Unified data model construction method and system for cae multi-physical field simulation

Abstract

This invention relates to the field of CAE multiphysics simulation, and discloses a unified data model construction method and system for CAE multiphysics simulation. The method includes: constructing a hierarchical data model structure, comprising a core entity layer, a physics adaptation layer, a coupling relationship layer, and a simulation process layer; creating solver adaptation entities and process node entities in the simulation process layer; defining mapping rules from model parameters to solver instructions through the solver adaptation entities; and defining pre-dependent nodes, post-dependent nodes, and data dependencies of simulation process nodes through the process node entities. This invention breaks down data silos in multiphysics, achieving a unified representation of all elements from geometry to mesh to material to boundary to coupling to results, with semantic consistency of multiphysics parameters ≥99.9% and accuracy of coupling relationship description ≥99%.

Description

Technical Field

[0001] This invention relates to the field of CAE multiphysics simulation, and in particular to a unified data model construction method and system for CAE multiphysics simulation. Background Technology

[0002] In the CAE multiphysics simulation process for complex products, the existing data management model has three major pain points that severely restrict simulation efficiency and accuracy: 1. Fragmentation of Physics Field Data Models: Multiphysics simulations across different disciplines rely on proprietary data models (e.g., ANSYS APDL for structural simulations, FLUENTCase for fluid simulations, and Maxwell Project for electromagnetic simulations), lacking a unified representation standard. For example, in motor simulations, structural fields require multiphysics simulations of "material elastic modulus and shaft stress boundaries," thermal fields require multiphysics simulations of "material thermal conductivity and winding temperature loads," and electromagnetic fields require multiphysics simulations of "material dielectric constant and coil current loads." These three types of data are stored in different file formats (.cdb / .cas / .mph), requiring manual mapping for data association. The mapping error rate exceeds that of multiphysics simulations by 15%, and data preparation before multiphysics coupling simulations takes 24 hours more than multiphysics simulations.

[0003] 2. Lack of semantic integrity: Existing data models are only designed for single physics fields and cannot describe multiphysics coupling relationships without loss. For example, in the fluid-solid coupling scenario of multiphysics simulation, the multiphysics simulation of "force transmission of fluid pressure on solid wall" needs to simultaneously characterize the multiphysics simulation of "pressure distribution" of the fluid field and the multiphysics simulation of "force boundary response" of the structural field. However, existing models can only store these two types of data separately. The multiphysics simulation of "physical quantity mapping rules (such as pressure → surface force conversion formula)" of the coupling interface needs to be manually defined by writing scripts. The coupling accuracy depends on the engineer's experience, and the simulation results deviate by more than 20% compared to multiphysics simulation.

[0004] 3. Data Disruptions in Simulation Process: In the multiphysics simulation process (CAD multiphysics simulation geometry import → multiphysics attribute assignment → multi-solver distribution → coupling result fusion), data needs to be repeatedly converted between different tools. Each conversion step requires redefining physical properties, making it difficult to guarantee data consistency. For example, in battery pack simulation, if the thermal field modifies the thermal conductivity of the cell material, the structural field needs to be manually updated to synchronize this parameter. Otherwise, it will lead to the multiphysics simulation error of "thermal stress calculation based on old parameters," resulting in a rework rate exceeding 30% for multiphysics simulations.

[0005] Existing technologies cannot meet the requirements of multiphysics simulation for "unified data, semantic integrity, and seamless process". There is an urgent need for a highly abstract, semantically lossless, and multi-field coupling unified data model to serve as the central data hub for multi-disciplinary multiphysics simulation CAE. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention provides a unified data model construction method for CAE multiphysics simulation, comprising the following steps: A hierarchical data model structure is constructed, which includes a core entity layer, a physical field adaptation layer, a coupling relationship layer, and a simulation process layer. In the core entity layer, a general core entity is created, including geometric entities, mesh entities, material entities, boundary condition entities, analysis step entities, and result entities. A globally unique identifier and SI basic attributes are configured for each of the general core entities. In the physical field adaptation layer, a physical field parameter container is attached to the general core entity, and multi-physics-specific parameters are stored in the physical field parameter container of the corresponding general core entity in the form of key-value pairs; In the coupling layer, coupling pair entities, mapping rule entities, and coupling boundary entities are created; the source physical field, target physical field, and coupling type are defined through the coupling pair entities; the source physical field parameters, target physical field parameters, transformation formulas, and spatial mapping methods are defined through the mapping rule entities; and the coupling boundary entities define a list of coupling boundary objects containing the globally unique identifier and the data synchronization frequency. In the simulation process layer, solver adapter entities and process node entities are created; the mapping rules from model parameters to solver instructions are defined through the solver adapter entities; and the pre-dependent nodes, post-dependent nodes, and data dependencies of the simulation process nodes are defined through the process node entities. The hierarchical data model structure is stored using a hybrid database architecture: the data of the core entity layer and the physical field adaptation layer are stored in a relational database, the data of the coupling relationship layer are stored in a graph database in the topological form of nodes and edges, and the data of the simulation process layer are stored in a document database. In response to the modification command for the multiphysics-specific parameters, the data in the relational database and the graph database are updated synchronously based on the cross-database transaction mechanism and the globally unique identifier.

