A multi-stage rolling bearing thermal stability analysis method based on a thermal network model

Abstract

The application discloses a kind of based on multi-stage rolling bearing thermal stability analysis method of heat grid model, it is related to computer-aided engineering technical field, including, extracting analysis task parameters from the natural language task description input by user, generates heat grid analysis task plan;Call preset parameterized modeling library, execute heat grid analysis task plan, generate multi-stage rolling bearing three-dimensional geometric model;From temperature field, heat flow field and thermal deformation field, extract each stage bearing feature point temperature, maximum temperature, interstage heat flow data and thermal working clearance data, form thermal stability key index data;Thermal stability key index data is analyzed, and structured analysis report is generated.The application generates heat grid analysis task plan by receiving and analyzing natural language task description, realizes the conversion from user's fuzzy and unstructured engineering analysis demand to accurate and executable computer simulation instruction sequence.

Description

Technical Field

[0001] This invention relates to the field of computer-aided engineering technology, and in particular to a method for analyzing the thermal stability of multi-stage rolling bearings based on a thermal mesh model. Background Technology

[0002] Numerical simulation based on thermal mesh models is a core technique for analyzing the thermal stability of multi-stage rolling bearings. This method establishes a discretized thermal model of the bearing and solves the three-dimensional transient heat conduction equations, enabling the prediction of temperature field distribution, thermal deformation, and thermal equilibrium state, thereby assessing the risks of thermal instability such as seizure and lubrication failure. With the development of Computer-Aided Engineering (CAE) technology, parametric modeling, automated scripting, and computational resource scheduling techniques have been applied to improve the local efficiency of the simulation process.

[0003] However, existing analysis workflows heavily rely on manual operation by engineers and expert experience. From requirements understanding, model preprocessing, solution calculation to result postprocessing and report generation, almost every step requires manual intervention, using various specialized software and writing complex scripts. This strong reliance on manual intervention results in low efficiency and poor repeatability of the entire analysis workflow, and places extremely high demands on the professional background of the operators, severely restricting the application of this technology in scenarios such as rapid iteration of new products and online status assessment. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a method for analyzing the thermal stability of multi-stage rolling bearings based on a thermal mesh model, which solves the problems of low efficiency and poor consistency of analysis process caused by the reliance on manual operation in the analysis of the thermal stability of multi-stage rolling bearings.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a method for analyzing the thermal stability of multi-stage rolling bearings based on a thermal mesh model. The method includes: extracting analysis task parameters from a user-input natural language task description to generate a thermal mesh analysis task plan; calling a pre-set parametric modeling library to execute the thermal mesh analysis task plan and generate a three-dimensional geometric model of the multi-stage rolling bearing; discretizing the three-dimensional geometric model of the multi-stage rolling bearing to generate an initial thermal mesh model; performing quality checks and repairs on the initial thermal mesh model based on preset mesh quality rules to obtain a qualified thermal mesh model; assigning material properties, heat sources, heat transfer coefficients, and thermal constraints to the qualified thermal mesh model to construct a multi-stage bearing coupled thermal mesh solution model; submitting and running the multi-stage bearing coupled thermal mesh solution model, monitoring the solution process, diagnosing and adjusting solution anomalies, obtaining the temperature field and heat flow field; obtaining the thermal deformation field based on the temperature field, combined with material thermal expansion parameters and structural constraints; extracting characteristic point temperatures, maximum temperatures, interstage heat flow data, and hot-state working clearance data of each bearing stage from the temperature field, heat flow field, and thermal deformation field to form key thermal stability index data; and analyzing the key thermal stability index data to generate a structured analysis report.

[0007] As a preferred embodiment of the multi-stage rolling bearing thermal stability analysis method based on a thermal mesh model according to the present invention, the steps for generating the thermal mesh analysis task plan are as follows: Receive and normalize the natural language task description to obtain normalized text, and extract the bearing model, structural parameters, operating conditions and thermal stability analysis objectives from the normalized text; The bearing model, structural parameters, operating conditions, and thermal stability analysis objectives are matched and mapped with a predefined bearing parameter specification library to generate analysis task parameter objects. The task logic is parsed and the task sequence is arranged for the analysis task parameter object to generate a hot grid analysis task plan.

[0008] As a preferred embodiment of the multi-stage rolling bearing thermal stability analysis method based on the thermal mesh model described in this invention, the thermal mesh analysis task plan includes parametric geometric modeling instructions, mesh generation and quality inspection instructions, physical boundary condition setting instructions, solver submission and monitoring instructions, thermal deformation calculation instructions, and result data extraction instructions.

[0009] As a preferred embodiment of the multi-stage rolling bearing thermal stability analysis method based on a thermal mesh model according to the present invention, the steps for generating the three-dimensional geometric model of the multi-stage rolling bearing are as follows: Retrieve and instantiate feature templates corresponding to parametric geometry modeling instructions from the pre-set parametric modeling library to generate single-level parametric geometry for each bearing level; Based on the axial arrangement relationship and installation spacing defined in the parametric geometry modeling command, spatial transformation and Boolean union operations are performed on the single-level parametric geometry of each bearing to generate a multi-level bearing assembly; Geometric cleaning and repair of the multi-stage bearing assembly were performed to form a three-dimensional geometric model of the multi-stage rolling bearing.

[0010] As a preferred embodiment of the multi-stage rolling bearing thermal stability analysis method based on a thermal mesh model according to the present invention, the steps for generating the initial thermal mesh model are as follows: The mesh generation and quality inspection instructions are analyzed to identify the key geometric features of each bearing level from the three-dimensional geometric model of the multi-level rolling bearing, and to generate a set of key geometric features. Based on the key geometric feature set and the parameters in the mesh generation instruction, an adaptive mesh size field is generated that is refined in the expected high temperature gradient region and the heat flow concentration region. Based on the adaptive mesh size field, the three-dimensional geometric model of the multi-stage rolling bearing is divided into parallel volume meshes to generate an initial thermal mesh model.

[0011] As a preferred embodiment of the multi-stage rolling bearing thermal stability analysis method based on a thermal mesh model according to the present invention, the steps for obtaining a qualified thermal mesh model are as follows: Calculate the Jacobian ratio and aspect ratio of each grid in the initial thermal grid model; Based on the Jacobian ratio, aspect ratio, and preset mesh quality rules, node smoothing and local reconstruction operations are performed on the initial hot mesh model to generate the repaired mesh. The size deviation of the repaired mesh is corrected based on the adaptive mesh size field to obtain a qualified thermal mesh model.

