Wheel dual-axis performance evaluation method and device based on numerical calculation and server

Abstract

This invention provides a method, apparatus, and server for evaluating the performance of a dual-axle wheel based on numerical computation, relating to the technical field of cloud simulation applications. The method includes: associating and building various functional components according to the performance evaluation requirements of the dual-axle wheel to generate a dual-axle wheel simulation application flow; acquiring available supercomputing resources at the current moment, and combining the available supercomputing resources with the queue information of components to be executed and the component dependency information corresponding to the dual-axle wheel simulation application flow to form scheduling state information; using a reinforcement learning model to process the scheduling state information to generate the execution order of target components; calling various functional components according to the execution order of target components; generating a dual-axle equivalent loading model based on a three-dimensional wheel model; and performing dual-axle performance evaluation on the dual-axle equivalent loading model to obtain the dual-axle wheel performance evaluation result. This invention can significantly improve the efficiency and accuracy of dual-axle wheel performance evaluation.

Description

Technical Field

[0001] This invention relates to the technical field of cloud simulation applications, and in particular to a method, device, and server for evaluating the dual-axle performance of wheels based on numerical calculation. Background Technology

[0002] During service, wheels are simultaneously subjected to the coupled effects of radial rolling loads and axial bending loads, and their fatigue life directly affects the safety of the entire vehicle. Currently, related technologies suggest that existing wheel biaxial performance evaluation mainly relies on physical biaxial fatigue testing equipment to simulate actual working conditions through physical loading. However, such testing equipment is expensive, requires a large space, and has a long test preparation cycle, making it difficult to meet the needs of rapid iteration and multi-solution comparison during the product design phase.

[0003] With the development of numerical computing technology, finite element simulation is increasingly used for predicting wheel fatigue performance. However, existing simulation methods have the following problems: First, professional personnel are required to manually establish the analysis process, which is complex and highly dependent on the experience of the personnel. The modeling methods of different engineers vary greatly, resulting in insufficient consistency of analysis results. Second, existing simulations often calculate radial rolling conditions and bending conditions separately, and then superimpose stress results through post-processing. This approach fails to achieve biaxial load coupling in a unified time domain, making it difficult to accurately reflect the actual stress process and thus affecting the accuracy of fatigue life prediction. Finally, existing cloud simulation platforms are mostly single-function modules and lack the ability to build automated processes for biaxial performance evaluation of wheels. Users still need to manually configure each analysis step, making it difficult to quickly deploy complex simulation processes. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a method, device and server for evaluating the dual-axle performance of wheels based on numerical calculation, which can significantly improve the efficiency and accuracy of the evaluation of dual-axle performance of wheels.

[0005] In a first aspect, embodiments of the present invention provide a method for evaluating the dual-axle performance of a wheel based on numerical computation. The method includes: acquiring a three-dimensional model of the wheel and the dual-axle performance evaluation requirements; and, based on the dual-axle performance evaluation requirements, calling corresponding functional components from a preset functional component library to perform association and construction processing on each functional component, generating a dual-axle simulation application flow corresponding to the dual-axle performance evaluation requirements; acquiring available supercomputing resources at the current moment, and combining the available supercomputing resources with the queue information of components to be executed and the component dependency information corresponding to the dual-axle simulation application flow to form scheduling state information; performing decision processing on the scheduling state information through a reinforcement learning model to generate the execution order of target components; calling each functional component according to the execution order of target components; generating a dual-axle equivalent loading model based on the three-dimensional model of the wheel; and performing dual-axle performance evaluation processing on the dual-axle equivalent loading model to obtain the dual-axle performance evaluation result.

[0006] In one implementation, the step of calling various functional components according to the execution order of the target components and generating a biaxial equivalent loading model based on the wheel 3D model includes: rendering the wheel 3D model to obtain an interactive 3D model, and determining the flange mounting surface area information, rim edge area information, and tire bead contact area information based on the user's interaction information with the interactive 3D model; automatically identifying the wheel rotation axis based on the flange mounting surface area information, rim edge area information, tire bead contact area information, and the execution order of the target components through a geometric recognition component and a coordinate establishment component, and establishing a local polar coordinate system to generate a biaxial equivalent loading model.

