Virtual restoration simulation method and system for cultural relics based on digital twinning

Abstract

The application relates to the technical field of digital twinning, and discloses a cultural relic virtual restoration simulation method and system based on digital twinning. The method comprises the following steps: a cultural relic digital twinning body is constructed by fusing geometric shapes and physical properties; collision contact between a virtual tool and a model is detected in real time; stress, deformation and reaction force in a contact area are calculated through real-time finite element analysis; the reaction force is decomposed into joint torque of a multi-degree-of-freedom force feedback device and a high-frequency vibration signal, dynamic matching material property characteristic force tactile feedback is provided; and the model physical parameters and geometric shapes are updated according to the stress overrun condition, so that the visual simulation of damage accumulation and fracture is realized. The system comprises a digital twinning construction module, a user interaction interface module, a physical simulation engine module, a force feedback rendering module and a model state updating module. Through the high-frequency closed-loop physical-force sensation coupling mechanism, the application improves the authenticity, immersion and training reliability of virtual restoration.

Description

Technical Field

[0001] This invention belongs to the field of digital twin technology, specifically relating to a method and system for virtual restoration simulation of cultural relics based on digital twins. Background Technology

[0002] With the increasing application of digital twin technology in the field of cultural heritage protection, virtual reality-based artifact restoration simulation systems have become an important tool to assist experts in conducting non-contact restoration research. These systems, through high-precision 3D modeling and interactive visualization, can reproduce the geometric shape and surface texture of artifacts in digital space, providing support for the virtual splicing of damaged structures, intelligent completion of missing areas, and the rehearsal of restoration plans.

[0003] However, the core interactive elements of such systems have long been limited to visual and auditory feedback, lacking a realistic simulation of the tactile dimension. This results in restorers being unable to perceive the physical properties of the artifacts themselves when operating virtual tools, especially key mechanical characteristics such as material stiffness, brittleness, and local stress response.

[0004] Digital twin-based virtual restoration simulation of cultural relics focuses on building a physically consistent interactive environment, aiming to make the virtual operation process as close as possible to the tactile experience of real restoration. This approach relies on the modeling capabilities of a high-fidelity physics engine for the mechanical behavior of cultural relic materials and needs to work in conjunction with force feedback devices to achieve two-way human-computer interaction.

[0005] Ideally, when a virtual carving knife or tweezers comes into contact with a fragment of pottery, the system should calculate the contact force in real time based on parameters such as the elastic modulus, yield strength, and fracture toughness of the material in the database, and drive the force feedback device to output the corresponding resistance, thereby forming a closed-loop tactile perception.

[0006] In existing technologies, mainstream virtual restoration platforms generally use general physics engines to simulate rigid bodies or simple elastic bodies, but they do not establish refined constitutive models for the diversity of cultural relic materials; at the same time, commercial force feedback devices can only provide vibration feedback with fixed intensity or preset modes, and cannot dynamically adjust the magnitude and direction of the output force according to the material properties of the contact point.

[0007] This static feedback mechanism can easily lead restorers to apply excessive force during virtual operations, causing non-physical deformation or even structural collapse of the digital model, severely undermining the credibility of the simulation results. Especially when dealing with highly sensitive artifacts such as thin-walled objects or fragile painted layers, the lack of adaptive tactile feedback not only reduces operational precision but may also mislead restoration decisions.

[0008] Therefore, there is an urgent need for an adaptive tactile simulation method that integrates the physical properties of materials with force feedback devices to achieve a realistic, safe, and reliable force-feedback interactive experience during the virtual restoration of cultural relics. Summary of the Invention

[0009] The technical problem this invention aims to solve is that existing virtual cultural relic restoration technologies lack a high-fidelity force-sensing interaction mechanism based on the real physical characteristics of cultural relics. Restoration personnel cannot perceive the material stiffness, surface texture, and mechanical response of the contact points between virtual tools and cultural relic models through touch. Existing force feedback devices can only provide fixed and non-specific vibration prompts, leading to improper force application by operators. This can easily cause irreversible virtual damage to the digital model during the simulation process, exceeding the laws of real physics, and even causing the collapse of the model's data structure, thereby severely weakening the training value of virtual restoration and the reliability of scheme simulation.

[0010] To address the technical deficiencies in the aforementioned background technology, this invention provides a method and system for virtual restoration simulation of cultural relics based on digital twins, the core technical solution of which is as follows:

[0011] A digital twin model of a cultural relic is constructed, integrating its geometric shape, physical material properties, and damage accumulation state. Through a high-frequency closed-loop physical simulation and force-feedback rendering process, the contact force, material deformation, and damage evolution process under the interaction between the virtual repair tool and the digital twin model are calculated in real time. The calculated multi-dimensional force-feedback data is then rendered in real time to a multi-degree-of-freedom force feedback device, providing the operator with dynamic force, vibration, and damping feedback that precisely matches the material properties of specific parts of the cultural relic. This achieves an immersive and high-fidelity virtual repair operation experience and ensures that the mechanical behavior of the virtual repair process conforms to real physical laws.