[0012] As a preferred embodiment of the multi-stage rolling bearing thermal stability analysis method based on the thermal mesh model described in this invention, the steps for constructing the multi-stage bearing coupled thermal mesh solution model are as follows: Based on the qualified thermal grid model, operating conditions and physical boundary conditions, the heat source action area, heat conduction and transfer area and heat transfer boundary area are identified, and characteristic interaction relationships are constructed according to the thermal coupling relationship between each area. Based on the feature interaction relationship, material properties, heat source, heat transfer coefficient and thermal constraint conditions are assigned to qualified thermal grid models, and the heat source, heat transfer coefficient and thermal constraint conditions are adjusted in a coordinated manner. The adjusted heat source, heat transfer coefficient, and thermal constraint conditions are integrated into a qualified thermal mesh model to generate a multi-level bearing coupled thermal mesh solution model.

[0013] As a preferred embodiment of the multi-stage rolling bearing thermal stability analysis method based on the thermal mesh model described in this invention, the steps for obtaining the temperature field and heat flow field, and calculating the thermal deformation field based on the temperature field, combined with material thermal expansion parameters and structural constraints, are as follows: Start the thermal solver to load the multi-stage bearing coupled thermal mesh solution model, start the embedded computing agent in parallel, and establish real-time data monitoring; The embedded computing agent calculates a health status index that characterizes the convergence state of the thermal solver based on real-time monitoring data. When an abnormality in the health status index is diagnosed, the parameters of the thermal solver are adjusted. After the parameter adjustment is executed, the convergence of the thermal solver solution process is verified by real-time data monitoring. Once the thermal solver converges, the temperature field and heat flow field are extracted. Based on the thermal deformation calculation instructions, the temperature field is matched according to the mapping category corresponding to the material thermal expansion parameters and structural constraints in the preset temperature field-thermal deformation field mapping library, and the corresponding thermal deformation field is retrieved.

[0014] As a preferred embodiment of the multi-stage rolling bearing thermal stability analysis method based on a thermal mesh model described in this invention, the steps for generating key thermal stability index data are as follows: The analysis result data extraction command extracts characteristic point temperature, maximum temperature, interstage heat flow data and hot working clearance data of bearings at each stage from the temperature field, heat flow field and thermal deformation field; Based on the characteristic point temperature, maximum temperature, interstage heat flow data and hot working clearance data of each bearing stage, the comprehensive thermal stability risk index of each bearing stage is calculated. The characteristic point temperature, maximum temperature, interstage heat flow data, hot working clearance data of each bearing stage, and comprehensive thermal stability risk index are packaged into key thermal stability indicator data.

[0015] As a preferred embodiment of the multi-stage rolling bearing thermal stability analysis method based on the thermal mesh model described in this invention, the steps for generating the structured analysis report are as follows: The key thermal stability index data are analyzed and evaluated, and potential failure modes are identified based on predefined risk assessment rules, generating multidimensional risk characterization results. Based on operating conditions and key thermal stability data, the correlation between operating conditions and thermal stability is analyzed. A structured analysis report is generated based on potential failure modes, correlations, and multidimensional risk characterization results.

[0016] The beneficial effects of this invention are as follows: By receiving and parsing natural language task descriptions, a hot grid analysis task plan is generated, realizing the transformation from users' fuzzy and unstructured engineering analysis needs to precise and executable computer simulation instruction sequences; it deeply integrates natural language processing and automatic task planning, enabling it to understand complex engineering semantics, decompose task logic, and orchestrate a complete simulation workflow; it not only liberates engineers from tedious and error-prone software operations and script writing, but also lowers the technical threshold for using professional CAE tools, ensuring the standardization, automation, and efficient startup of the analysis process, laying a solid foundation for fully automated simulation execution and analysis. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a method for analyzing the thermal stability of multi-stage rolling bearings based on a thermal mesh model.

[0019] Figure 2 A flowchart for generating a task plan for natural language parsing and hot grid analysis.

[0020] Figure 3 A flowchart for constructing a multi-stage bearing coupled thermal mesh solution model.

[0021] Figure 4 A flowchart for generating a structured analysis report. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 This is one embodiment of the present invention, which provides a method for analyzing the thermal stability of multi-stage rolling bearings based on a thermal mesh model, including the following steps: S1. Extract analysis task parameters from the natural language task description input by the user and generate a hot grid analysis task plan.

[0026] S1.1: Receive and standardize the natural language task description to obtain standardized text, and extract the bearing model, structural parameters, operating conditions and thermal stability analysis targets from the standardized text; Specifically, after receiving the natural language task description, the system performs operations such as removing extra spaces, standardizing punctuation, and correcting spelling errors, and converting full-width characters to half-width characters to complete text cleaning. The cleaned text is then reorganized according to a preset sentence structure, and standardized terminology is used to form a standardized text. Based on the standardized text, keyword matching and semantic recognition are used to extract bearing models, structural parameters, operating conditions, and thermal stability analysis targets.

[0027] It should be noted that the pre-set sentence structure refers to the uniform reorganization of the cleaned text into a fixed word order format of "bearing model + structural parameters + operating conditions + thermal stability analysis target"; The bearing model refers to the specific bearing specification name obtained by identifying and matching it with the standard bearing model library from the specification text; Structural parameters refer to numerical descriptions of geometric dimensions such as the number of rollers, contact angle, and clearance extracted from the specification text; Operating conditions refer to numerical or range descriptions of operating states such as speed, load, and ambient temperature captured from the specification text; The objectives of thermal stability analysis refer to the specific requirements and statements regarding temperature field distribution, thermal deformation, or stability criteria extracted from the standard text.

