corresponding horizontal and vertical coordinate components. When calculating the cumulative density increment of each voxel node in the vertical column of space, the dynamic rendering module uses the bottom vertical coordinate of the three-dimensional spatial bounding box as the lower limit of integration and the vertical coordinate component of the corresponding voxel node as the upper limit of integration. It performs an algebraic operation of the difference between the scalar parameter pseudo density and the preset pseudo density benchmark value from bottom to top along the direction parallel to the vertical coordinate axis. When calculating the cumulative density increment of the columnar column at the top of the spatial vertical column, the dynamic rendering module uses the bottom vertical coordinate of the three-dimensional spatial bounding box as the lower limit of integration and the top vertical coordinate of the three-dimensional spatial bounding box as the upper limit of integration. It performs an integral accumulation algebra operation on the difference between the scalar parameter pseudo density and the preset pseudo density benchmark value from bottom to top along a direction parallel to the vertical coordinate axis. Each voxel node records the cumulative density increment of the columnar column calculated for the corresponding spatial vertical column. The formula is expressed as: ; In the formula, For the horizontal coordinate components With vertical coordinate components The columnar cumulative density increment of the spatial vertical columns divided to match the baseline; The spatial center coordinates, composed of horizontal, vertical, and triangular coordinate components, have corresponding horizontal coordinate components. and vertical coordinate components With vertical coordinate components The pseudo-density increment of the initial level sparse octree voxel mesh in the leaf node state of the voxel node. For having corresponding horizontal coordinate components and vertical coordinate components With vertical coordinate components The geometric height of the initial level sparse octree voxel mesh in the leaf node state of the voxel node in the vertical coordinate axis direction; The vertical coordinate components of the lowest-level voxel node in the spatial vertical column; The vertical coordinate component of the top voxel node in the spatial vertical column.

[0043] S42, the dynamic rendering module sets a preset volume scaling factor with a value range of zero to one based on the soil compression coefficient provided in the engineering geological survey report and the Poisson's ratio parameter of the measured physical soil. The preset volume scaling factor is used to convert the columnar cumulative density increment into the spatial deformation size of the grid vertices of the irregular triangular mesh. The dynamic rendering module multiplies the cumulative density increment of the columnar column with the preset volume scaling factor to calculate the vertical geometric displacement of each voxel node in the vertical column of space and the vertical geometric displacement of the top of the vertical column of space. The direction of the vertical geometric displacement is from the top vertical coordinate of the three-dimensional space bounding box to the bottom vertical coordinate. The dynamic rendering module records the vertical geometric displacement of each voxel node as the spatial displacement of each voxel node. The formula is expressed as: ; In the formula, For the horizontal coordinate components and vertical coordinate components The vertical geometric displacement of the top of the corresponding spatial vertical column; Preset volume scaling factor; This represents the cumulative density increment of the corresponding spatial vertical column.

[0044] S43, After the mesh reconstruction module completes spatial splitting or spatial merging, the dynamic rendering module re-establishes the spatial index mapping binding relationship between the mesh vertices of the irregular triangular mesh and the voxel nodes in the leaf node state of the uppermost sparse octree voxel mesh in space, based on the horizontal and vertical coordinate components of the mesh vertices of the irregular triangular mesh. The dynamic rendering module applies the calculated vertical geometric displacement to the vertical coordinate components of the corresponding mesh vertices of the irregular triangular mesh recorded in the initialization phase, based on the spatial index mapping binding relationship, and updates the coordinates of the corresponding mesh vertices. The dynamic rendering module recalculates the normal vectors of each mesh face and the smoothed normal vectors of each mesh vertex based on the updated coordinates of the irregular triangular mesh using a mesh normal recalculation algorithm. The updated normal vectors are then input into the fragment shader program of the graphics processor to perform lighting calculations and output the continuously deformed image of the irregular triangular mesh.