[0007] This invention also provides a unified data model construction system for CAE multiphysics simulation based on the aforementioned method, comprising: The data model management module is used to construct a hierarchical data model structure, which includes a core entity layer, a physics adaptation layer, a coupling relationship layer, and a simulation process layer. In the core entity layer, general core entities including geometric entities, mesh entities, material entities, boundary condition entities, analysis step entities, and result entities are created. Each general core entity is configured with a globally unique identifier and SI basic attributes. A hybrid database architecture is used to store the hierarchical data model structure. The data of the core entity layer and the physics adaptation layer are stored in a relational database, the data of the coupling relationship layer is stored in a graph database in the topological form of nodes and edges, and the data of the simulation process layer is stored in a document database. In response to modification commands for multiphysics-specific parameters, the data in the relational database and the graph database are updated synchronously based on a cross-database transaction mechanism and the globally unique identifier. The multiphysics attribute assignment module is used to attach a physical field parameter container to the general core entity in the physical field adaptation layer, and store the multiphysics-specific parameters in the form of key-value pairs into the physical field parameter container of the corresponding general core entity. The coupling relationship definition module is used to create coupling pair entities, mapping rule entities, and coupling boundary entities in the coupling relationship layer; define the source physical field, target physical field, and coupling type through the coupling pair entities; define the source physical field parameters, target physical field parameters, transformation formulas, and spatial mapping methods through the mapping rule entities; and define a list of coupling boundary objects containing the globally unique identifier and the data synchronization frequency through the coupling boundary entities. The solver collaborative scheduling module is used to create solver adapter entities and process node entities in the simulation process layer; define the mapping rules from model parameters to solver instructions through the solver adapter entities; and define the pre-dependent nodes, post-dependent nodes, and data dependencies of the simulation process nodes through the process node entities.

[0008] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Data Unification and Semantic Integrity: Breaking down data silos in multiphysics, achieving unified representation of all elements including geometry, mesh, material, boundary, coupling, and results; semantic consistency of multiphysics parameters ≥99.9%; and accuracy of coupling relationship description ≥99%. 2. Improved efficiency of multi-solver collaboration: The preparation time for multiphysics simulation data is reduced by 90%, the error rate of data conversion between solvers is reduced from 15% to below 0.1%, and the automation rate of the simulation process is increased to 95%; 3. Guaranteed coupling accuracy: The conversion error of physical quantities in multi-physics coupling is ≤3%, which is 17 percentage points higher than the accuracy of traditional manual coupling simulation; 4. Scalability and reusability: Adding new physics fields only requires expanding the parameters of the physics field adaptation layer, without modifying the core structure, and the expansion cycle is ≤1 week; the material / boundary condition parameter template reuse rate is ≥70%, reducing the workload of repeated definition; 5. Full-process traceability: The data change log fully records the modification of multiphysics parameters and the adjustment of coupling rules, improving the efficiency of version backtracking by 90%, and the simulation results can be traced back to the original data source, meeting the needs of engineering auditing. Attached Figure Description

[0009] Figure 1 A flowchart provided for an embodiment of the present invention. Detailed Implementation

[0010] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” or “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0011] like Figure 1 A unified data model construction method for CAE multiphysics simulation includes the following steps: A hierarchical data model structure is constructed, which includes a core entity layer, a physical field adaptation layer, a coupling relationship layer, and a simulation process layer. In the core entity layer, a general core entity is created, including geometric entities, mesh entities, material entities, boundary condition entities, analysis step entities, and result entities. A globally unique identifier and SI basic attributes are configured for each of the general core entities. In the physical field adaptation layer, a physical field parameter container is attached to the general core entity, and multi-physics-specific parameters are stored in the physical field parameter container of the corresponding general core entity in the form of key-value pairs; In the coupling layer, coupling pair entities, mapping rule entities, and coupling boundary entities are created; the source physical field, target physical field, and coupling type are defined through the coupling pair entities; the source physical field parameters, target physical field parameters, transformation formulas, and spatial mapping methods are defined through the mapping rule entities; and the coupling boundary entities define a list of coupling boundary objects containing the globally unique identifier and the data synchronization frequency. In the simulation process layer, solver adapter entities and process node entities are created; the mapping rules from model parameters to solver instructions are defined through the solver adapter entities; and the pre-dependent nodes, post-dependent nodes, and data dependencies of the simulation process nodes are defined through the process node entities. The hierarchical data model structure is stored using a hybrid database architecture: the data of the core entity layer and the physical field adaptation layer are stored in a relational database, the data of the coupling relationship layer are stored in a graph database in the topological form of nodes and edges, and the data of the simulation process layer are stored in a document database. In response to the modification command for the multiphysics-specific parameters, the data in the relational database and the graph database are updated synchronously based on the cross-database transaction mechanism and the globally unique identifier.