[0028] S1.2: Match and map the bearing model, structural parameters, operating conditions, and thermal stability analysis objectives with the predefined bearing parameter specification library to generate analysis task parameter objects; Specifically, the bearing model, structural parameters, operating conditions, and thermal stability analysis objectives are sequentially compared with standard entries in a predefined bearing parameter specification library. The bearing model is locked in the predefined bearing parameter specification library through complete character matching or minimum difference correction. Structural parameters are described based on geometric dimensions such as roller count, contact angle, and clearance, and corresponding numerical ranges are found in the predefined bearing parameter specification library to establish mapping relationships. Operating conditions are described using operating states such as speed, load, and ambient temperature, and are associated and matched with typical operating condition records in the predefined bearing parameter specification library. Thermal stability analysis objectives are located in the predefined bearing parameter specification library based on specific requirements of temperature field distribution, thermal deformation, or stability criteria. After matching is completed, the locked specific specification name, established mapping relationship, associated typical operating condition records, and located analysis indicator definitions are integrated to generate the analysis task parameter object.

[0029] It should be noted that the predefined bearing parameter specification library refers to a complete database that stores all standard bearing specification names, geometric dimension ranges, typical operating condition records, and analysis index definitions.

[0030] S1.3: Perform task logic parsing and task sequence arrangement on the analysis task parameter object to generate a hot grid analysis task plan.

[0031] Specifically, the specific specification names, mapping relationships, typical working condition records, and analysis index definitions contained in the analysis task parameter objects are logically sorted out to determine the execution order required for the thermal grid analysis task plan. Based on the specific specification names, the geometric modeling steps are determined, the meshing strategy is established using the mapping relationships, the boundary condition application process is set in combination with the typical working condition records, and the result extraction process is planned with reference to the analysis index definitions. The geometric modeling steps, meshing strategy, boundary condition application process, and result extraction process are arranged in chronological order to form the thermal grid analysis task plan.

[0032] It should be noted that the geometric modeling step refers to the process of generating bearing entity geometric data based on specific specification names; the mesh generation strategy refers to the method of converting bearing entity geometric data into a discrete model using mapping relationships; the boundary condition application process refers to the operation of setting operating state constraints on the discrete model in combination with typical working condition records; and the result extraction step refers to obtaining data such as temperature field distribution or thermal deformation from the completed state by referring to the definition of analysis indicators.

[0033] S1.4: The thermal mesh analysis task plan includes parametric geometric modeling instructions, mesh generation and quality check instructions, physical boundary condition setting instructions, solver submission and monitoring instructions, thermal deformation calculation instructions, and result data extraction instructions.

[0034] It should be noted that parametric geometric modeling commands refer to control commands that automatically generate three-dimensional solid geometric data of the bearing based on the bearing model and structural parameters. Mesh generation and quality inspection commands refer to control commands that discretize the generated 3D solid geometric data and verify whether the mesh quality meets the requirements. Physical boundary condition setting commands refer to control commands that apply rotational speed, load, and temperature constraints to the discretized mesh model based on the operating conditions. Solver submission and monitoring commands refer to the control commands that send the constrained mesh model to the computational core for calculation and track the calculation status in real time; The thermal deformation calculation command is a control command that derives the thermal displacement of various parts of the bearing based on the temperature field distribution data during the calculation process. The result data extraction command is a control command that filters out values ​​that meet the objectives of thermal stability analysis from the completed thermal displacement and temperature field distribution data.

[0035] S2. Call the preset parametric modeling library, execute the thermal mesh analysis task plan, and generate a three-dimensional geometric model of a multi-stage rolling bearing; discretize the three-dimensional geometric model of the multi-stage rolling bearing to generate an initial thermal mesh model.

[0036] S2.1: Retrieve and instantiate the feature template corresponding to the parametric geometry modeling command from the preset parametric modeling library to generate single-level parametric geometry of each bearing level; Specifically, the system retrieves matching feature templates from a pre-built parametric modeling library based on the bearing model, and matches the specific specification names contained in the parametric geometric modeling instructions with the parameter placeholders in the feature templates. The system then replaces the parameter placeholders in the feature templates with the dimensional values ​​determined by the specific specification names, generating a single-level parametric geometric data structure containing complete geometric information. The single-level parametric geometry of each bearing level is composed of multiple single-level parametric geometric data structures, and the combination process determines the relative positions of each single-level parametric geometric data structure based on the bearing hierarchy.

[0037] S2.2: Based on the axial arrangement relationship and installation spacing defined in the parametric geometry modeling command, perform spatial transformation and Boolean union operation on the single-level parametric geometry of each bearing to generate a multi-level bearing assembly; Specifically, the relative position order of the single-level parametric geometry of each bearing level in space is determined according to the axial arrangement relationship defined in the parametric geometry modeling command. The single-level parametric geometry of each bearing level is translated using the installation spacing value to adjust the distance between adjacent components. After the position adjustment is completed, the single-level parametric geometry of each bearing level is directly used as the operation object of Boolean union operation. Through Boolean union operation, multiple independent single-level parametric geometries are merged into a multi-level bearing assembly.

[0038] S2.3: Perform geometric cleaning and repair on the multi-stage bearing assembly to form a three-dimensional geometric model of the multi-stage rolling bearing; Specifically, a geometric cleanup operation is performed on the multi-stage bearing assembly to remove the tiny fragments and redundant vertices remaining after Boolean union operations, ensuring the integrity of the surface topology. For gaps or overlapping defects in the contact area of ​​adjacent components, geometric repair is used to fill the gaps and correct the interference, making the multi-stage bearing assembly a closed and continuous entity. The repaired entity is directly used as the three-dimensional geometric model of the multi-stage rolling bearing.

[0039] S2.4: Parse the mesh generation and quality inspection instructions to identify the key geometric features of each bearing level from the three-dimensional geometric model of the multi-level rolling bearing and generate a set of key geometric features; Specifically, the process involves reading mesh generation and quality inspection instructions, extracting keywords describing geometric features related to contact areas, chamfers, and fillets, and using these keywords to perform matching searches on the surface of the multi-stage rolling bearing's 3D geometric model. This identifies specific locations such as the contact surfaces between rollers and raceways, the edges of cage pockets, and sealing lips. Successfully matched surface areas are marked as key geometric feature regions, and the rate of curvature change and adjacent edge topological relationships of these regions are recorded. Based on these records, high stress concentration areas and non-critical transition areas are distinguished, determining the local ranges in each bearing stage that require mesh refinement. For the identified key geometric feature regions, an independent set of key geometric features is generated.