[0045] In a specific implementation, S5 provided by the present invention may include the following sub-steps: S51, the pipeline linkage calculation module extracts the pipeline control points of the three-dimensional vector model of the underground heterogeneous pipeline network in the virtual three-dimensional spatial coordinate system. The anisotropic stiffness tensor matrix represents the resistance to deformation in the direction of force. The anisotropic stiffness tensor matrix is ​​constructed based on the material mechanical properties of the three-dimensional vector model of the underground heterogeneous pipeline network. The pipeline linkage calculation module extracts the spatial displacement of voxel nodes including pipeline control points and the spatial displacement of multiple adjacent voxel nodes within the spatial range, and calculates the arithmetic mean of the spatial displacement of multiple adjacent voxel nodes. When adjacent voxel nodes are at different levels, the pipeline linkage calculation module calculates the spatial displacement of the corresponding spatial position based on the trilinear interpolation algorithm. The pipeline linkage calculation module calculates the difference between the spatial displacement of the voxel node of the pipeline control point and the arithmetic mean of the spatial displacement of multiple adjacent voxel nodes in the spatial range, and uses the calculation result as the spatial displacement difference vector. The pipeline linkage calculation module uses the anisotropic stiffness tensor matrix to process the spatial displacement difference vector and calculates the following spatial displacement vector of the bending of the pipeline control point. The pipeline linkage calculation module performs an algebraic addition operation on the following spatial displacement vector of the bending and the arithmetic mean of the spatial displacement of multiple adjacent voxel nodes to obtain the following spatial displacement vector of the pipeline control point, which includes translation and bending. The formula is expressed as: ; In the formula, The spatial displacement vector of the control points of the underground heterogeneous pipe network in a three-dimensional vector model; The anisotropic stiffness tensor matrix is ​​constructed based on the material mechanical properties of a three-dimensional vector model of an underground heterogeneous pipe network. This is the spatial displacement difference vector generated by the spatial displacement of the voxel nodes in the leaf node state of the initial level sparse octree voxel mesh of the pipeline control points of the adjacent three-dimensional vector model surrounding the underground heterogeneous pipeline network.

[0046] S52, the pipeline network linkage calculation module, based on the spatial index mapping binding relationship between the pipeline network control points and the pipeline surface mesh vertices established in S12, performs algebraic addition operations on the spatial center coordinates of the pipeline surface mesh vertices in the initialization phase and the following spatial displacement vectors of the corresponding pipeline network control points, updates the spatial center coordinates of the pipeline surface mesh vertices of the three-dimensional vector model of the underground heterogeneous pipeline network, and inputs the updated pipeline surface mesh vertices into the vertex shader program of the graphics processor. The vertex shader program of the graphics processor extracts the triangular mesh primitives adjacent to the vertices of the pipeline surface mesh based on the topological adjacency relationship in the graphics processor's memory buffer, and calculates the discrete Gaussian curvature of the pipeline surface mesh vertices after coordinate update based on the principle of discrete differential geometry; the vertex shader program writes the discrete Gaussian curvature into the graphics processor's memory buffer; after the discrete Gaussian curvature is written, the pipeline linkage calculation module triggers the memory barrier instruction. After the memory barrier instruction is released, the graphics processor's compute shader program reads the discrete Gaussian curvature of multiple pipeline surface mesh vertices based on topological adjacency. When the discrete Gaussian curvature of each pipeline surface mesh vertex with topological adjacency is greater than the discrete Gaussian curvature limit threshold of the corresponding pipeline entity, or less than zero minus the discrete Gaussian curvature limit threshold of the corresponding pipeline entity, and the multiple pipeline surface mesh vertices are connected sequentially to form a closed structure based on topological adjacency, the compute shader program determines the multiple pipeline surface mesh vertices as the same pipeline breakage location. The graphics processor's compute shader program reads the fracture status marker data of multiple pipeline surface mesh vertices at the same pipeline fracture location. It adds fracture status marker data only to pipeline surface mesh vertices that do not have fracture status marker data attached, and writes the fracture status marker data into the graphics processor's video memory buffer. After the fracture status marker data is written, the pipeline linkage calculation module triggers the video memory barrier instruction again. The formula is expressed as: ; In the formula, The pipeline surface mesh of the three-dimensional vector model of the underground heterogeneous pipe network. Discrete Gaussian curvature of each vertex; For the pipeline surface mesh of the three-dimensional vector model of the underground heterogeneous pipe network The total number of triangular mesh primitives with adjacent vertices; For the pipeline surface mesh of the three-dimensional vector model of the underground heterogeneous pipe network The vertex adjacent to the first The interior angles of a triangular mesh primitive; For the pipeline surface mesh of the three-dimensional vector model surrounding the underground heterogeneous pipe network, The local measure of surface area at each vertex.