[0012] This invention provides a virtual restoration simulation method for cultural relics based on digital twins, comprising the following steps:

[0013] A three-dimensional digital twin model of a cultural relic is constructed. The digital twin model includes a geometric data layer that represents the geometric shape and internal structure of the cultural relic, and a physical property data layer that is spatially registered with the geometric data layer and records the physical properties of the materials of each part of the cultural relic.

[0014] The physical property data layer discretizes the geometric data layer into a finite element voxel mesh and assigns a set of preset physical parameters to each voxel element, including Young's modulus, Poisson's ratio, material density, and fracture toughness threshold.

[0015] Receive and process the position, orientation, and active force data of the virtual repair tool in three-dimensional space, as well as the data of the active force applied by the operator, which are output in real time by the multi-degree-of-freedom force feedback operation device;

[0016] Real-time detection of collisions between the virtual repair tool and the geometric data layer of the digital twin model, and determination of the set of voxel units involved in the collision contact area;

[0017] The physical simulation engine is triggered, and based on the active force data and the physical parameters of the voxel elements in the collision contact area, the internal stress distribution, strain tensor, and deformation displacement of the voxel elements are calculated using the real-time finite element analysis method; at the same time, the reaction force vector acting on the virtual repair tool is calculated and generated.

[0018] The reaction force vector is decomposed into target torque signals applied to each drive joint of the multi-degree-of-freedom force feedback operating device and target high-frequency vibration signals applied to the end effector of the device, and the device is driven to generate force and tactile feedback corresponding to the reaction force vector.

[0019] The physical property data layer of the digital twin model is updated in real time based on the internal stress values ​​of the voxel units calculated by the physical simulation engine.

[0020] When the internal stress value of any voxel unit exceeds its preset fracture toughness threshold, its Young's modulus parameter is dynamically adjusted to the preset failure value to simulate permanent damage or fracture of the material, and the geometric data layer of the digital twin model is updated simultaneously to present the corresponding damage morphology in the visualization interface.

[0021] As one embodiment of the present invention, the construction of the three-dimensional digital twin model of the cultural relic specifically includes:

[0022] Multi-layer slice image data or high-density point cloud data of cultural relics are obtained through industrial-grade computer tomography equipment or structured light 3D scanning equipment.

[0023] The acquired raw data is denoised, registered and fused, and the moving cube algorithm or Poisson reconstruction algorithm is used to generate the initial high-precision polygonal surface mesh model of the cultural relic, which constitutes the geometric data layer;

[0024] Based on the surface mesh model, a voxelized finite element mesh filled with tetrahedral or hexahedral elements is generated.

[0025] Based on the material analysis report of the cultural relic or the non-destructive testing results of samples of similar materials, each voxel element in the finite element mesh is assigned its corresponding Young's modulus, Poisson's ratio, material density, and fracture toughness threshold in a database manner, thus completing the construction of the physical property data layer.

[0026] In one embodiment of the present invention, the multi-degree-of-freedom force feedback operating device is a six-degree-of-freedom serial robotic arm structure, with a pen-shaped operating tool with a high-frequency piezoelectric ceramic vibrator fixedly connected to its end; the receiving and processing of its output data specifically includes:

[0027] The rotation angle of each joint is obtained by an absolute encoder installed at each joint of the serial robotic arm at a sampling frequency greater than 1000 Hz, and the three-dimensional position and orientation of the pen-shaped operating tool end in the global coordinate system are determined in real time by forward kinematics calculation.

[0028] The triaxial force sensor, based on strain gauge technology and built into the pen-shaped tool, measures in real time the multi-dimensional active forces exerted by the operator's hand on the tool, such as pushing, pulling, and torsion, and converts them into digital signals.

[0029] As one embodiment of the present invention, the calculation of the internal stress distribution, strain tensor, and deformation displacement of the voxel element using real-time finite element analysis specifically executes the following calculation process:

[0030] When a collision is detected, the active force is applied as a boundary condition to the voxel unit nodes in the contact area;

[0031] Establish and solve the global stiffness matrix equation corresponding to the digital twin model. The equation is in the form that the global force vector is equal to the product of the global stiffness matrix and the global nodal displacement vector.

[0032] The global stiffness matrix is ​​assembled from the element stiffness matrix based on the geometry of all voxel elements and their Young's modulus and Poisson's ratio parameters.

[0033] By using iterative solvers such as the conjugate gradient method or the multigrid method, the displacement vectors of all nodes can be efficiently solved, thereby obtaining the deformation field of the entire model.