[0040] S2.5: Based on the key geometric feature set and the parameters in the mesh generation instruction, generate an adaptive mesh size field that is refined in the expected high temperature gradient region and heat flow concentration region; Specifically, the key geometric feature set and parameters in the mesh generation command are read, and the positioning markers of the expected high temperature gradient region and heat flow concentration region are extracted. Based on the positioning markers, the spatial range that needs to be refined is delineated on the surface of the multi-stage rolling bearing 3D geometric model. The delineated range is used as the reference region for the generation of the size field. Combining the minimum and maximum mesh sizes set in the mesh generation command, the minimum mesh size value is assigned in the reference region, and in the transition zone far from the reference region, it is linearly increased to the maximum mesh size value according to the distance ratio, forming a smooth size gradient distribution. The assigned mesh size values ​​are mapped to each vertex and edge of the multi-stage rolling bearing 3D geometric model to construct an adaptive mesh size field covering the entire geometry.

[0041] S2.6: Based on the adaptive mesh size field, perform parallel volume mesh generation on the three-dimensional geometric model of the multi-stage rolling bearing to generate the initial thermal mesh model; Specifically, the adaptive mesh size field is read to obtain the mesh size values ​​corresponding to each vertex and edge of the multi-stage rolling bearing 3D geometric model. Based on the mesh size values, the multi-stage rolling bearing 3D geometric model is divided into multiple independent subdomains to ensure that the mesh size values ​​in each subdomain change smoothly. An independent computation process is allocated to each independent subdomain. A parallel algorithm is used to fill tetrahedral meshes in all independent subdomains simultaneously to generate local mesh fragments. All local mesh fragments are matched and merged at the boundaries of the subdomains to eliminate overlapping parts and form a continuous overall mesh structure. The continuous overall mesh structure is assigned a thermal conductivity attribute label to form the initial thermal mesh model.

[0042] It should be noted that a parallel algorithm is a technique that decomposes a complex task into multiple subtasks and uses multiple processors or computing cores to execute multiple subtasks simultaneously. By dividing a large-scale problem into mutually independent or loosely coupled parts, parallel algorithms can shorten the overall running time and improve computational efficiency.

[0043] S3. Based on the preset mesh quality rules, the initial thermal mesh model is checked and repaired to obtain a qualified thermal mesh model. Material properties, heat sources, heat transfer coefficients and thermal constraints are assigned to the qualified thermal mesh model to construct a multi-level bearing coupled thermal mesh solution model.

[0044] S3.1: Calculate the Jacobian ratio and aspect ratio of each grid in the initial thermal grid model; The Jacobian ratio of each grid in the initial thermal grid model is calculated using the following expression: ; In the formula, Indicates the first Jacobian ratio of each grid; Indicates a grid index; Indicates the index of the Gaussian integral point; Let represent the set of Gaussian integration points, where The total number of integration points is preset according to the mesh type (e.g., 4 for tetrahedrons and 8 for hexahedrons); This indicates traversing all cells within the grid. For each integration point, select the one with the largest absolute value of the Jacobian determinant; This indicates traversing all cells within the grid. For each integration point, select the one with the smallest absolute value of the Jacobian determinant; This represents the determinant value of the Jacobian matrix, i.e., the nth matrix. The grid in the first Scaling ratio of the local volume (or area) at each Gaussian integration point; Indicates the first The grid in the first The Jacobian matrix at each Gaussian integration point, which is derived from the objective physical coordinates of the grid. It is calculated by combining the shape function derivative mapping of the corresponding mesh type.

[0045] The aspect ratio of each grid in the initial thermal mesh model is calculated using the following expression: ; In the formula, Indicates the first The aspect ratio of each grid; Indicates the first The radius of the circumscribed sphere of a grid is obtained by calculating the radius of the smallest sphere that encloses all nodes of that grid. Indicates the first The radius of the inscribed sphere of a grid is obtained by calculating the perpendicular distance from the geometric center of the grid to all its constituent surfaces and taking the minimum value among them. For severely distorted grids, an iterative method is used to find the maximum radius of the inscribed sphere.

[0046] S3.2: Based on the Jacobian ratio, aspect ratio, and preset mesh quality rules, perform node smoothing and local reconstruction operations on the initial hot mesh model to generate the repaired mesh; Specifically, based on the Jacobian ratio and aspect ratio of each mesh in the initial hot mesh model, meshes whose Jacobian ratio or aspect ratio does not meet the preset mesh quality rules are selected as meshes to be repaired. For the meshes to be repaired, the node coordinates of the meshes to be repaired are extracted, and a node smoothing operation is performed. By adjusting the node coordinate positions, the Jacobian ratio of the meshes to be repaired is reduced and the aspect ratio of the meshes to be repaired is optimized. If the Jacobian ratio or aspect ratio of the meshes to be repaired still does not meet the preset mesh quality rules after the node smoothing operation, a local reconstruction operation is performed on the meshes to be repaired. The original topological connections of the meshes to be repaired are deleted, and a new mesh is regenerated based on the nodes around the meshes to be repaired to replace the meshes to be repaired. The node smoothing operation and the local reconstruction operation are repeated until the Jacobian ratio and aspect ratio of all meshes in the initial hot mesh model meet the preset mesh quality rules, and the repaired mesh is generated.

[0047] It should be noted that the mesh quality rule refers to the upper limit standard that the Jacobian ratio and aspect ratio of each mesh in the initial thermal mesh model must meet. This is set based on the convergence requirements of finite element calculations and the accuracy requirements of thermal analysis. The setting steps include: determining the tolerance limit of the thermal analysis task for mesh distortion; referring to the quality inspection standards of industry-standard finite element software; and determining specific thresholds based on the characteristics of the mesh type (such as tetrahedron or hexahedron) in the initial thermal mesh model. An exemplary range is a Jacobian ratio less than 10 and an aspect ratio less than 5. The basis for this value is that when the Jacobian ratio exceeds 10, the change in the determinant of the Jacobian matrix at the integration point is too large, leading to an increase in numerical integration error; when the aspect ratio exceeds 5, the difference between the circumscribed sphere radius and the inscribed sphere radius is too large, indicating that the mesh shape is extremely stretched, which will reduce the accuracy of heat conduction calculations and affect the solver's convergence speed.