[0047] S53, the pipeline network linkage calculation module passes all pipeline surface mesh vertices with additional fracture state marker data to the geometry shader program of the graphics processor. The geometry shader program of the graphics processor intercepts the continuous primitives of the pipeline surface mesh containing fracture state marker data in each simulation frame, stops outputting the corresponding continuous primitives of the pipeline surface mesh, so that the corresponding continuous primitives of the pipeline surface mesh do not participate in the subsequent pixel output calculation. At the pipeline breakage location, the graphics processor's geometry shader program, based on topological adjacency, writes the updated spatial center coordinates of multiple pipeline surface mesh vertices that are sequentially connected to form a closed structure into the corresponding ghosted vertex resources, and extracts the corresponding ghosted vertex resources. The graphics processor's geometry shader program instantiates and generates two closed end cap polygonal meshes corresponding to the pipeline surface mesh continuous primitives that stop output on both sides of the pipeline breakage location, respectively, according to the direction and reverse direction of sequentially connecting multiple ghosted vertex resources based on topological adjacency. For multiple pipeline surface mesh vertices whose fracture state marking data is updated from unattached to attached state in the current simulation frame, the pipeline linkage calculation module extracts the spatial center coordinates of the two closed end cap polygon meshes corresponding to each pipeline fracture location, and generates a globally unique identifier for the graphical fracture event by combining the pipeline coding identifier and the simulation system timestamp. The pipeline linkage calculation module asynchronously writes the globally unique identifier of the graphic fracture event, the type of the graphic fracture event, the simulation system timestamp, the pipeline code identifier, the spatial center coordinates of the two closed end cap polygonal meshes, the spatial center coordinates of the vertices of multiple pipeline surface meshes at the same pipeline fracture location, the spatial index corresponding to each pipeline surface mesh vertex, and the fracture status marker data into the graphics processor's video memory buffer. After the graphics fracture event information is written, the pipeline linkage calculation module triggers the memory barrier instruction; after the memory barrier instruction is released, the pipeline linkage calculation module transmits the graphics fracture event information to the main memory space and writes the globally unique identifier of the graphics fracture event as a hash key value into the topology event hash table.

[0048] In specific implementations, S6 provided by the present invention may include the following sub-steps: S61, after a single simulation frame is rendered, the historical state backtracking module extracts a set of input feature parameters including the coordinates of the construction action center, the energy level of the construction equipment, the type of construction technology, the effective influence space radius and the standard deviation of the influence range attenuation. At the same time, it extracts the corresponding simulation system timestamp and the globally unique identifiers of the spatial splitting, spatial merging and graphic breakage events that occurred in the current simulation frame. The historical state backtracking module encapsulates the input feature parameter set, the corresponding simulation system timestamp, and the globally unique identifiers of spatial splitting, spatial merging, and graphical fracture events into a lightweight data frame according to the last-in-first-out rule, and pushes the lightweight data frame into the parameter stack established by S14.