[0007] In one implementation, after the step of generating the biaxial equivalent loading model, the method includes: by calling the boundary condition construction component, setting the flange mounting surface area corresponding to the flange mounting surface area information as a fixed area, and constraining the translational and rotational degrees of freedom within the fixed area to generate boundary constraint conditions, wherein the boundary constraint conditions are used to simulate the force state of the wheel connected to the axle through the flange.

[0008] In one embodiment, the step of performing biaxial performance evaluation processing on a biaxial equivalent loading model to obtain the biaxial performance evaluation result of the wheel includes: acquiring the radial pressure field and bending load, and simultaneously superimposing the radial pressure field and bending load in the same time integration step to generate the equivalent stress field and principal stress distribution; performing fatigue life calculation processing based on the equivalent stress field and principal stress distribution to obtain the biaxial performance evaluation result of the wheel.

[0009] In one embodiment, the step of obtaining the bending load includes: applying a time-varying bending moment to the rim edge region using a reverse reaction force loading model; converting the bending moment into an equivalent distributed force based on the time-varying bending moment and rim radius information; and applying the equivalent distributed force to the rim edge node to generate a bending load uniformly distributed along the rim circumference.

[0010] In one embodiment, before the step of simultaneously superimposing and calculating the radial pressure field and bending load to generate the equivalent stress field and principal stress distribution, the method includes: acquiring the body structure information of the wheel bolt rod, and performing geometric classification and recognition processing on the body structure information to obtain a cylinder recognition result; based on the cylinder recognition result, extracting the cylinder side surface information from the body structure information, and performing axis fitting processing on the cylinder side surface information to obtain the bolt rod axis direction information; using the bolt rod axis direction information as the preload application direction, and constructing the bolt preload based on the preload application direction, so as to use the bolt preload as a load boundary condition to participate in the simultaneous superimposition and calculation processing to generate the equivalent stress field and principal stress distribution.

[0011] In one embodiment, the step of calculating fatigue life based on the equivalent stress field and principal stress distribution to obtain the dual-axle performance evaluation result of the wheel includes: acquiring standardized road spectrum input parameter information, decomposing the standardized road spectrum input parameter information to generate various sub-working condition information; arranging the equivalent stress field and principal stress distribution corresponding to each time step in the various sub-working condition information in chronological order to obtain the stress time history information corresponding to each sub-working condition information; performing cyclic identification and damage accumulation processing based on the stress time history information to generate comprehensive accumulated damage information and wheel predicted fatigue life information, and using the comprehensive accumulated damage information and wheel predicted fatigue life information as the dual-axle performance evaluation result of the wheel.

[0012] Secondly, embodiments of the present invention also provide a wheel dual-axle performance evaluation device based on numerical calculation. The device includes: a component association module, which acquires a three-dimensional wheel model and wheel dual-axle performance evaluation requirements, and calls corresponding functional components from a preset functional component library according to the wheel dual-axle performance evaluation requirements to perform association and construction processing on each functional component, generating a wheel dual-axle simulation application process corresponding to the wheel dual-axle performance evaluation requirements; a reinforcement learning module, which acquires available supercomputing resources information at the current moment, and combines the available supercomputing resources information with the queue information of components to be executed and the component dependency information corresponding to the wheel dual-axle simulation application process to form scheduling state information, so as to perform decision processing on the scheduling state information through a reinforcement learning model to generate the execution order of the target components; and a dual-axle performance evaluation module, which calls each functional component according to the execution order of the target components, generates a dual-axle equivalent loading model based on the wheel three-dimensional model, and performs dual-axle performance evaluation processing on the dual-axle equivalent loading model to obtain the wheel dual-axle performance evaluation result.

[0013] Thirdly, embodiments of the present invention also provide a server, including a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement any of the methods provided in the first aspect.

[0014] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing computer-executable instructions, which, when invoked and executed by a processor, cause the processor to implement any of the methods provided in the first aspect.

[0015] The embodiments of the present invention bring the following beneficial effects: This invention provides a method, apparatus, and server for evaluating the dual-axle performance of a wheel based on numerical computation. The method first obtains a three-dimensional model of the wheel and the dual-axle performance evaluation requirements. Based on these requirements, it calls corresponding functional components from a pre-set functional component library to associate and build the components, generating a dual-axle simulation application flow corresponding to the performance evaluation requirements. Next, it obtains the available supercomputing resources at the current moment and combines this information with the queue information of components to be executed and the component dependency information corresponding to the dual-axle simulation application flow to form scheduling state information. A reinforcement learning model is then used to process this scheduling state information to generate the execution order of the target components. Finally, based on the execution order, the method calls the functional components, generates a dual-axle equivalent loading model based on the three-dimensional wheel model, and performs dual-axle performance evaluation on the model to obtain the evaluation result. This invention significantly improves the efficiency and accuracy of dual-axle performance evaluation.