[0048] S3.3: Based on the adaptive mesh size field, the size deviation of the repaired mesh is corrected to obtain a qualified thermal mesh model; Specifically, the geometric center coordinates of each grid in the repaired mesh are extracted and mapped to the adaptive mesh size field to obtain the corresponding target size value. The circumscribed sphere radius of each grid in the repaired mesh is measured as the current size value. The current size value is compared with the target size value. If the current size value is greater than the target size value, a mesh refinement operation is performed on the corresponding grid in the repaired mesh, generating more nodes by splitting the edges to reduce the circumscribed sphere radius. If the current size value is less than the target size value, a mesh coarsening operation is performed on the corresponding grid in the repaired mesh, reducing the number of grids by merging adjacent nodes to increase the circumscribed sphere radius. The size comparison, mesh refinement, and mesh coarsening operations are repeated until the circumscribed sphere radius of all grids in the repaired mesh is consistent with the target size value at the corresponding position in the adaptive mesh size field, thus obtaining a qualified thermal mesh model.

[0049] S3.4: Based on the qualified thermal grid model, operating conditions and physical boundary conditions, the instructions identify the heat source action area, heat conduction and transfer area and heat exchange boundary area, and construct characteristic interaction relationships according to the thermal coupling relationship between each area; Specifically, the description of the energy application location in the operating conditions is read, and the set of grids that bear the energy input in the qualified thermal grid model is marked as the heat source action area. According to the definition of heat flow path in the physical boundary condition setting instructions, the set of grids that connect the heat source action area and are responsible for energy conduction is identified and marked as the heat conduction transfer area. According to the settings of surface-environment interaction in the physical boundary condition setting instructions, the set of grid surfaces on the outer surface of the qualified thermal grid model that exchange energy is extracted and marked as the heat transfer boundary area. Combining the spatial adjacency relationship of the heat source action area, the heat conduction transfer area, and the heat transfer boundary area, the connection relationship of the heat source action area to the heat transfer area is established, and the connection relationship of the heat conduction transfer area to the heat transfer boundary area is established, forming a characteristic interaction relationship.

[0050] S3.5: Based on the feature interaction relationship, material properties, heat source, heat transfer coefficient and thermal constraint conditions are assigned to qualified thermal grid models, and the heat source, heat transfer coefficient and thermal constraint conditions are adjusted in a coordinated manner. Specifically, based on the feature interaction relationship, material properties are assigned to the mesh corresponding to the heat transfer region in the qualified thermal mesh model. According to the location of the heat source action region marked in the feature interaction relationship, the heat source is applied to the corresponding mesh. According to the range of the heat transfer boundary region marked in the feature interaction relationship, the heat transfer coefficient is assigned to the corresponding mesh surface. According to the physical constraint requirements, thermal constraint conditions are applied to the surface of the qualified thermal mesh model at a specified fixed temperature. If the temperature of the heat source action region is too high, the heat source intensity is reduced and the heat transfer coefficient of the heat transfer boundary region is increased. If the temperature gradient of the heat transfer region is abnormal, the thermal conductivity parameters in the material properties are corrected and the location of the thermal constraint conditions is reassigned. The application and adjustment are repeated until the spatial distribution of the heat source, heat transfer coefficient and thermal constraint conditions completely matches the feature interaction relationship.

[0051] It should be noted that material properties refer to the set of physical parameters that describe the heat conduction capacity of the heat transfer region in a qualified thermal mesh model; A heat source refers to the external excitation condition that injects energy into a qualified thermal grid model within the area of ​​the heat source's influence. The heat transfer coefficient is a proportional constant that characterizes the efficiency of heat exchange between the heat transfer boundary region and the surrounding environment. Thermal constraints refer to the boundary settings that restrict the temperature of a specific surface in a qualified thermal mesh model to remain constant.

[0052] S3.6: Integrate the coordinated heat source, heat transfer coefficient and thermal constraint conditions into the qualified thermal mesh model to generate a multi-level bearing coupled thermal mesh solution model; Specifically, the coordinated heat source is mapped to the grid corresponding to the heat source's action area in the qualified thermal grid model. The coordinated heat transfer coefficient is assigned to the grid surface corresponding to the heat transfer boundary region in the qualified thermal grid model. The coordinated thermal constraint conditions are applied to the grid surface with a fixed specified temperature in the qualified thermal grid model. After completing the attribute assignment, based on the physical connection relationship between the components in the multi-level bearing structure, the temperature continuity and heat flow balance connection of the adjacent component grids at the contact surface are established. By connecting the temperature continuity and heat flow balance connection of each level of bearing assembly in series, a global heat conduction network covering the entire bearing assembly is formed. The complete data structure containing material properties, heat source, heat transfer coefficient, thermal constraint conditions and global heat conduction network is encapsulated to generate a multi-level bearing coupled thermal grid solution model.

[0053] S4. Submit and run the multi-level bearing coupled thermal mesh solution model, monitor the solution process, diagnose and adjust solution anomalies, obtain the temperature field and heat flow field, and obtain the thermal deformation field based on the temperature field, combined with the material thermal expansion parameters and structural constraints.

[0054] S4.1: Start the thermal solver to load the multi-stage bearing coupled thermal mesh solution model, start the embedded computing agent in parallel, and establish real-time data monitoring; Specifically, a loading command is sent to the thermal solver to load the multi-stage bearing coupled thermal mesh solution model into memory. The embedded computing agent is then started in parallel, establishing a communication link between the embedded computing agent and the thermal solver. A real-time data monitoring channel is created through the embedded computing agent, which directly connects to the heat source action area, heat conduction transfer area, and heat exchange boundary area in the multi-stage bearing coupled thermal mesh solution model. The real-time data monitoring channel continuously collects the energy injection status of the heat source action area, the temperature distribution status of the heat conduction transfer area, and the energy exchange status of the heat exchange boundary area. The embedded computing agent converts the collected status data into visual monitoring signals and displays the running progress of the multi-stage bearing coupled thermal mesh solution model on the monitoring interface.

[0055] It should be noted that the embedded computing agent is a program entity that starts in parallel and establishes a communication link with the thermal solver to create a real-time data monitoring channel to collect state data of the heat source action area, heat conduction transfer area and heat transfer boundary area in the multi-level bearing coupled thermal mesh solution model.