[0034] Based on nodal displacements, the strain tensor of each voxel element is calculated using shape function differentiation, and then the corresponding stress tensor is calculated according to the constitutive relation, i.e. Hooke's law, to obtain the internal stress distribution.

[0035] As one embodiment of the present invention, decomposing the reaction force vector into a target torque signal and a target high-frequency vibration signal specifically includes:

[0036] The reaction force vector is decomposed into translational force components in three orthogonal directions and rotational torque components on three orthogonal axes; the translational force components are used to simulate the pushing and resistance sensations during contact, and the rotational torque components are used to simulate the torsional damping when the tool is embedded in the material.

[0037] By transposing the inverse dynamic Jacobian matrix, the translational force component and the rotational torque component are mapped to the target torque required to be output by the drive motors of each joint of the serial robotic arm, and the drive motors are driven by a closed-loop torque controller to generate precise force feedback.

[0038] Extract the high-frequency components from the reaction force vector, or query a preset material texture database based on the material property data of the voxel unit in the contact area, to generate a sinusoidal or composite waveform vibration control signal with a specific frequency and amplitude;

[0039] The vibration control signal is applied to the piezoelectric ceramic vibrator inside the pen-shaped operating tool, causing it to generate tactile vibration feedback corresponding to the micro-texture or material particle feel of the artifact's surface.

[0040] As one embodiment of the present invention, updating the physical attribute data layer and geometric data layer of the digital twin model specifically includes:

[0041] Establish a damage accumulation model, in which each voxel unit is associated with a damage state variable, with an initial value of 0;

[0042] Within each physical simulation time step, if the equivalent stress of the voxel element, such as the von Mises stress, is greater than its fracture toughness threshold, then its damage state variables are accumulated according to the preset damage evolution rule.

[0043] When the damage state variable reaches a preset critical value of 1, the Young's modulus parameter of the voxel unit is multiplied by a decay coefficient less than 1 to simulate the degradation of material stiffness.

[0044] Meanwhile, in the visualization rendering module, the damage is visualized by adjusting the color and transparency or applying a preset crack texture map based on the damage state variable value of the voxel unit.

[0045] When the damage state variable reaches 1, indicating complete failure, the system removes the voxel element from the global stiffness matrix of the finite element analysis and generates a geometric discontinuity surface representing crack propagation in the geometric data layer, thereby simultaneously reflecting the macroscopic fracture of the material visually and mechanically.

[0046] According to another aspect of the present invention, a virtual restoration simulation system for cultural relics based on digital twins is provided, comprising:

[0047] The digital twin construction module for cultural relics is used to collect three-dimensional geometric data and material physical property data of cultural relics, and generate a three-dimensional digital twin model containing a geometric data layer and a physical property data layer. The physical property data layer is a finite element voxel mesh registered with the geometric data layer, and each voxel unit is assigned physical parameters including Young's modulus, Poisson's ratio, material density and fracture toughness threshold.

[0048] The user interaction interface module, coupled with a multi-degree-of-freedom force feedback operation device, is used to acquire the position, posture and active force input applied by the operator to the virtual repair tool in real time, and output force and tactile feedback signals to the operator.

[0049] The physical simulation engine module is configured to, when a collision is detected between the virtual repair tool and the digital twin model, calculate the internal stress and strain distribution and reaction force vector of the model using real-time finite element analysis based on the active force input and the physical parameters of the voxel elements in the contact area.

[0050] The force feedback rendering module, which connects the physical simulation engine module and the user interaction interface module, is used to convert the reaction force vector into the target torque and high-frequency vibration control signal that drives the multi-degree-of-freedom force feedback operation device.

[0051] The model state update module adjusts the physical parameters of the voxel units in the digital twin model that are subjected to excessive stress in real time according to the calculation results of the physical simulation engine module to simulate damage accumulation and fracture, and updates their geometric data to achieve damage visualization.

[0052] As one embodiment of the present invention, the digital twin construction module for cultural relics includes an industrial computer tomography data interface, a three-dimensional point cloud processing unit, a polygon mesh generation unit, and a finite element voxelization unit.

[0053] The physical simulation engine module is deployed on a heterogeneous computing platform where a central processing unit (CPU) and a graphics processing unit (GPU) work together, with the GPU being used to accelerate the assembly and solution of the global stiffness matrix in parallel.

[0054] As one embodiment of the present invention, the multi-degree-of-freedom force feedback operation device in the user interaction interface module includes a base, a serial robotic arm composed of multiple joints and links, a DC servo motor and a high-resolution encoder installed on each joint, a pen-shaped operation tool with a built-in triaxial force sensor connected to the end, and a piezoelectric ceramic vibrator integrated into the tip of the pen-shaped operation tool.