[0049] S62, after receiving the historical state backtracking instruction carrying the target backtracking simulation system timestamp, the historical state backtracking module obtains a lightweight data frame from the parameter stack through a loop pop operation. The lightweight data frame corresponding to the target backtracking simulation system timestamp is greater than the target backtracking simulation system timestamp and not greater than the current simulation system timestamp. The lightweight data frame corresponding to the target backtracking simulation system timestamp does not perform a pop operation. The historical state backtracking module processes lightweight data frames one by one from the current simulation system timestamp back to the target simulation system timestamp according to the simulation system timestamp, and queries the topology event hash table based on the globally unique identifiers of the spatial splitting events and spatial merging events recorded in the lightweight data frames; Within each lightweight data frame, the historical state backtracking module first processes the spatial merging event, then processes the spatial splitting event. When processing the spatial merging event, the historical state backtracking module performs spatial splitting and restores the spatial center coordinates of the eight child voxel nodes, the scalar parameter pseudo-density, and the pseudo-density increment of the current construction batch. When processing the spatial splitting event, the historical state backtracking module performs spatial merging and restores the spatial center coordinates of the parent voxel node, the scalar parameter pseudo-density, and the pseudo-density increment of the current construction batch. After restoring the spatial splitting and merging states corresponding to the current lightweight data frame, the historical state backtracking module configures the pseudo-density increment of the current construction batch of the voxel nodes in the leaf node state of the sparse octree voxel mesh to zero, and parses the input feature parameter set contained in the lightweight data frame; the historical state backtracking module maps the input feature parameter set to the three-dimensional spatial Gaussian decay function used by the input parameter mapping module, and calculates the pseudo-density increment of the current construction batch of each voxel node within the discrete time step corresponding to the current lightweight data frame; The historical state backtracking module performs a reverse inverse algebraic operation on the calculated pseudo-compaction increment of the current construction batch to obtain the historical state recovery compensation amount for each voxel node. Then, it performs an algebraic addition operation on the scalar parameter pseudo-compaction and the historical state recovery compensation amount to complete the scalar parameter pseudo-compaction recovery corresponding to the current lightweight data frame. After completing the processing of the current lightweight data frame, the historical state backtracking module configures the pseudo-compaction increment of the current construction batch corresponding to the current lightweight data frame to zero. After the scalar parameter pseudo-density recovery is completed, the historical state backtracking module instructs the dynamic rendering module to recalculate the spatial displacement of each voxel node and the grid vertex coordinates of the irregular triangular mesh according to S41 to S43, and instructs the pipeline linkage calculation module to recalculate the following spatial displacement vector of the pipeline control point according to S51, and then recalculate the pipeline surface mesh vertex coordinates according to the pipeline surface mesh vertex spatial center coordinate update operation in S52. The formula is expressed as: ; In the formula, The compensation amount is used to restore the historical state of voxel nodes in the leaf node state of the initial level sparse octree voxel mesh. This is the current timestamp of the simulation system; The simulation system timestamp is used to trace back to the target. It is the discrete time step between the current simulation system timestamp and the target backtracking simulation system timestamp; To the corresponding discrete time step The set of input feature parameters at time; A pre-defined attenuation function model is constructed based on the negative exponential empirical formula for geological compaction effect.

[0050] S63, the historical state backtracking module processes the globally unique identifier of each graph fracture event recorded in each popped lightweight data frame, and generates a hash key value based on the globally unique identifier of each graph fracture event. When a matching hash key value exists in the topology event hash table, the historical state backtracking module uses the hash key value to query the topology event hash table and extracts the spatial index corresponding to each pipeline surface mesh vertex at the same pipeline fracture location. The history state backtracking module sends a control command to the graphics processor to reset the marker; the graphics processor determines the corresponding pipeline surface mesh vertex in the video memory vertex buffer object according to the spatial index, and clears the fracture state marker data on the corresponding pipeline surface mesh vertex. The graphics processor's geometry shader program stops generating the two closed end cap polygon meshes corresponding to the pipeline fracture position recorded in the popped lightweight data frame, and restores the normal rendering output of the corresponding pipeline surface mesh continuous primitives. After processing all graphics breakage events recorded in the popped lightweight data frames, the graphics processor retains the unremoved breakage state marker data and writes the recalculated spatial center coordinates of the pipeline surface mesh vertices with unremoved breakage state marker data into the corresponding ghosting vertex resources. The graphics processor's geometry shader program stops outputting continuous primitives of the pipeline surface mesh containing unremoved breakage state marker data and generates two closed endcap polygon meshes corresponding to each pipeline breakage location. During the historical state backtracking process, the graphics breakage events are not written back into the topology event hash table.

[0051] To aid in understanding the technical solution of this invention, the following is an application example of a three-dimensional modeling and simulation system for port area road and storage yard construction.