[0056] S4.2: The embedded computing agent calculates a health status index that characterizes the convergence state of the thermal solver based on real-time monitoring data. When an abnormality in the health status index is diagnosed, the parameters of the thermal solver are adjusted. Specifically, when the health status index fails to reach the preset convergence threshold within a preset number of iterations (e.g., 300 to 400 steps), or when the health status index shows a continuous decline, the thermal solver is determined to be in an abnormal state. The embedded computing agent then sends an adjustment command to the thermal solver to reduce the energy injection intensity in the heat source action area and increase the energy exchange efficiency in the heat exchange boundary area.

[0057] Specifically, the steps for setting the convergence threshold are as follows: collect data from the entire normal convergence process multiple times (e.g., 50 times) under the same operating conditions, and extract the time when the convergence standard is reached each time. The values ​​form a sample set, and the mean and standard deviation of the sample set are calculated. Finally, the difference between the mean and three times the standard deviation is taken as the convergence threshold. The exemplary value range is usually 80~120. The basis for the value is to ensure that the threshold is located in the safe isolation zone between the "convergence stagnation zone" and the "normal convergence completion zone". This can avoid misjudging slow convergence as a fault and ensure that an alarm is triggered immediately when the solver really diverges or stagnates.

[0058] The health state index, which characterizes the convergence state of the thermal solver, is calculated using the following expression: ; In the formula, The health status index represents the convergence state of the thermal solver. The baseline value is 1.0 (representing the initial state). The larger the value, the higher the degree of convergence of the solver and the healthier the state. Indicates the sensitivity amplification factor; This represents the initial composite residual vector, which is formed by normalizing and splicing together the energy injection residual in the heat source region, the temperature distribution residual in the heat conduction region, and the energy exchange residual in the heat transfer boundary region during the 0th iteration of the thermal solver. This represents the current composite residual vector, which is formed by splicing together the energy injection residual in the heat source action area, the temperature distribution residual in the heat conduction area, and the energy exchange residual in the heat transfer boundary area at the current real-time monitoring moment by the heat solver after the same normalization process. Represents the L2 norm; It represents a very small positive real number.

[0059] It should be noted that the sensitivity amplification factor is set based on the statistical characteristics of the "residual decline rate in the later stage of convergence" in historical normal convergence curves. An exemplary value of 5 is used, and the basis for this value is to ensure that the health index is maintained when the relative convergence progress reaches the critical engineering interval of 90% to 99%. It can exhibit a nonlinear leap, thus numerically distinguishing between the "near convergence" and "fully converged" states. At this point... The value will rapidly increase from the initial 1.0 to over 100, therefore the judgment logic is revised to monitor... Whether the value crosses the convergence threshold within a limited number of steps, rather than a real-time comparison. The value and the threshold.

[0060] S4.3: After the parameter adjustment is executed, continue to verify the convergence of the thermal solver solution process through real-time data monitoring. After the thermal solver converges, extract the temperature field and heat flow field. Specifically, after the parameter adjustment is executed, the current comprehensive residual vector during the thermal solver iteration process is continuously collected, and the health status index of the thermal solver convergence is calculated. When the health status index exceeds the preset stability threshold, the thermal solver is determined to have completed convergence. After the convergence determination is completed, the temperature field and heat flow field are extracted from the thermal solver. The temperature field covers the entire heat source action area and heat conduction transfer area, and the heat flow field includes energy exchange data of all heat transfer boundary areas.

[0061] S4.4: Based on the thermal deformation calculation command, in the preset temperature field-thermal deformation field mapping library, according to the mapping category corresponding to the material thermal expansion parameters and structural constraints, the temperature field is characterized and the corresponding thermal deformation field is retrieved. Specifically, based on the thermal deformation calculation instructions, the material thermal expansion parameters (such as the linear expansion coefficient) and structural constraints (such as the bearing housing fixing method and shaft support boundary) under the current working conditions are extracted. These two types of parameters are combined as a primary index key to locate a specific mapping category sub-library in the pre-set temperature field-thermal deformation field mapping library. Spatial features of the temperature field are extracted (including the location of the highest temperature point, the distribution pattern of the temperature gradient, and the coordinates of the thermal center). A feature vector containing the normalized temperature distribution vector, the principal direction of the thermal gradient tensor, and the coordinates of the thermal center is constructed. A hybrid feature matching algorithm based on weighted Euclidean distance and cosine similarity is executed in the determined mapping category sub-library to find the pre-stored template with the highest feature similarity to the current temperature field. After a successful match, the thermal deformation field associated with the template is directly retrieved.

[0062] It should be noted that the temperature field-thermal deformation field mapping library is a database pre-established through high-precision offline simulation. It stores standard thermal deformation fields corresponding to various typical temperature field distributions under different material thermal expansion parameters and different structural constraints. The mapping rule is defined as follows: hard classification and screening are performed based on material properties and boundary conditions, and soft matching is performed in the sub-class library based on the temperature field topological features. The feature matching algorithm refers to the process of comprehensively determining whether the real-time temperature field feature vector and the template feature vector in the mapping library belong to the same physical response mode by calculating the weighted Euclidean distance (used to measure the difference in amplitude) and cosine similarity (used to measure the difference in distribution pattern) between them. The weight coefficient of the temperature gradient distribution pattern is set to 0.6, the weight of the highest temperature point position is 0.3, and the weight of the thermal center coordinates is 0.1 to ensure that the matching result reflects the driving mechanism of thermal deformation first rather than local noise.

[0063] S5. Extract the characteristic point temperature, maximum temperature, interstage heat flow data and hot working clearance data of each bearing from the temperature field, heat flow field and thermal deformation field to form key thermal stability index data; analyze the key thermal stability index data to generate a structured analysis report.

[0064] S5.1: Data extraction command for analysis results, extracts the characteristic point temperature, maximum temperature, interstage heat flow data and hot working clearance data of each bearing from the temperature field, heat flow field and thermal deformation field; Specifically, the system identifies the range of results from the multi-level bearing coupled thermal mesh solution model specified in the data extraction command, confirms the four types of targets required by the command: bearing characteristic point temperature, maximum temperature, interstage heat flow data, and hot working clearance data. It locates the temperature field associated with the command, extracts the temperature values ​​of corresponding coordinate points from the heat source action area and heat transfer area covered by the temperature field as the bearing characteristic point temperature for each level, and scans the entire temperature field data to lock the maximum temperature. It also locates the heat flow field associated with the command, filters the energy exchange values ​​between adjacent bearing stage surfaces from the energy exchange data of the heat transfer boundary region included in the heat flow field as interstage heat flow data, and locates the thermal deformation field associated with the command. Finally, it measures the relative displacement of the inner and outer rings of the bearing in the direction limited by structural constraints from the thermal deformation field as the hot working clearance data.