[0012] The step of attaching a physics parameter container to the general core entity in the physics adaptation layer and storing multiphysics-specific parameters in key-value pairs into the physics parameter container of the corresponding general core entity includes: Establish a three-level semantic dictionary containing physical field categories, parameter names, and corresponding SI units; Extract the key names of the multiphysics-specific parameters to be input from the three-level semantic dictionary; The extracted key name is used as the key, and the received parameter value is used as the value to form the key-value pair, which is then written into the physical field parameter container attached to the material entity or the boundary condition entity.

[0013] The step of storing the data of the coupling relationship layer in the graph database in the topological form of nodes and edges includes: The source physical field and the target physical field defined in the coupling pair entity are respectively mapped to physical field nodes in the graph database; The mapping rule entity is mapped to a mapping rule edge connecting the two corresponding physical field nodes, and the source physical field parameters, the target physical field parameters, the transformation formula and the spatial mapping method are stored as attribute data in the mapping rule edge; Map the list of coupling boundary objects defined in the coupling boundary entity to attribute nodes, and attach the attribute nodes to the corresponding mapping rule edges.

[0014] The step of synchronously updating the data in the relational database and the graph database in response to the modification instruction of the multiphysics-specific parameters, based on the cross-database transaction mechanism and the globally unique identifier, includes: Obtain modification instructions for the multiphysics-specific parameters of the target general core entity in the relational database; Within the same cross-database transaction context, the modified values ​​of the multiphysics-specific parameters are overwritten into the corresponding physical parameter container in the relational database; Extract the globally unique identifier of the target general core entity, and retrieve the mapping rule edge associated with the attribute node corresponding to the globally unique identifier in the graph database; The modified values ​​of the multiphysics-specific parameters are substituted into the transformation formula on the retrieved mapping rule edge for update operation, and the cross-database transaction is submitted after the update operation is completed in both the relational database and the graph database.

[0015] The method further includes: The data dependencies of the process node entities in the document database are parsed, and the multiphysics solver process is triggered sequentially. During the execution of the multiphysics solver process, when the simulation time step reaches the data synchronization frequency defined in the coupled boundary entity, the result data output by the source physics solver is extracted; Based on the transformation formula and spatial mapping method recorded on the corresponding mapping rule edge in the graph database, the result data is converted into input load data for the target physics solver. The input load data is written into the input configuration file of the target physics solver by using the mapping rules defined in the solver adaptation entity.

[0016] Specifically, the core of this method is to achieve unified representation and lossless transfer of multiphysics data through "hierarchical structure design + generalized semantic definition + explicit coupling relationship". The specific steps are as follows: 1. Hierarchical structure design of the data model The system adopts a four-layer architecture: "core entity layer - physics adaptation layer - coupling relationship layer - simulation process layer". Each layer is connected through standardized interfaces to ensure a balance between "general abstraction" and "physics specificity". Core Entity Layer Define a general core entity for CAE simulation that is independent of physical fields.

[0017] • Geometry: geo_id (unique identifier), geo_type (B-Rep / Mesh / point cloud), coord_sys (Cartesian / cylindrical / spherical coordinates), unit (SI / imperial units); • Mesh entity: mesh_id, elem_type (SOLID / SHELL / BEAM), node_list (set of node IDs), intg_order (integral order); Material entity: mat_id, mat_name, base_param (density ρ / general property), physics_param (multiphysics parameter container); • Boundary Condition Entities: bc_id, target_obj_id (associated geometry / mesh ID), bc_category (constraint / load), physics_info (physics parameter container); • AnalysisStep Entity: step_id, step_type (static / transient / steady-state), time_param (time step / total duration), solver_ref (associated solver ID); • Result entity: result_id, source_step_id (associated analysis step), result_type (stress / temperature / velocity), data_format (scalar / vector / tensor).

[0018] Core attributes are independent of physical fields; for example, geo_id and mesh_id are globally unique identifiers that are applicable to all physical fields.

[0019] Physics adaptation layer Add multiphysics-specific parameters to the core entity; Structural field adaptation: Add E (elastic modulus), ν (Poisson's ratio), and σ_y (yield strength) to the material physics_param; add constraint_type (fixed / displaced) and force_magnitude (force magnitude) to the boundary conditions physics_info. • Thermal field adaptation: Add k (thermal conductivity), c_p (specific heat capacity), and α (coefficient of thermal expansion) to the material physics_param; add temperature (temperature value) and heat_flux (heat flux density) to the boundary conditions physics_info. • Fluid field adaptation: Add μ (viscosity), ρ (fluid density), and c (velocity of sound) to the material physics_param; add velocity (flow rate) and pressure (pressure) to the boundary conditions physics_info. Electromagnetic field adaptation: Add ε (dielectric constant), μ (permeability), and σ (conductivity) to the material physics_param; add voltage (voltage) and current (current) to the boundary conditions physics_info.

[0020] It supports 8+ physics fields including structural, thermal, fluid, electromagnetic, and acoustic. Adding a new physics field only requires expanding the adaptation parameters, without modifying the core entity.

[0021] Coupling layer Explicitly describe the multiphysics coupling logic and parameter mapping.