[0055] As one embodiment of the present invention, the force feedback rendering module includes an inverse dynamics calculation unit and a texture vibration generation unit; the inverse dynamics calculation unit is used to map the reaction force vector in Cartesian space to the joint torque space of the robotic arm;

[0056] The texture vibration generation unit has a built-in editable material texture library, which establishes a mapping relationship between material names and specific vibration waveform parameters, including frequency, amplitude and envelope.

[0057] As one embodiment of the present invention, the model state update module further includes a visualization rendering unit, which is bidirectionally coupled to the physical simulation engine module;

[0058] It not only renders the updated geometric data layer into a three-dimensional image, but also displays the internal stress or strain distribution of the model calculated by the physical simulation engine module in real time on the surface of the cultural relic model in the form of a pseudo-color cloud map, providing the operator with an insight into the internal mechanical state that surpasses real vision.

[0059] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0060] 1. By constructing a digital twin model containing precise physical properties and using real-time finite element analysis for physical simulation, this invention elevates virtual repair operations from purely geometric interaction to a mechanical interaction level based on physical laws, greatly enhancing the realism of the simulation.

[0061] 2. This invention establishes a high-frequency closed-loop system from the operator's application of force to physical simulation calculation and then to force feedback rendering, enabling the operator to perceive the stiffness, elasticity, brittleness and surface texture of different cultural relics in real time through touch, solving the problem of missing tactile feedback or single feedback information in the existing technology.

[0062] 3. The proposed damage accumulation and fracture model based on stress threshold can scientifically simulate the complete physical process of cultural relic materials from elastic deformation and plastic deformation to final fracture when the repair tool applies excessive force. It avoids the drawbacks of the model in the prior art that does not realistically collapse suddenly or have no damage response, and provides a reliable basis for verifying the mechanical feasibility of the repair scheme.

[0063] 4. By finely decomposing the reaction force vector into the torque signal driving the robotic arm and the tactile signal driving the vibrator, the separation rendering of force and touch is achieved, providing the operator with a rich and multi-dimensional interactive experience, enhancing the immersion of virtual repair operations and the effectiveness of skills training. Attached Figure Description

[0064] Figure 1 This is a schematic diagram of the overall technical solution architecture of the virtual restoration simulation method and system for cultural relics based on digital twins proposed in this invention;

[0065] Figure 2 This is a schematic diagram of the core principle framework of physical simulation and force feedback closed-loop rendering based on real-time finite element analysis in this invention.

[0066] Figure 3 This is a logical flowchart of the construction stage of the digital twin model of cultural relics in this invention;

[0067] Figure 4 This is a logical flowchart of the collision detection and physical response calculation stage under the interaction between the virtual repair tool and the digital twin in this invention.

[0068] Figure 5 This is a schematic diagram of the multi-level interaction relationship and data flow between the multi-degree-of-freedom force feedback operation device and various modules of the system in this invention;

[0069] Figure 6 This is a logical flowchart of the damage accumulation and geometric-physical state synchronous update mechanism based on stress threshold in this invention. Detailed Implementation

[0070] Please refer to Figures 1 to 6 This invention provides a virtual restoration simulation method and system for cultural relics based on digital twins, aiming to solve the problem in existing virtual restoration technologies where the lack of realistic tactile feedback prevents operators from perceiving the material stiffness, surface texture, and mechanical response of cultural relics.

[0071] Existing force feedback devices can only output vibration signals of fixed intensity and cannot dynamically adjust the feedback content according to the local physical characteristics of cultural relics. This causes virtual tools to overload, leading to the collapse of digital model structures or damage behaviors that do not conform to physical laws, which seriously weakens the reliability and effectiveness of virtual restoration in skills training and scheme simulation.

[0072] To overcome the above-mentioned shortcomings, this invention constructs a digital twin model of cultural relics that integrates geometric shape, physical properties and damage state, and realizes real-time calculation and multi-dimensional force feedback of contact force, material deformation and damage evolution during virtual restoration through a high-frequency closed-loop physical simulation and force perception rendering mechanism.

[0073] The digital twin-based virtual restoration simulation method for cultural relics includes the following steps:

[0074] Constructing a three-dimensional digital twin model of cultural relics;

[0075] Receive and process the position, attitude, and active force data of the virtual repair tool output by the multi-degree-of-freedom force feedback operation device;

[0076] Real-time detection of collisions and contact between the virtual repair tool and the digital twin model;

[0077] The physical simulation engine is triggered to calculate the stress, strain, and displacement of the voxel elements in the contact area and generate a reaction force vector. The reaction force vector is decomposed into a target torque signal and a high-frequency vibration signal, which drives the force feedback device to generate precise force and tactile feedback.

[0078] The physical properties and geometry of the digital twin model are dynamically updated based on the simulation results to simulate material damage and fracture processes.

[0079] First, perform step S1: Construct a three-dimensional digital twin model of the cultural relic.

[0080] The model consists of two core data layers: a geometric data layer and a physical property data layer.