[0065] S5.2: Based on the characteristic point temperature, maximum temperature, interstage heat flow data, and hot operating clearance data of each bearing stage, calculate the comprehensive thermal stability risk index for each bearing stage. The expression is: ; In the formula, Indicates the first Comprehensive thermal stability risk index of grade bearings; Indicates the bearing level index; Indicates the first Characteristic point temperature of a first-stage bearing; This indicates the critical alarm temperature threshold (e.g., 80°C) determined based on the tempering softening temperature of the bearing material and the dropping point of the grease. Indicates the first Interstage heat flow data for multistage bearings; This indicates the maximum allowable heat flux carrying capacity (e.g., 1500W) calculated through thermal balance simulation based on the bearing's rated dynamic load, limiting speed, and heat dissipation boundary conditions. Indicates the first Hot operating clearance data for grade A bearings; This represents the minimum limit of safe clearance (e.g., 15 μm) calculated based on bearing clearance standards and minimum oil film thickness theory to maintain normal lubrication without mechanical interference.

[0066] S5.3: Encapsulate the characteristic point temperature, maximum temperature, interstage heat flow data, hot working clearance data, and comprehensive thermal stability risk index of each bearing stage into key thermal stability indicator data; Specifically, the temperature values ​​of characteristic points of each bearing stage, the highest temperature value, the interstage heat flow data value, the hot working clearance data value, and the comprehensive thermal stability risk index value are recorded to form key thermal stability indicator data.

[0067] S5.4: Analyze and evaluate the key thermal stability index data, identify potential failure modes based on predefined risk assessment rules, and generate multidimensional risk characterization results; Specifically, key thermal stability index data are read, a comprehensive thermal stability risk index is extracted, and potential failure mode types are identified. If the comprehensive thermal stability risk index exceeds the first-level threshold, it is marked as a "temperature overload" potential failure mode; if it exceeds the second-level threshold, it is marked as a "heat flow congestion" potential failure mode; and if it exceeds the third-level threshold, it is marked as a "clearance disappearance" potential failure mode. Combining the contribution ratios of characteristic point temperatures, maximum temperatures, interstage heat flow data, and hot-state operating clearance data of each bearing stage in each potential failure mode, a multi-dimensional risk characterization result is generated. The multi-dimensional risk characterization result includes the failure mode category, risk level, and dominant risk factor.

[0068] It should be noted that the potential failure mode classification is based on the specific thermal failure mechanism; the first-level threshold is set according to the critical point of bearing steel hardness reduction, corresponding to the risk of material performance degradation; the second-level threshold is set according to the critical point of lubricant carbonization, corresponding to the risk of lubricating medium failure; and the third-level threshold is set according to the critical point of metal contact seizure, corresponding to the risk of mechanical structure lockup. For example, the value range is as follows: the first-level threshold range is 0.5 to 0.7, based on the material yield strength retention rate not less than 80%; the second-level threshold range is 0.7 to 0.9, based on the critical point of lubricant oxidation and deterioration and the heat flux density saturation limit; and the third-level threshold range is 0.9 to 0.95, based on the lubricating film thickness being less than the total surface roughness.

[0069] S5.5: Based on operating conditions and key thermal stability index data, analyze the correlation between operating conditions and thermal stability; Specifically, operating condition data and key thermal stability index data are collected. Operating conditions such as speed, load, and ambient temperature are time-series aligned with characteristic point temperatures, maximum temperatures, interstage heat flow data, and hot-state operating clearance data of each bearing stage. The fluctuation ranges of these data under different combinations of operating conditions are compared to establish a mapping relationship between changes in operating conditions and changes in key thermal stability index data. If an increase in speed is accompanied by a sharp increase in interstage heat flow data, speed is determined to be the dominant factor in heat flow congestion. If an increase in load leads to a rapid reduction in hot-state operating clearance data, load is determined to be the dominant factor in clearance disappearance. The response lag time of characteristic point temperatures, maximum temperatures, interstage heat flow data, and hot-state operating clearance data of each bearing stage to changes in operating conditions is comprehensively considered to quantify the correlation between operating conditions and thermal stability.

[0070] S5.6: Generate a structured analysis report based on potential failure modes, correlations, and multidimensional risk characterization results; Specifically, the report integrates the correlations between various potential failure modes, operating conditions, and thermal stability, as well as the results of multi-dimensional risk characterization. It arranges the abnormal evolution processes of characteristic point temperatures, maximum temperatures, interstage heat flow data, and hot-state operating clearance data for each bearing level in chronological order. It matches the potential failure mode types with the dominant factors identified in the correlations to form a failure cause description paragraph. It fills the risk level determination column with the comparison results of the comprehensive thermal stability risk index value from the multi-dimensional risk characterization results and the first, second, and third level thresholds. Based on the quantified response lag time in the correlations, it calculates the remaining time for fault development and writes it into the early warning suggestion section. It summarizes the extreme value records of characteristic point temperatures, maximum temperatures, interstage heat flow data, and hot-state operating clearance data for each bearing level as supporting data attachments, generating a structured analysis report.

[0071] In summary, this invention achieves the transformation from users' fuzzy and unstructured engineering analysis needs to precise and executable computer simulation instruction sequences by receiving and parsing natural language task descriptions and generating hot grid analysis task plans. It deeply integrates natural language processing and automatic task planning, enabling it to understand complex engineering semantics, decompose task logic, and orchestrate a complete simulation workflow. This not only liberates engineers from tedious and error-prone software operations and script writing but also lowers the technical threshold for using professional CAE tools, ensuring the standardization, automation, and efficient startup of the analysis process, laying a solid foundation for fully automated simulation execution and analysis.