[0022] • Coupling Pair: couple_id, source_physics, target_physics, couple_type (one-way / two-way coupling); • Mapping Rule Entities: rule_id, source_param (source physics parameters, such as fluid pressure), target_param (target physics parameters, such as structural surface forces), conversion_formula (conversion formula, such as F=P×S, where S is the area of ​​action), spatial_mapping (spatial mapping method, such as node-node / face-face); • CouplingBoundary: cb_id, couple_ref_id (associated coupling pair), target_obj_list (list of coupling boundary object IDs), data_sync_freq (data synchronization frequency, such as every 10 time steps).

[0023] It supports 15+ coupling types, including fluid-solid, thermal-structural, and electromagnetic-thermal, and the coupling rules can be customized via formulas or scripts.

[0024] Simulation process layer Define the relationship and flow logic between the data model and multiple solvers.

[0025] SolverAdapter: solver_id, solver_name (e.g., ANSYS / FLUENT / Maxwell), data_mapping (model parameters → solver command mapping), input_output (solver input and output file association); ProcessNode: node_id, node_type (geometry import / mesh generation / solution / result fusion), prev_node (previous node ID), next_node (next node ID), data_dependency (data dependency); DataLog: log_id, operation_time (operation time), operator (operator), data_change (data modification record, e.g., "material 1 thermal conductivity changed from 300W / (m·K) to 320W / (m·K)").

[0026] It supports multi-solver collaborative process definition, and data flow is automatically triggered by log recording, allowing for traceability of data changes throughout the entire process.

[0027] 2. Generalized semantic definition and guarantee of physical field independence By "generalizing entity attributes + containerizing physical field parameters + standardizing semantic mapping", we ensure that the data model is decoupled from specific physical fields and solvers. • Generalization of entity attributes: The basic attributes of core entities (such as geo_id, mesh_id, mat_id) adopt globally unique identifiers (UUID), and the units are uniformly set to SI system (such as length in m, force in N, and temperature in K) to avoid semantic ambiguity caused by differences in units or identification rules; • Physics parameters are containerized: The physics-specific parameters of the core entity are stored in "parameter containers" (such as physics_param for materials and physics_info for boundary conditions). The containers adopt a key-value pair structure. When adding a new physics field, only the corresponding key needs to be added (such as "acoustic impedance Z" for acoustic fields), without modifying the entity structure. • Semantic mapping standardization: Define a three-level semantic dictionary of "physical field-parameter-unit", for example: Structural field - elastic modulus - a material's ability to resist elastic deformation - Pa-E_struct Thermal field - thermal conductivity - the ability of a material to transfer heat - W / (m·K) - k_thermal Fluid field - dynamic viscosity - fluid's ability to resist shear deformation - Pa·s - mu_fluid Electromagnetic field - relative permittivity - a measure of a material's polarizability in an electric field - dimensionless - eps_r_em 3. Explicitization and lossless description of coupling relationships The explicit storage and lossless transfer of multiphysics coupling relationships are achieved through a three-level structure of "coupled pairs - mapping rules - coupling boundaries". • Definition of coupling pairs: Clearly define the source / target physical fields and the coupling direction (e.g., unidirectional coupling of "thermal field → structural field", "bidirectional coupling of "flow field ↔ structural field"). • Quantification of mapping rules: The transformation logic of coupled physical quantities is stored in the form of "formula + parameters", such as the mapping rule of "fluid pressure → structural surface force" in fluid-structure interaction: source_param: pressure_fluid (fluid pressure, in Pa); target_param: force_struct (structural surface force, in N); conversion_formula: F = P × S (S is the area of ​​the coupling boundary, which is automatically extracted from the geometric entity); spatial_mapping: surface-to-surface mapping (topological matching between fluid computation surfaces and structural mesh surfaces); • Coupling Boundary Association: Specifies the geometry / mesh objects involved in the coupling and the data synchronization frequency.

[0028] 4. Data Model Storage Methods It adopts a hybrid storage architecture of "structured database + graph database" to balance the efficiency of entity attribute query and the ability to associate coupled relationships: • Core entity storage: The structured data of the core entity layer and the physical field adaptation layer are stored in a relational database and linked using a table structure; • Coupling relationship storage: Store the associated data of the coupling relationship layer in a graph database, and describe the coupling logic in the form of "node-edge"; • Simulation process storage: Store process nodes and solver adapter data in a document-based database, supporting efficient storage and retrieval of unstructured data; • Data consistency guarantee: A transaction mechanism is adopted to ensure cross-database data synchronization. For example, when the thermal conductivity of a material is modified, the physics_param in the relational database and the coupling rules in the graph database that depend on the material are updated simultaneously, with a data update delay of ≤100ms.