[0081] The geometric data layer represents the external outline and internal structure of the cultural relic, while the physical attribute data layer is strictly registered with the geometric data layer in space to record the physical parameters of the materials in each part of the cultural relic.

[0082] In practice, multi-layer slice image data or high-density point cloud data of cultural relics are obtained through industrial-grade computer tomography equipment or structured light 3D scanning equipment.

[0083] After denoising, registration, and fusion processing, the original scan data is used to generate an initial high-precision polygonal surface mesh model using the moving cube algorithm or the Poisson reconstruction algorithm, which forms the basis of the geometric data layer.

[0084] Based on this surface mesh model, the system further generates an internally filled finite element voxel mesh, which is composed of tetrahedral or hexahedral units and covers the entire volume area of ​​the artifact.

[0085] Each voxel unit is assigned a set of physical parameters, including Young's modulus, Poisson's ratio, material density, and fracture toughness threshold. These parameters are derived from non-destructive testing reports of the artifact itself, or from standardized datasets obtained through laboratory testing of similar historical material samples.

[0086] All parameters are stored in a database and indexed to the spatial coordinates of the corresponding voxel elements to ensure quick retrieval and retrieval in subsequent physical simulations.

[0087] The resulting digital twin model not only possesses high-fidelity geometric form but also incorporates mechanical behavior characteristics consistent with real cultural relics, laying the foundation for subsequent high-precision interactive simulation.

[0088] Then, step S2 is executed: receiving and processing the position, orientation, and active force data of the virtual repair tool in three-dimensional space, as well as the data of the active force applied by the operator, which are output in real time by the multi-degree-of-freedom force feedback operation device.

[0089] The multi-degree-of-freedom force feedback operation device adopts a six-degree-of-freedom serial robotic arm structure, with its base fixed to the operating table and a pen-shaped operating tool fixed to its end.

[0090] This pen-shaped tool integrates a triaxial force sensor based on strain gauge technology to measure the pushing, pulling, and torsional torques applied by the operator's hand.

[0091] High-resolution absolute encoders are installed at each joint of the robotic arm, with a sampling frequency greater than 1000 Hz.

[0092] The system converts the rotation angles of each joint into the three-dimensional position and Euler angle orientation of the pen-shaped tool tip in the global coordinate system through forward kinematics calculation.

[0093] Meanwhile, the analog signal output by the triaxial force sensor is converted into a digital force vector after analog-to-digital conversion, which contains force components in three orthogonal directions.

[0094] Position, attitude, and force data are synchronously packaged into timestamp-aligned data frames and transmitted to the central processing unit at fixed intervals.

[0095] This data frame serves as the main input boundary condition for the physical simulation, and its high sampling rate and low latency characteristics ensure the real-time performance and stability of the entire interactive closed loop.

[0096] Next, step S3 is executed: real-time detection of collisions between the virtual repair tool and the geometric data layer of the digital twin model, and determination of the set of voxel units involved in the collision contact area.

[0097] The system uses a hierarchical bounding box acceleration structure to spatially divide the digital twin model. Combined with a continuous collision detection algorithm, it determines whether the geometric proxy model of the pen-shaped tool (usually simplified to a cylinder or capsule) penetrates or contacts the mesh on the surface of the artifact in each simulation time step.

[0098] Once a collision event is detected, the system immediately locates the precise position of the contact point in the global coordinate system and queries the corresponding voxel unit through a spatial indexing mechanism.

[0099] Because the voxel mesh is strictly aligned with the surface mesh, the contact point can be uniquely mapped to one or more adjacent voxel units.

[0100] The system organizes these voxel units and their neighboring units into local subdomains, which serve as the computational domain for subsequent physical simulations.

[0101] The selection of this local subdomain balances computational efficiency with boundary effect control, ensuring the integrity of the stress transfer path.

[0102] Then, step S4 is executed: the physical simulation engine is triggered, and based on the active force data and the physical parameters of the voxel elements in the collision contact area, the internal stress distribution, strain tensor and deformation displacement of the voxel elements are calculated using the real-time finite element analysis method, and the reaction force vector acting on the virtual repair tool is generated.

[0103] The physics simulation engine is deployed on a heterogeneous computing platform where the central processing unit and the graphics processing unit work together, with the graphics processing unit responsible for parallel acceleration of critical computing tasks.

[0104] After a collision is detected, the system applies an active force as a Dirichlet boundary condition to the nodes of the voxel elements in the contact region.

[0105] The engine then calculates the local stiffness matrix of each element using standard finite element shape functions based on the geometry of all voxel elements and their Young's modulus and Poisson's ratio parameters, and assembles them into a global stiffness matrix.

[0106] The global stiffness matrix equation is expressed as:

[0107] ;

[0108] For the global force vector, The global stiffness matrix. Let be the global nodal displacement vector to be solved.