[0052] Reference Appendix Figure 5 This invention is applied to the foundation reinforcement project of newly built container road stacking site in coastal port area. The initialization module generates an initial level of sparse octree voxel mesh and an irregular triangular mesh as a surface visualization model in three-dimensional space based on geological exploration elevation data and engineering geological survey report. The initialization module imports the three-dimensional vector model of underground heterogeneous pipe network. The initialization module allocates a data recording area in the main memory space and establishes a parameter stack and a topological event hash table. The parameter mapping module receives the set of process characteristic parameters including the coordinates of the construction action center and the construction operation energy level transmitted back by the tracked dynamic compaction machine. The parameter mapping module calculates the pseudo-compaction increment of each voxel node based on the process feature parameter set. The mesh reconstruction module monitors the density gradient change of the foundation soil in the impact compaction area based on the Euclidean modulus of the three-dimensional gradient vector of the calculated pseudo-compaction scalar field. The mesh reconstruction module dynamically generates sub-voxel nodes in areas where the Euclidean modulus of the three-dimensional gradient vector exceeds the preset splitting threshold. The dynamic rendering module performs integral calculation along the vertical coordinate axis on the difference between the scalar parameter pseudo-compaction and the preset pseudo-compaction benchmark value and updates the grid vertex coordinates of the irregular triangular mesh. The pipeline linkage calculation module transfers the foundation deformation to the pipeline control points after processing the anisotropic stiffness tensor matrix based on the spatial displacement of the voxel nodes. The pipeline linkage calculation module determines that pipeline fracture has occurred when the discrete Gaussian curvature of the grid vertex on the local pipeline surface of the drainage pipeline exceeds the discrete Gaussian curvature limit threshold. The pipeline network linkage calculation module marks the vertices of the pipeline surface mesh corresponding to the pipeline fracture location in the video memory buffer and issues an alarm. The geometry shader program of the graphics processor stops drawing continuous primitives of the pipeline surface mesh corresponding to the pipeline fracture location based on the fracture state marking data and generates a closed end cap polygon mesh. After receiving the alarm, the console issues a historical state backtracking command carrying the target backtracking timestamp. The historical state backtracking module extracts lightweight data frames from the parameter stack according to the last-in-first-out rule and uses the topological event hash table to restore the spatial merging state and spatial splitting state of the sparse octree voxel mesh. The historical state backtracking module performs inverse calculation of the three-dimensional spatial scalar field to restore the pseudo-density of the scalar parameters and instructs the graphics processor to clear the fracture state marking data. The simulation system backtracks to the state corresponding to the target backtracking timestamp.

[0053] While specific embodiments of the present invention have been described above, those skilled in the art should understand that these specific embodiments are merely illustrative. Those skilled in the art can omit, substitute, and modify the details of the above methods and systems in various ways without departing from the principles and essence of the present invention. For example, combining the above method steps to perform substantially the same function and achieve substantially the same result using substantially the same method falls within the scope of the present invention.

Claims

1. A three-dimensional modeling and simulation system for the construction of port area road storage yards, characterized in that, include: The initialization module is used to construct sparse octree voxel meshes and irregular triangular meshes, import the 3D vector model of underground heterogeneous pipe networks, and establish parameter stacks and topological event hash tables. The parameter mapping module is used to receive construction operation parameters, calculate the pseudo-compactness increment based on the three-dimensional Gaussian decay function, and update the scalar parameter pseudo-compactness. The mesh reconstruction module is used to compare the Euclidean magnitude of the 3D gradient vector with the preset splitting threshold and preset merging threshold, and perform spatial splitting or spatial merging. The dynamic rendering module is used to calculate the vertical geometric displacement based on the difference between the scalar parameter pseudo-density and the preset pseudo-density benchmark value, and update the grid vertex coordinates of the irregular triangular mesh. The pipeline linkage calculation module is used to calculate the following spatial displacement vector of the pipeline control points using the anisotropic stiffness tensor matrix, and to determine the fracture location of the same pipeline based on the discrete Gaussian curvature of the grid vertices on the pipeline surface, the discrete Gaussian curvature limit threshold, and the topological adjacency relationship. The historical state backtracking module is used to recover the spatial splitting state, spatial merging state, and scalar parameter pseudo-density of the sparse octree voxel mesh based on the historical state backtracking instructions, parameter stack, and topological event hash table.