[0029] CAE Multiphysics Simulation System Based on Unified Data Model A unified data model construction system for CAE multiphysics simulation based on the aforementioned methods includes: The data model management module is used to construct a hierarchical data model structure, which includes a core entity layer, a physics adaptation layer, a coupling relationship layer, and a simulation process layer. In the core entity layer, general core entities including geometric entities, mesh entities, material entities, boundary condition entities, analysis step entities, and result entities are created. Each general core entity is configured with a globally unique identifier and SI basic attributes. A hybrid database architecture is used to store the hierarchical data model structure. The data of the core entity layer and the physics adaptation layer are stored in a relational database, the data of the coupling relationship layer is stored in a graph database in the topological form of nodes and edges, and the data of the simulation process layer is stored in a document database. In response to modification commands for multiphysics-specific parameters, the data in the relational database and the graph database are updated synchronously based on a cross-database transaction mechanism and the globally unique identifier. The multiphysics attribute assignment module is used to attach a physical field parameter container to the general core entity in the physical field adaptation layer, and store the multiphysics-specific parameters in the form of key-value pairs into the physical field parameter container of the corresponding general core entity. The coupling relationship definition module is used to create coupling pair entities, mapping rule entities, and coupling boundary entities in the coupling relationship layer; define the source physical field, target physical field, and coupling type through the coupling pair entities; define the source physical field parameters, target physical field parameters, transformation formulas, and spatial mapping methods through the mapping rule entities; and define a list of coupling boundary objects containing the globally unique identifier and the data synchronization frequency through the coupling boundary entities. The solver collaborative scheduling module is used to create solver adapter entities and process node entities in the simulation process layer; define the mapping rules from model parameters to solver instructions through the solver adapter entities; and define the pre-dependent nodes, post-dependent nodes, and data dependencies of the simulation process nodes through the process node entities.

[0030] The system also includes: The CAD geometry import and adaptation module is used to parse the input computer-aided design 3D model file, extract 3D geometric topology data and generate the corresponding geometric entity in the core entity layer, and generate the corresponding mesh entity in the core entity layer from the mesh data generated based on spatial subdivision. The simulation result fusion module is used to extract the calculation result files output by multiple solvers, match the extracted calculation result data to the corresponding general core entity based on the globally unique identifier, and write the calculation result data back to the result entity in the core entity layer.

[0031] The multiphysics attribute assignment module is configured to establish a three-level semantic dictionary containing physics field categories, parameter names, and corresponding SI units, extract the key names of the multiphysics-specific parameters to be input from the three-level semantic dictionary, combine the extracted key names as keys and the received parameter values ​​as values ​​to form key-value pairs, and write them into the physics parameter container attached to the material entity or the boundary condition entity.

[0032] The data model management module is configured to map the source physical field and the target physical field defined in the coupling pair entity to physical field nodes in the graph database, map the mapping rule entity to mapping rule edges connecting the two corresponding physical field nodes, and store the source physical field parameters, the target physical field parameters, the transformation formula and the spatial mapping method on the mapping rule edges; The solver collaborative scheduling module is configured to parse the process node entity, trigger the multiphysics solver process sequentially according to the data dependency relationship; when the data synchronization frequency is reached, it extracts the result data output by the source physics solver, substitutes the result data into the transformation formula and spatial mapping method recorded on the corresponding mapping rule edge in the graph database to convert it into the input load data of the target physics solver, and writes the input load data into the input configuration file of the target physics solver through the mapping rule defined in the solver adaptation entity.

[0033] The system also includes a data traceability and version management module; The simulation process layer also includes a data log entity; The data traceability and version management module is used to listen for modification operations on data in the hybrid database architecture, collect the timestamp of the operation, the operator's identifier, and the data difference record before and after the modification, encapsulate the collected information into the data log entity, and append the data log entity to the document database.

[0034] Specifically, the system is the hardware and software implementation of the aforementioned data model, with a unified data model as the central hub, connecting the entire CAE multiphysics simulation process, and includes 7 core functional modules: Data Model Management Module Java Spring Boot framework, PostgreSQL / Neo4j / MongoDB; enables unified data model creation, modification, querying, and deletion, and manages entity associations and coupling relationships; entity query response time ≤50ms; supports storage of 1 million+ entities; data update consistency ≥99.9%.

[0035] CAD geometry import adapter module OpenCASCADE 7.8 and STEP / IGES parsing library import CAD geometric models (STEP / IGES) into a unified data model, automatically extracts geometric attributes and generates geometry entities; importing a 100MB STEP model takes ≤3 minutes; geometric attribute extraction accuracy is ≥99%.

[0036] Multiphysics property assignment module Web front-end framework (Vue3), parameter editing component; add multi-physics parameters to core entities, support parameter template reuse; multi-physics parameter assignment efficiency is improved by 80% compared to manual; parameter template reuse rate ≥70%.

[0037] Coupling relationship definition module Neo4j graph visualization component and formula editor; visualizes the definition of multiphysics coupling pairs, mapping rules, and coupling boundaries, supports formula input and spatial mapping preview. The time taken to define coupling rules is ≤10 minutes per rule; the accuracy of mapping rule verification is ≥99%.

[0038] Solver Co-scheduling Module A distributed task scheduling framework and solver SDK (ANSYS / FLUENT / Maxwell); solver adaptation rules based on data models automatically generate solver input files, schedule multiple solvers for parallel computation, and synchronously couple data; supports 8+ mainstream solvers; solver instruction generation accuracy ≥98%; data synchronization latency ≤100ms.