[0109] Since the global stiffness matrix is ​​sparse and symmetric positive definite, the system uses the conjugate gradient method or multigrid method for iterative solution to obtain the displacement solution of all nodes in milliseconds.

[0110] After obtaining the displacement field, the system uses the spatial derivatives of the shape functions to calculate the strain tensor of each voxel element. Then, the stress tensor is calculated based on the generalized Hooke's law. :

[0111] ;

[0112] The elastic matrix is ​​composed of Young's modulus and Poisson's ratio. The reaction force vector is derived from the stress state at the contact node using the principle of virtual work; its direction is opposite to that of the active force, and its magnitude is determined by the local stiffness and deformation.

[0113] This reaction force vector contains translational force components and rotational torque components, which fully describe the mechanical resistance of the artifact to the virtual tool.

[0114] Then, step S5 is executed: the reaction force vector is decomposed into the target torque signal applied to each drive joint of the multi-degree-of-freedom force feedback operating device and the target high-frequency vibration signal applied to the end effector of the device, and the device is driven to generate corresponding force and tactile feedback.

[0115] The system first transforms the reaction force vector from Cartesian space to the joint space of the robotic arm. This transformation is achieved by transposing the inverse dynamics Jacobian matrix:

[0116] ;

[0117] Here are the target torque vectors for each joint. For the current joint configuration The Jacobian matrix below, The reaction force vector in Cartesian space. This is a transpose.

[0118] The target torque signal is sent to the closed-loop torque controller, which drives the DC servo motors of each joint to output precise resistance torque, allowing the operator to feel the real pushing resistance and torsional damping.

[0119] Meanwhile, the system extracts the high-frequency fluctuation components from the reaction force vector, or queries the built-in material texture database based on the material type of the voxel unit in the contact area. This database stores vibration waveform parameters corresponding to different materials, including fundamental frequency, harmonic components, amplitude envelope, and duration.

[0120] The system generates a composite sinusoidal vibration control signal with a specific frequency and amplitude, which is then applied to a piezoelectric ceramic vibrator at the tip of a pen-shaped tool. The vibrator produces high-frequency mechanical vibrations with micron-level amplitudes to simulate the microscopic granular texture or the frictional feel of cracks on the surface of cultural relics, achieving a high-fidelity reproduction of tactile texture.

[0121] Finally, step S6 is executed: based on the internal stress values ​​of the voxel units calculated by the physical simulation engine, the physical property data layer of the digital twin model is updated in real time; when the internal stress value of any voxel unit is greater than its preset fracture toughness threshold, its Young's modulus parameter is dynamically adjusted to the preset failure value, and the geometric data layer is updated synchronously to present the damage morphology.

[0122] The system maintains damage state variables for each voxel unit. The initial value is 0.

[0123] Within each simulation time step, if the von Mises equivalent stress of a certain voxel element... Greater than its fracture toughness threshold Then update according to the following damage evolution rules. :

[0124] ;

[0125] For the updated damage state variables, For the current damage state variable, This is the simulation time step.

[0126] when When the critical value of 1 is reached, the system multiplies the Young's modulus of the voxel element by a decay factor (e.g., 0.1) to simulate material stiffness degradation. If If the value continues to accumulate until it equals 1, the element is considered completely failed. Its Young's modulus is set to a minimum value close to 0, and it is removed from the global stiffness matrix, no longer participating in subsequent mechanical calculations. At the visualization level, the system... The values ​​dynamically adjust the color, transparency, or overlay crack texture maps of the corresponding areas.

[0127] When multiple adjacent voxel units fail simultaneously, the system generates a geometric discontinuity surface in the geometric data layer to represent the crack propagation path. This surface is smoothly embedded into the original model through a subdivision surface algorithm to ensure visual consistency.

[0128] This mechanism enables the simulation of the entire process from microscopic stress accumulation to macroscopic fracture phenomena, making the damage behavior in virtual repair conform to real physical laws.

[0129] The digital twin-based virtual restoration simulation system for cultural relics includes a digital twin construction module for cultural relics, a user interaction interface module, a physical simulation engine module, a force feedback rendering module, and a model state update module.

[0130] The digital twin construction module for cultural relics includes an industrial computer computed tomography (CT) data interface, a 3D point cloud processing unit, a polygon mesh generation unit, and a finite element voxelization unit, which is responsible for completing the entire process of building a voxel mesh model with physical properties from raw scan data.

[0131] The user interaction interface module is coupled with a six-degree-of-freedom force feedback operation device, which consists of a base, a serial robotic arm, a joint servo motor, a high-resolution encoder, a pen-shaped tool with a built-in triaxial force sensor, and a piezoelectric ceramic vibrator, to achieve high-precision input acquisition and multi-dimensional force feedback output.

[0132] The physics simulation engine module is deployed on a heterogeneous computing platform, utilizing graphics processors to accelerate the assembly and solution of stiffness matrices in parallel, ensuring real-time performance.