[0072] It should be noted that 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 preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for analyzing the thermal stability of multi-stage rolling bearings based on a thermal mesh model, characterized in that, include: Extract analysis task parameters from the natural language task description input by the user, and generate a hot grid analysis task plan; Call the pre-set parametric modeling library, execute the thermal mesh analysis task plan, and generate a three-dimensional geometric model of a multi-stage rolling bearing; discretize the three-dimensional geometric model of the multi-stage rolling bearing to generate an initial thermal mesh model; The initial thermal mesh model is checked and repaired based on preset mesh quality rules to obtain a qualified thermal mesh model. Material properties, heat sources, heat transfer coefficients, and thermal constraints are then assigned to the qualified thermal mesh model to construct a multi-level bearing-coupled thermal mesh solution model. The steps are as follows: Based on the qualified thermal grid model, operating conditions and physical boundary conditions, the heat source action area, heat conduction and transfer area and heat transfer boundary area are identified, and characteristic interaction relationships are constructed according to the thermal coupling relationship between each area. Based on the feature interaction relationship, material properties, heat source, heat transfer coefficient and thermal constraint conditions are assigned to qualified thermal grid models, and the heat source, heat transfer coefficient and thermal constraint conditions are adjusted in a coordinated manner. The heat source, heat transfer coefficient and thermal constraint conditions after coordinated adjustment are integrated into a qualified thermal mesh model to generate a multi-level bearing coupled thermal mesh solution model. Submit and run the multi-level bearing coupled thermal mesh solution model, monitor the solution process, diagnose and adjust solution anomalies, obtain the temperature field and heat flow field, and based on the temperature field, combine the material thermal expansion parameters and structural constraints to obtain the thermal deformation field. The steps are as follows: Start the thermal solver to load the multi-stage bearing coupled thermal mesh solution model, start the embedded computing agent in parallel, and establish real-time data monitoring; The embedded computing agent calculates a health status index that characterizes the convergence state of the thermal solver based on real-time monitoring data. When an abnormality in the health status index is diagnosed, the parameters of the thermal solver are adjusted. After the parameter adjustment is executed, the convergence of the thermal solver solution process is verified by real-time data monitoring. Once the thermal solver converges, the temperature field and heat flow field are extracted. Based on the thermal deformation calculation instructions, the temperature field is matched according to the mapping category corresponding to the material thermal expansion parameters and structural constraints in the preset temperature field-thermal deformation field mapping library, and the corresponding thermal deformation field is retrieved. The key thermal stability index data are extracted from the temperature field, heat flow field, and thermal deformation field, including characteristic point temperature, maximum temperature, interstage heat flow data, and hot working clearance data of each bearing stage. The steps are as follows: The analysis result data extraction command extracts characteristic point temperature, maximum temperature, interstage heat flow data and hot working clearance data of bearings at each stage from the temperature field, heat flow field and thermal deformation field; Based on the characteristic point temperature, maximum temperature, interstage heat flow data and hot working clearance data of each bearing stage, the comprehensive thermal stability risk index of each bearing stage is calculated. The characteristic point temperature, maximum temperature, interstage heat flow data, hot working clearance data of each bearing stage, and comprehensive thermal stability risk index are packaged into key thermal stability indicator data. Analyze the key thermal stability indicators and generate a structured analysis report.

2. The method for analyzing the thermal stability of multi-stage rolling bearings based on a thermal mesh model as described in claim 1, characterized in that, The steps for generating the thermal mesh analysis task plan are as follows: Receive and normalize the natural language task description to obtain normalized text, and extract the bearing model, structural parameters, operating conditions and thermal stability analysis objectives from the normalized text; The bearing model, structural parameters, operating conditions, and thermal stability analysis objectives are matched and mapped with a predefined bearing parameter specification library to generate analysis task parameter objects. The task logic is parsed and the task sequence is arranged for the analysis task parameter object to generate a hot grid analysis task plan.

3. The method for analyzing the thermal stability of multi-stage rolling bearings based on a thermal mesh model as described in claim 2, characterized in that, The thermal mesh analysis task plan includes parametric geometric modeling instructions, mesh generation and quality check instructions, physical boundary condition setting instructions, solver submission and monitoring instructions, thermal deformation calculation instructions, and result data extraction instructions.

4. The method for analyzing the thermal stability of multi-stage rolling bearings based on a thermal mesh model as described in claim 3, characterized in that, The steps for generating the three-dimensional geometric model of the multi-stage rolling bearing are as follows: Retrieve and instantiate feature templates corresponding to parametric geometry modeling instructions from the pre-set parametric modeling library to generate single-level parametric geometry for each bearing level; Based on the axial arrangement relationship and installation spacing defined in the parametric geometry modeling command, spatial transformation and Boolean union operations are performed on the single-level parametric geometry of each bearing to generate a multi-level bearing assembly; Geometric cleaning and repair of the multi-stage bearing assembly were performed to form a three-dimensional geometric model of the multi-stage rolling bearing.

5. The method for analyzing the thermal stability of multi-stage rolling bearings based on a thermal mesh model as described in claim 3, characterized in that, The steps for generating the initial thermal mesh model are as follows: The mesh generation and quality inspection instructions are analyzed to identify the key geometric features of each bearing level from the three-dimensional geometric model of the multi-level rolling bearing, and to generate a set of key geometric features. Based on the key geometric feature set and the parameters in the mesh generation instruction, an adaptive mesh size field is generated that is refined in the expected high temperature gradient region and the heat flow concentration region. Based on the adaptive mesh size field, the three-dimensional geometric model of the multi-stage rolling bearing is divided into parallel volume meshes to generate an initial thermal mesh model.

6. The method for analyzing the thermal stability of multi-stage rolling bearings based on a thermal mesh model as described in claim 5, characterized in that, The steps to obtain a qualified thermal mesh model are as follows: Calculate the Jacobian ratio and aspect ratio of each grid in the initial thermal grid model; Based on the Jacobian ratio, aspect ratio, and preset mesh quality rules, node smoothing and local reconstruction operations are performed on the initial hot mesh model to generate the repaired mesh. The size deviation of the repaired mesh is corrected based on the adaptive mesh size field to obtain a qualified thermal mesh model.

7. The method for analyzing the thermal stability of multi-stage rolling bearings based on a thermal mesh model as described in claim 1, characterized in that, The steps for generating the structured analysis report are as follows: The key thermal stability index data are analyzed and evaluated, and potential failure modes are identified based on predefined risk assessment rules, generating multidimensional risk characterization results. Based on operating conditions and key thermal stability data, the correlation between operating conditions and thermal stability is analyzed. A structured analysis report is generated based on potential failure modes, correlations, and multidimensional risk characterization results.

Images