[0039] Simulation Result Fusion Module Python Pandas / Numpy and ParaView visualization libraries; receive output results from multiple solvers and write them back to the result entity of a unified data model to achieve multiphysics result fusion and visualization; fusion of 10GB result data takes ≤30 minutes; result visualization response time ≤1 second / frame.

[0040] Data traceability and version management module Git version control and data log auditing components; record all changes to the data model, support version rollback and difference comparison; data change log integrity ≥ 100%; version rollback time ≤ 1 minute; difference comparison accuracy ≥ 99%.

[0041] Taking the thermal-electrical-structural multiphysics coupling simulation of a new energy vehicle battery pack as an example, this paper details the construction and application process of a unified data model: 1. Implementation Background and Requirements • Simulation objective: Analyze the multi-field coupled response of the battery pack during charging and discharging (electromagnetic heating → thermal conduction → structural thermal stress), which requires correlation of three types of physical fields: electromagnetic (current / voltage), thermal (temperature / heat flow), and structural (stress / deformation); Input data: Battery pack CAD model (STEP format, including cells, casing, and cooling pipes), material parameters (dielectric constant / thermal conductivity / elastic modulus of cells, thermal expansion coefficient of casing), boundary conditions (charge / discharge current 100A, ambient temperature 25℃, casing fixed constraints); • Solver configuration: Electromagnetic solver (Maxwell), thermal solver (ANSYSIcepak), structural solver (ANSYSMechanical), which needs to achieve unidirectional coupling of "electromagnetic → thermal → structural" (electromagnetic heating is used as the heat source of the thermal field, and the thermal field temperature is used as the thermal load of the structural field).

[0042] 2. Construction of a unified data model Step 1: Core Entity Creation (Data Model Management Module) 1. Geometric Entity: Import the STEP model, generate a geometry entity (geo_id=UUID-001, geo_type=B-Rep, unit=m), and automatically extract the battery cell (volume 0.001m³). 3 The geometric properties of the casing (thickness 0.005m) and cooling pipes (diameter 0.01m); 2. Mesh Entities: Divide the geometric entities into multiphysics meshes (electromagnetic mesh: tetra4 elements, thermal mesh: hexa8 elements, structural mesh: tetra10 elements), generating 3 mesh entities (mesh_id=UUID-101 / 102 / 103), and associate the geometric entities with geo_id; 3. Material Entities: Create two material entities (cell: mat_id=UUID-201; casing: mat_id=UUID-202), and uniformly set the density to base_param (cell 2500kg / m³). 3 Shell weight 7850 kg / m 3 ).

[0043] Step 2: Multiphysics property adaptation (property assignment module) Electromagnetic field properties: Add ε=10 (relative permittivity) and σ=5e5S / m (conductivity) to the physics_param of the cell material; create electromagnetic boundary conditions (bc_id=UUID-301, target_obj_id=UUID-101 (cell mesh), physics_info={current=100A, voltage=3.7V}). 2. Thermal Field Properties: Add k=150W / (m·K) (thermal conductivity) and c_p=1000J / (kg·K) (specific heat capacity) to the physics_param of the cell material; add k=45W / (m·K) and α=12e-6 / K (coefficient of thermal expansion) to the shell; create thermal boundary conditions (bc_id=UUID-302, target_obj_id=UUID-102, physics_info={ambient_temperature=298K,heat_flux=0W / m 2}); Structural field properties: Add E=206e9Pa (elastic modulus) and ν=0.3 (Poisson's ratio) to the physics_param of the shell material; create structural boundary conditions (bc_id=UUID-303, target_obj_id=UUID-103, physics_info={constraint_type=fixed,displacement=0m}).

[0044] Step 3: Define Coupling Relationships (Coupling Relationship Definition Module) 1. Electromagnetic-thermal coupling pair: Create couple_id=UUID-401, source_physics=electromagnetic, target_physics=thermal, couple_type=unidirectional; Mapping rules: rule_id=UUID-501, source_param=electromagnetic loss power P, target_param=heat source power Q, conversion_formula=Q=P (loss power is fully converted into heat), spatial_mapping=node-node; Coupling boundary: cb_id=UUID-601, target_obj_list=[UUID-101,UUID-102] (cell electromagnetic grid and thermal grid), data_sync_freq=1 charge / discharge cycle; 1. Thermal-structural coupling pair: Create couple_id=UUID-402, source_physics=thermal, target_physics=structural, couple_type=unidirectional; Mapping rules: rule_id=UUID-502, source_param=thermal temperature difference ΔT, target_param=structural thermal stress σ, conversion_formula=σ=α×E×ΔT (thermal stress formula), spatial_mapping=surface-to-surface; Coupled boundary: cb_id=UUID-602, target_obj_list=[UUID-102,UUID-103] (hot mesh and structured mesh), data_sync_freq=every 100s.