2. The three-dimensional modeling and simulation system for port area road and storage yard construction according to claim 1, characterized in that, The initialization module assigns a preset pseudo-density benchmark value to the voxel nodes of the sparse octree voxel mesh at the initial level, parses the three-dimensional vector model of the underground heterogeneous pipe network with the data structure stored using the boundary representation method, extracts the pipe network control points and the mesh vertices on the pipe surface, and establishes a spatial index mapping binding relationship. Pre-allocate ghost vertex resources for pipeline surface mesh vertices within the graphics processor's video memory vertex buffer object, and set the initial graphics pipeline rendering state of the ghost vertex resources to hibernation and hiding. Based on the engineering physical material properties of the three-dimensional vector model, anisotropic stiffness tensor matrices are configured for each pipeline entity.

3. The three-dimensional modeling and simulation system for port area road and storage yard construction according to claim 1, characterized in that, The parameter mapping module receives the actual construction operation parameter set in the port area, extracts the coordinates of the construction action center, the energy level of the construction equipment, the construction process type and the effective influence space radius, converts the coordinates of the construction action center into three-dimensional spatial floating point coordinates in the graphics rendering pipeline, and determines the three-dimensional spatial scalar energy level and the influence range attenuation standard deviation. Before each construction batch begins, the pseudo-density increment of the voxel node in the current construction batch is initialized. The computation shader program of the graphics processor is called to calculate the pseudo-density increment of the voxel node within the effective spatial radius in parallel. The in-situ update of the scalar parameter pseudo-density is completed through parallel algebraic addition.

4. The three-dimensional modeling and simulation system for port area road and storage yard construction according to claim 1, characterized in that, The mesh reconstruction module uses the central difference method to solve the first-order partial derivatives of the scalar parameter pseudo-density along the horizontal, vertical and vertical coordinate axes, and calculates the Euclidean modulus of the three-dimensional gradient vector. When the Euclidean modulus exceeds the preset splitting threshold, a dynamic splitting operation is performed on the corresponding voxel node to generate a child voxel node, and the scalar parameter pseudo-compactness and the pseudo-compactness increment of the current construction batch are allocated according to the trilinear interpolation algorithm. When the Euclidean modulus of all child voxel nodes generated from the same parent voxel node is lower than the preset merging threshold, spatial merging is performed. The arithmetic mean of the pseudo-density of the scalar parameter of each child voxel node and the arithmetic mean of the pseudo-density increment are assigned to the recovered upper-level voxel node, and the spatial splitting event and spatial merging event are written into the topology event hash table.

5. The three-dimensional modeling and simulation system for port area road and storage yard construction according to claim 1, characterized in that, The dynamic rendering module extracts the scalar parameter pseudo density of voxel nodes in the leaf node state, calculates the difference between the scalar parameter pseudo density and the preset pseudo density benchmark value, and divides the spatial vertical column using the horizontal coordinate component and vertical coordinate component of the spatial center coordinate of the voxel node as the matching benchmark. The difference is integrated and accumulated algebraically along the vertical coordinate axis from bottom to top to obtain the columnar cumulative density increment. The columnar cumulative density increment is multiplied by a preset volume scaling factor to obtain the vertical geometric displacement and is recorded as the spatial displacement. After spatial splitting or merging, the spatial index mapping binding relationship between the grid vertices of the irregular triangular mesh and the voxel nodes in the uppermost leaf node state in space is re-established. Based on the relationship, the grid vertex coordinates of the irregular triangular mesh are updated and the terrain geometry deformation rendering screen is output.

6. The three-dimensional modeling and simulation system for port area road and storage yard construction according to claim 2, characterized in that, The pipeline linkage calculation module extracts the spatial displacement of voxel nodes including pipeline control points and the spatial displacement of multiple adjacent voxel nodes, and calculates the arithmetic mean of the spatial displacement of the multiple adjacent voxel nodes and the spatial displacement difference vector between the spatial displacement of the voxel node and the arithmetic mean. The spatial displacement difference vector is processed using the anisotropic stiffness tensor matrix to obtain the bending following spatial displacement vector of the pipeline control point. The bending following spatial displacement vector is then algebraically added to the arithmetic mean to obtain the following spatial displacement vector of the pipeline control point that includes both translation and bending. Update the spatial center coordinates of the pipeline surface mesh vertices based on the spatial index mapping binding relationship.