[0016] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained in accordance with the structures particularly pointed out in the description, claims and drawings.

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0018] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating a numerical calculation-based method for evaluating the performance of a dual-axle wheel, provided in an embodiment of the present invention; Figure 2 A schematic diagram of a local coordinate system provided in an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating a dual-axis simulation effect provided in an embodiment of the present invention; Figure 4 A schematic diagram of the axis of a bolt rod provided in an embodiment of the present invention; Figure 5 A schematic diagram of a wheel dual-axle performance evaluation device based on numerical calculation provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of the structure of a server provided in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Currently, with the continuous development of lightweight automotive design and complex road operating environments, wheel structures simultaneously bear the coupled effects of radial rolling loads and axial bending loads during service, and their fatigue life directly affects the safety and reliability of the entire vehicle. Traditional wheel performance evaluation mainly relies on biaxial fatigue testing equipment, which simulates actual working conditions through physical loading. However, such biaxial testing equipment is expensive, occupies a large space, has a complex test preparation process, and a long test cycle, making it difficult to meet the needs of rapid iteration and multi-scheme comparison in the product design stage. Especially for small and medium-sized manufacturing enterprises, the high equipment investment and testing costs limit the efficiency of wheel structure optimization design.

[0022] With the development of numerical computing technology, finite element simulation has been gradually used for wheel fatigue performance prediction. However, existing simulation methods usually require professionals to manually establish complex analysis processes, including geometric identification, coordinate establishment, biaxial load application, and fatigue life calculation. This is not only complicated to operate, but also results in significant differences in modeling methods among different engineers, leading to insufficient consistency in analysis results.

[0023] Therefore, a numerical computation-based intelligent agent system for dual-axis wheel performance evaluation is proposed. By modularizing dual-axis simulation-related functions and utilizing an intelligent agent to automatically combine and schedule processes, a dedicated dual-axis performance evaluation application is generated in the cloud. This achieves unified coupled analysis of radial rolling and axial bending conditions of the wheel, and combines road spectrum data for comprehensive fatigue life prediction. This technology reduces reliance on expensive testing equipment, improves simulation analysis efficiency and consistency, and shortens wheel product development cycles, possessing significant engineering application value for enhancing wheel structure design capabilities and digital R&D levels.

[0024] Existing biaxial performance evaluation methods for wheels primarily rely on physical biaxial fatigue testing equipment, which simulates actual service conditions of wheels by simultaneously applying radial rolling loads and axial bending loads. While this type of testing can realistically reflect the stress state of the wheel, the equipment is expensive, the test preparation cycle is long, and it requires complex clamping and calibration procedures. This makes it difficult to conduct multiple rounds of verification during the product design phase, resulting in low development efficiency. Furthermore, physical testing is mainly used in the final verification stage and cannot meet the needs of early structural optimization and rapid solution selection.

[0025] To reduce testing costs, some companies use finite element numerical simulation (FEM) to predict wheel fatigue performance. However, existing simulation methods typically require engineers to manually complete steps such as model processing, coordinate system establishment, load definition, and fatigue analysis. This process is complex and highly dependent on personnel experience. Differences in modeling methods among different engineers affect the consistency of results. Furthermore, existing simulation methods often calculate radial rolling and bending conditions separately and then superimpose stress results through post-processing. This fails to achieve biaxial load coupling within a unified time domain, making it difficult to accurately reflect the stress change process under actual biaxial test conditions, thus affecting the accuracy of fatigue life prediction.

[0026] On the other hand, existing cloud simulation platforms are mostly single-function modules, lacking the ability to build automated workflows for dual-axle performance evaluation of wheels. Users still need to manually configure each analysis step, making it difficult to quickly deploy complex simulation workflows. Therefore, existing solutions still have significant limitations in terms of cost, efficiency, and analytical consistency.