[0133] The force feedback rendering module includes an inverse dynamics calculation unit and a texture vibration generation unit. The former completes the mapping of force vectors to joint torques, while the latter generates tactile vibration signals based on the material type.

[0134] The model status update module includes a visualization rendering unit, which not only updates the geometry but also overlays and displays the internal stress or strain distribution on the surface of the artifact in real time in the form of a pseudo-color cloud map, providing operators with mechanical insights beyond the naked eye.

[0135] The entire system operates under a unified time synchronization framework, with each module exchanging data at fixed intervals to form a high-frequency closed loop of "perception-simulation-feedback-update".

[0136] This closed-loop mechanism ensures that every tiny movement of the operator triggers a physically compliant model response and provides real-time feedback through force and touch channels, thereby achieving an immersive, high-fidelity virtual restoration experience.

[0137] This technical solution fundamentally solves the problems of tactile loss and model behavior distortion in traditional virtual restoration, providing a scientific and reliable digital platform for cultural relic protection, restoration training, and scheme verification.

[0138] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish an entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.

[0139] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A virtual restoration simulation method for cultural relics based on digital twins, characterized in that, include: Construct a three-dimensional digital twin model of the cultural relic, wherein the digital twin model includes a geometric data layer and a physical attribute data layer; Receive and process the position, orientation, and active force data of the virtual repair tool in three-dimensional space, as well as the data of the active force applied by the operator, which are output in real time by the multi-degree-of-freedom force feedback operation device; Real-time detection of collisions between the virtual repair tool and the geometric data layer of the digital twin model, and determination of the set of voxel units involved in the collision contact area; The physical simulation engine is triggered, and based on the active force data and the physical parameters of the voxel elements in the collision contact area, the internal stress distribution, strain tensor, and deformation displacement of the voxel elements are calculated using the real-time finite element analysis method; at the same time, the reaction force vector acting on the virtual repair tool is calculated and generated. The reaction force vector is decomposed into target torque signals applied to each drive joint of the multi-degree-of-freedom force feedback operating device and target high-frequency vibration signals applied to the end effector of the device, and the device is driven to generate force and tactile feedback corresponding to the reaction force vector, specifically including: The reaction force vector is transformed from Cartesian space to the joint space of the robotic arm; the transformation is achieved by transposing the inverse dynamics Jacobian matrix: ； Here are the target torque vectors for each joint. For the current joint configuration The Jacobian matrix below, The reaction force vector in Cartesian space. For transpose; The target torque signal is sent to the closed-loop torque controller, which drives the DC servo motors of each joint to output precise resistance torque; At the same time, the high-frequency fluctuation components in the reaction force vector are extracted, or the built-in material texture database is queried according to the material type of the voxel unit in the contact area; the material texture database stores the vibration waveform parameters corresponding to different materials, including fundamental frequency, harmonic components, amplitude envelope and duration; A composite sinusoidal vibration control signal is generated and applied to the tip of a pen-shaped operating tool; the piezoelectric ceramic vibrator generates high-frequency mechanical vibration with a micrometer-level amplitude. Based on the internal stress values ​​of the voxel units calculated by the physical simulation engine, the physical property data layer of the digital twin model is updated in real time. When the internal stress value of any voxel unit is greater than its preset fracture toughness threshold, its Young's modulus parameter is dynamically adjusted to the preset failure value to simulate permanent damage or fracture of the material, and the geometric data layer of the digital twin model is updated simultaneously to present the corresponding damage morphology in the visualization interface.

2. The method for virtual restoration simulation of cultural relics based on digital twins according to claim 1, characterized in that, The construction of a three-dimensional digital twin model of a cultural relic specifically includes: Multi-layer slice image data or high-density point cloud data of cultural relics are obtained through industrial-grade computer tomography equipment or structured light 3D scanning equipment. The acquired raw data is denoised, registered and fused, and the moving cube algorithm or Poisson reconstruction algorithm is used to generate the initial high-precision polygonal surface mesh model of the cultural relic, which constitutes the geometric data layer; Based on the surface mesh model, a voxelized finite element mesh filled with tetrahedral or hexahedral elements is generated. Based on the material analysis report of the cultural relic or the non-destructive testing results of samples of similar materials, each voxel element in the finite element mesh is assigned its corresponding Young's modulus, Poisson's ratio, material density, and fracture toughness threshold in a database manner, thus completing the construction of the physical property data layer.