7. The three-dimensional modeling and simulation system for port area road and storage yard construction according to claim 6, characterized in that, The pipeline network linkage calculation module extracts triangular mesh primitives adjacent to the vertices of the pipeline surface mesh based on topological adjacency relationships, and calculates the discrete Gaussian curvature of the vertices of the pipeline surface mesh based on the principle of discrete differential geometry. When multiple pipeline surface mesh vertices with topological adjacency are sequentially connected to form a closed structure, and the discrete Gaussian curvature of each pipeline surface mesh vertex is greater than the discrete Gaussian curvature limit threshold of the corresponding pipeline entity, or less than the negative number of the limit threshold, the multiple pipeline surface mesh vertices are determined as the same pipeline fracture location, and fracture status labeling data is added to the pipeline surface mesh vertices that have not been attached with fracture status labeling data.

8. The three-dimensional modeling and simulation system for port area road and storage yard construction according to claim 7, characterized in that, The geometry shader program of the graphics processor stops outputting continuous primitives of the pipeline surface mesh containing fracture status marker data, writes the updated spatial center coordinates of the vertices of the pipeline surface mesh at the same pipeline fracture location into the corresponding ghost vertex resource, and uses the ghost vertex resource to instantiate and generate closed end cap polygon meshes on both sides of the pipeline fracture location. The pipeline linkage calculation module extracts the spatial center coordinates of each closed end cap polygonal grid, combines the pipeline code identifier and the simulation system timestamp to generate a globally unique identifier for the graphic fracture event, asynchronously writes the graphic fracture event information into the graphics processor's video memory buffer, transmits it to the main memory space after the video memory barrier instruction is released, and uses the globally unique identifier as a hash key to write the graphic fracture event information into the topology event hash table.

9. The three-dimensional modeling and simulation system for port area road and storage yard construction according to claim 1, characterized in that, After a single simulation frame is rendered, the historical state backtracking module extracts the input feature parameter set, the simulation system timestamp, and the globally unique identifiers of the spatial splitting event, spatial merging event, and graphic breakage event that occurred in the current simulation frame. Based on the last-in-first-out rule, it encapsulates these into a lightweight data frame and pushes it onto the parameter stack. After receiving the historical state backtracking instruction carrying the target backtracking simulation system timestamp, a lightweight data frame with a simulation system timestamp greater than the target backtracking simulation system timestamp and not greater than the current simulation system timestamp is obtained through a loop pop operation. The lightweight data frame corresponding to the target backtracking simulation system timestamp is not popped from the stack.

10. The three-dimensional modeling and simulation system for port area road and storage yard construction according to claim 9, characterized in that, The historical state backtracking module queries the topology event hash table based on the globally unique identifiers of the spatial splitting and spatial merging events recorded in the popped lightweight data frames, and processes the lightweight data frames one by one from the current simulation system timestamp back to the target simulation system timestamp. Perform spatial splitting on the spatial merging event, and restore the spatial center coordinates, scalar parameter pseudo-compactness, and pseudo-compactness increment of each sub-voxel node; Perform spatial merging on the spatial splitting event to restore the spatial center coordinates, scalar parameter pseudo-density, and the pseudo-density increment of the parent voxel node; Before and after each lightweight data frame is processed, the pseudo-density increment of the voxel nodes in the leaf node state is initialized. The input feature parameter set is parsed, the three-dimensional Gaussian decay function is called to recalculate the pseudo-density increment and perform inverse negation algebra operation to obtain the historical state recovery compensation amount, and the scalar parameter pseudo-density is recovered by algebraic addition operation. Based on the restored scalar parameter pseudo-compactness, the spatial displacement, the grid vertex coordinates of the irregular triangular mesh, and the grid vertex coordinates of the pipeline surface are recalculated. Based on the globally unique identifier of the graphic fracture event, query the topology event hash table and extract the spatial index, clear the fracture state marker data of the corresponding pipeline surface mesh vertex, stop generating the corresponding closed end cap polygon mesh and restore the continuous primitive output of the pipeline surface mesh; For pipeline surface mesh vertices whose fracture state marker data has not been cleared, the recalculated spatial center coordinates are written to the corresponding ghost vertex resources, the output of the corresponding pipeline surface mesh continuous primitives is stopped and a closed end cap polygon mesh is generated, and the graphic fracture event is not written to the topology event hash table again during the historical state backtracking process.

Images