[0045] Step 4: Simulation process definition (solver cooperative scheduling module) 1. Solver adaptation: Create 3 SolverAdapters (Maxwell: solver_id=UUID-701, Icepak: UUID-702, Mechanical: UUID-703) and define parameter mappings (e.g., electromagnetic loss P → Icepak heat source Q, thermal temperature T → Mechanical thermal load). 2. Process Node: Define ProcessNode (Geometric Import → Mesh Generation → Electromagnetic Solution → Thermal Solution → Structural Solution → Result Fusion), and set prerequisite / post-delay dependencies (e.g., "Thermal Solution" is preceded by "Electromagnetic Solution", and loss power data needs to be synchronized).

[0046] 3. Simulation Execution and Data Flow 1. Electromagnetic solution: Maxwell is scheduled to read the UUID-101 mesh, UUID-201 material electromagnetic properties, and UUID-301 boundary conditions, calculate the charging and discharging electromagnetic loss power (P=500W), and write the results back to the result entity (result_id=UUID-801). 2. Thermal solution: The system automatically triggers electromagnetic-thermal coupling data synchronization, uses P=500W of UUID-801 as the heat source of Icepak, and calculates the highest temperature of the cell (T=353K) by combining the UUID-102 mesh, the thermal properties of UUID-201 / 202, and the boundary conditions of UUID-302. The result is written back to result_id=UUID-802. Structural solution: Synchronize the temperature data of UUID-802 (ΔT=353K-298K=55K), calculate the thermal stress (shell σ=12e-6×206e9×55≈135MPa) through mapping rules, and write the result back to result_id=UUID-803; 4. Results Fusion: The results fusion module associates the electromagnetic / thermal / structural results of UUID-801 / 802 / 803 to generate a coupled simulation report (including loss power distribution, temperature field cloud map, and thermal stress cloud map).

[0047] 4. Implementation effect verification • Data consistency: When the thermal conductivity of the battery cell is modified (from 150→160W / (m・K)), the system automatically synchronizes it to the thermal solver and structural solver without manual updates, achieving 100% data consistency. • Coupling accuracy: Electromagnetic-to-thermal power conversion error ≤1%, thermal-structural stress calculation error ≤3% (compared with manual coupling simulation); Efficiency improvements: Multiphysics data preparation time has been reduced from 24 hours to 2 hours, and the automation rate of the simulation process has increased from 30% to 95%.

[0048] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A unified data model construction method for CAE multiphysics simulation, characterized in that, Includes the following steps: A hierarchical data model structure is constructed, which includes a core entity layer, a physical field adaptation layer, a coupling relationship layer, and a simulation process layer. In the core entity layer, a general core entity is created, including geometric entities, mesh entities, material entities, boundary condition entities, analysis step entities, and result entities. A globally unique identifier and SI basic attributes are configured for each of the general core entities. In the physical field adaptation layer, a physical field parameter container is attached to the general core entity, and multi-physics specific parameters are stored in the physical field parameter container of the corresponding general core entity in the form of key-value pairs; In the coupling layer, coupling pair entities, mapping rule entities, and coupling boundary entities are created; the source physical field, target physical field, and coupling type are defined through the coupling pair entities; the source physical field parameters, target physical field parameters, transformation formulas, and spatial mapping methods are defined through the mapping rule entities; and the coupling boundary entities define a list of coupling boundary objects containing the globally unique identifier and the data synchronization frequency. In the simulation process layer, solver adapter entities and process node entities are created; the mapping rules from model parameters to solver instructions are defined through the solver adapter entities; and the pre-dependent nodes, post-dependent nodes, and data dependencies of the simulation process nodes are defined through the process node entities. The hierarchical data model structure is stored using a hybrid database architecture: the data of the core entity layer and the physical field adaptation layer are stored in a relational database, the data of the coupling relationship layer are stored in a graph database in the topological form of nodes and edges, and the data of the simulation process layer are stored in a document database. In response to the modification command for the multiphysics-specific parameters, the data in the relational database and the graph database are updated synchronously based on the cross-database transaction mechanism and the globally unique identifier.

2. The method according to claim 1, characterized in that, The step of attaching a physics parameter container to the general core entity in the physics adaptation layer and storing multiphysics-specific parameters in key-value pairs into the physics parameter container of the corresponding general core entity includes: Establish a three-level semantic dictionary containing physical field categories, parameter names, and corresponding SI units; Extract the key names of the multiphysics-specific parameters to be input from the three-level semantic dictionary; The extracted key name is used as the key, and the received parameter value is used as the value to form the key-value pair, which is then written into the physical field parameter container attached to the material entity or the boundary condition entity.

3. The method according to claim 1, characterized in that, The step of storing the data of the coupling relationship layer in the graph database in the topological form of nodes and edges includes: The source physical field and the target physical field defined in the coupling pair entity are respectively mapped to physical field nodes in the graph database; The mapping rule entity is mapped to a mapping rule edge connecting the two corresponding physical field nodes, and the source physical field parameters, the target physical field parameters, the transformation formula and the spatial mapping method are stored as attribute data in the mapping rule edge; Map the list of coupling boundary objects defined in the coupling boundary entity to attribute nodes, and attach the attribute nodes to the corresponding mapping rule edges.