3. The method for virtual restoration simulation of cultural relics based on digital twins according to claim 2, characterized in that, The multi-degree-of-freedom force feedback operating device is a six-degree-of-freedom serial robotic arm structure, with a pen-shaped operating tool with a high-frequency piezoelectric ceramic vibrator fixedly attached to its end; the receiving and processing of its output data specifically includes: The rotation angle of each joint is obtained by an absolute encoder installed at each joint of the serial robotic arm, and the three-dimensional position and orientation of the pen-shaped tool end in the global coordinate system are determined in real time by forward kinematics calculation. The triaxial force sensor, based on strain gauge technology and built into the pen-shaped tool, measures in real time the multi-dimensional active forces exerted by the operator's hand on the tool, including pushing, pulling, and torsion, and converts them into digital signals.

4. The virtual restoration simulation method for cultural relics based on digital twins according to claim 3, characterized in that, The real-time finite element analysis method is used to calculate the internal stress distribution, strain tensor, and deformation displacement of the voxel element. The specific calculation process is as follows: When a collision is detected, the active force is applied as a boundary condition to the voxel unit nodes in the contact area; Establish and solve the global stiffness matrix equation corresponding to the digital twin model. The equation is in the form that the global force vector is equal to the product of the global stiffness matrix and the global nodal displacement vector. The global stiffness matrix is ​​assembled from the element stiffness matrix based on the geometry of all voxel elements and their Young's modulus and Poisson's ratio parameters. The displacement vectors of all nodes are solved by using an iterative solver, including the conjugate gradient method or the multigrid method, to obtain the deformation field of the entire model. Based on nodal displacements, the strain tensor of each voxel element is calculated using shape function differentiation, and then the corresponding stress tensor is calculated according to the constitutive relation, i.e. Hooke's law, to obtain the internal stress distribution.

5. The virtual restoration simulation method for cultural relics based on digital twins according to claim 4, characterized in that, Updating the physical attribute data layer and geometric data layer of the digital twin model specifically includes: Establish a damage accumulation model, in which each voxel unit is associated with a damage state variable, with an initial value of 0; Within each physical simulation time step, if the equivalent stress of the voxel element is greater than the fracture toughness threshold, the damage state variables are accumulated according to the preset damage evolution rule. When the damage state variable reaches a preset critical value of 1, the Young's modulus parameter of the voxel unit is multiplied by a decay coefficient less than 1 to simulate the degradation of material stiffness. Meanwhile, in the visualization rendering module, the damage is visualized by adjusting the color and transparency or applying a preset crack texture map based on the damage state variable value of the voxel unit. When the damage state variable reaches 1, it indicates complete failure. The voxel element is removed from the global stiffness matrix of the finite element analysis and a geometric discontinuity surface representing crack propagation is generated in the geometric data layer, thereby simultaneously reflecting the macroscopic fracture of the material visually and mechanically.

6. The virtual restoration simulation method for cultural relics based on digital twins according to claim 5, characterized in that, The damage evolution rule is defined as follows: ； For the updated damage state variables, For the current damage state variable, For the simulation time step, For the von Mises equivalent effect, This represents the fracture toughness threshold.

7. The method for virtual restoration simulation of cultural relics based on digital twins according to claim 6, characterized in that, The material texture database stores the mapping relationship between different material types and vibration waveform parameters, which include fundamental frequency, harmonic components, amplitude envelope, and duration.

8. The method for virtual restoration simulation of cultural relics based on digital twins according to claim 7, characterized in that, The physical simulation engine is deployed on a heterogeneous computing platform where a central processing unit (CPU) and a graphics processing unit (GPU) work together, with the GPU used to accelerate the assembly and solution of the global stiffness matrix in parallel.

9. A virtual restoration simulation system for cultural relics based on digital twins, characterized in that, The virtual restoration simulation of cultural relics is implemented using the digital twin-based method for cultural relic virtual restoration as described in any one of claims 1 to 8. The system includes: The digital twin construction module for cultural relics is used to collect three-dimensional geometric data and material physical property data of cultural relics, and generate a three-dimensional digital twin model containing a geometric data layer and a physical property data layer. The physical property data layer is a finite element voxel mesh registered with the geometric data layer, and each voxel unit is assigned physical parameters including Young's modulus, Poisson's ratio, material density and fracture toughness threshold. The user interaction interface module, coupled with a multi-degree-of-freedom force feedback operation device, is used to acquire the position, posture and active force input applied by the operator to the virtual repair tool in real time, and output force and tactile feedback signals to the operator. The physical simulation engine module is configured to, when a collision is detected between the virtual repair tool and the digital twin model, calculate the internal stress and strain distribution and reaction force vector of the model using real-time finite element analysis based on the active force input and the physical parameters of the voxel elements in the contact area. The force feedback rendering module, which connects the physical simulation engine module and the user interaction interface module, is used to convert the reaction force vector into the target torque and high-frequency vibration control signal that drives the multi-degree-of-freedom force feedback operation device. The model state update module adjusts the physical parameters of the voxel units in the digital twin model that are subjected to excessive stress in real time according to the calculation results of the physical simulation engine module to simulate damage accumulation and fracture, and updates their geometric data to achieve damage visualization.

Images