4. The method according to claim 1, characterized in that, The step of synchronously updating the data in the relational database and the graph database in response to the modification instruction of the multiphysics-specific parameters, based on the cross-database transaction mechanism and the globally unique identifier, includes: Obtain modification instructions for the multiphysics-specific parameters of the target general core entity in the relational database; Within the same cross-database transaction context, the modified values ​​of the multiphysics-specific parameters are overwritten into the corresponding physical parameter container in the relational database; Extract the globally unique identifier of the target general core entity, and retrieve the mapping rule edge associated with the attribute node corresponding to the globally unique identifier in the graph database; The modified values ​​of the multiphysics-specific parameters are substituted into the transformation formula on the retrieved mapping rule edge for update operation, and the cross-database transaction is submitted after the update operation is completed in both the relational database and the graph database.

5. The method according to claim 1, characterized in that, Also includes: The data dependencies of the process node entities in the document database are parsed, and the multiphysics solver process is triggered sequentially. During the execution of the multiphysics solver process, when the simulation time step reaches the data synchronization frequency defined in the coupled boundary entity, the result data output by the source physics solver is extracted; Based on the transformation formula and spatial mapping method recorded on the corresponding mapping rule edge in the graph database, the result data is converted into input load data for the target physics solver. The input load data is written into the input configuration file of the target physics solver by using the mapping rules defined in the solver adaptation entity.

6. A unified data model construction system for CAE multiphysics simulation based on the method of any one of claims 1 to 5, characterized in that, include: The data model management module is used to construct a hierarchical data model structure, which includes a core entity layer, a physics adaptation layer, a coupling relationship layer, and a simulation process layer. In the core entity layer, a general core entity is created, including geometric entities, mesh entities, material entities, boundary condition entities, analysis step entities, and result entities. A globally unique identifier and SI basic attributes are configured for each of the general core entities. The hierarchical data model structure is stored using a hybrid database architecture. The core entity layer and the physical field adaptation layer are stored in a relational database, the data of the coupling relationship layer are stored in a graph database in the topological form of nodes and edges, and the data of the simulation process layer is stored in a document database. In response to the modification instructions for multiphysics-specific parameters, the data in the relational database and the graph database are updated synchronously based on the cross-database transaction mechanism and the globally unique identifier; The multiphysics attribute assignment module is used to attach a physical field parameter container to the general core entity in the physical field adaptation layer, and store the multiphysics-specific parameters in the form of key-value pairs into the physical field parameter container of the corresponding general core entity. The coupling relationship definition module is used to create coupling pair entities, mapping rule entities, and coupling boundary entities in the coupling relationship layer; define the source physical field, target physical field, and coupling type through the coupling pair entities; define the source physical field parameters, target physical field parameters, transformation formulas, and spatial mapping methods through the mapping rule entities; and define a list of coupling boundary objects containing the globally unique identifier and the data synchronization frequency through the coupling boundary entities. The solver collaborative scheduling module is used to create solver adapter entities and process node entities in the simulation process layer; define the mapping rules from model parameters to solver instructions through the solver adapter entities; and define the pre-dependent nodes, post-dependent nodes, and data dependencies of the simulation process nodes through the process node entities.

7. The system according to claim 6, characterized in that, Also includes: The CAD geometry import and adaptation module is used to parse the input computer-aided design 3D model file, extract 3D geometric topology data and generate the corresponding geometric entity in the core entity layer, and generate the corresponding mesh entity in the core entity layer from the mesh data generated based on spatial subdivision. The simulation result fusion module is used to extract the calculation result files output by multiple solvers, match the extracted calculation result data to the corresponding general core entity based on the globally unique identifier, and write the calculation result data back to the result entity in the core entity layer.

8. The system according to claim 6, characterized in that: The multiphysics attribute assignment module is configured to establish a three-level semantic dictionary containing physics field categories, parameter names, and corresponding SI units, extract the key names of the multiphysics-specific parameters to be input from the three-level semantic dictionary, combine the extracted key names as keys and the received parameter values ​​as values ​​to form key-value pairs, and write them into the physics parameter container attached to the material entity or the boundary condition entity.

9. The system according to claim 6, characterized in that: The data model management module is configured to map the source physical field and the target physical field defined in the coupling pair entity to physical field nodes in the graph database, map the mapping rule entity to mapping rule edges connecting the two corresponding physical field nodes, and store the source physical field parameters, the target physical field parameters, the transformation formula and the spatial mapping method on the mapping rule edges; The solver collaborative scheduling module is configured to parse the process node entity, trigger the multiphysics solver process sequentially according to the data dependency relationship; when the data synchronization frequency is reached, it extracts the result data output by the source physics solver, substitutes the result data into the transformation formula and spatial mapping method recorded on the corresponding mapping rule edge in the graph database to convert it into the input load data of the target physics solver, and writes the input load data into the input configuration file of the target physics solver through the mapping rule defined in the solver adaptation entity.

10. The system according to claim 6, characterized in that, The system also includes a data traceability and version management module; The simulation process layer also includes a data log entity; The data traceability and version management module is used to listen for modification operations on data in the hybrid database architecture, collect the timestamp of the operation, the operator's identifier, and the data difference record before and after the modification, encapsulate the collected information into the data log entity, and append the data log entity to the document database.

Images