[0027] Based on this, the present invention provides a method, device, and server for evaluating the dual-axle performance of wheels based on numerical calculation. This allows for the construction of a unified dual-axle coupled calculation algorithm within a web-based 3D interactive environment. Through component-based encapsulation, road spectrum-driven approach, and a combined fatigue life algorithm with supercomputing cluster solution, it enables the construction of a simulation application for analyzing the dual-axle test strength and cumulative fatigue life of wheels under coupled radial rolling and axial bending conditions without the need for physical testing equipment. Furthermore, a multi-functional simulation component library can be built in the cloud, and the dual-axle performance evaluation simulation process can be automatically constructed and executed through an intelligent agent scheduling algorithm. Unlike traditional fixed-process simulation software, this invention automatically combines functional modules such as geometric recognition, boundary construction, dual-axle load loading, road spectrum generation, solution calculation, and fatigue life evaluation through a background intelligent agent, forming a cloud-based simulation application for dual-axle performance evaluation of wheels based on an intelligent agent-driven mechanism. This achieves a one-click intelligent evaluation capability without the need for manual construction and operation of the analysis process.

[0028] See Figure 1 The diagram shows a flowchart of a numerical calculation-based method for evaluating the performance of a dual-axle wheel. This method is applied to a numerical calculation-based dual-axle wheel performance evaluation system. The system includes: a functional component library, a task parsing module, a simulation process generation module, an intelligent scheduling module, and a numerical solution execution module. The method mainly includes the following steps S102 to S106: Step S102: Obtain the three-dimensional model of the wheel and the dual-axle performance evaluation requirements of the wheel. Based on the dual-axle performance evaluation requirements of the wheel, call the corresponding functional components from the preset functional component library to associate and build each of the functional components, and generate the dual-axle simulation application process of the wheel corresponding to the dual-axle performance evaluation requirements of the wheel.

[0029] In one implementation, firstly, the three-dimensional model of the wheel and the dual-axle performance evaluation requirements of the wheel are obtained. The three-dimensional model of the wheel refers to the digital geometric model of the wheel to be evaluated uploaded by the user, which is used for subsequent mesh generation and finite element calculation. The dual-axle performance evaluation requirements of the wheel refer to the evaluation target parameters set by the user, specifically including information such as radial load amplitude, bending moment range, loading frequency, and the type of road condition to be simulated. This requirement information can be obtained by the user through the interactive interface on the web page or by importing a configuration file.

[0030] After obtaining the above information, the system calls the corresponding functional components from the preset functional component library according to the dual-axle performance evaluation requirements. The preset functional component library refers to a pre-built collection of software modules stored in the cloud. This collection includes functional modules such as geometry recognition components, coordinate establishment components, boundary condition construction components, dual-axle load loading components, solution calculation components, and fatigue life assessment components. Each component encapsulates specific simulation calculation functions and provides a unified calling interface. The system automatically identifies the type and quantity of components to be called based on the working condition type and calculation objective in the evaluation requirements. For example, when the requirements include radial loading, the dual-axle load loading component is called; when the requirements include fatigue assessment, the fatigue life assessment component is called.

[0031] Subsequently, the system performs association and construction processing on the various functional components obtained from the calls. Association and construction refers to determining the execution order and data transmission path of components based on the input-output dependencies between them. For example, the output of the geometry recognition component serves as the input of the coordinate establishment component, and the output of the coordinate establishment component serves as the input of the dual-axis load loading component. By parsing the interface definitions of each component, the system automatically establishes the above dependency links, generating a complete dual-axis wheel simulation application process. This process is an executable, structured computational workflow that includes all computational steps from model input to result output and their order.

[0032] Step S104: Obtain the available supercomputing resources at the current moment, and combine the available supercomputing resources with the queue information of the components to be executed and the component dependency information corresponding to the dual-axle wheel simulation application process to form scheduling state information, so as to generate the target component execution order by making decision processing on the scheduling state information through a reinforcement learning model.

[0033] In one implementation, after constructing the dual-axle wheel simulation application process, the system obtains the available supercomputing resources at the current moment. Available supercomputing resources refer to parameters such as the number of idle and available computing nodes in the supercomputing cluster, the number of CPU cores on each node, memory capacity, and network bandwidth. This information is obtained in real-time through the supercomputing cluster's resource management interface and is used to determine the currently schedulable computing capacity. Simultaneously, the system obtains the queue information of components to be executed and the component dependency information corresponding to the simulation application process. The queue information of components to be executed refers to a list of functional components that have not yet been executed in the current dual-axle wheel simulation application process. This list records the identifier, type, and estimated computational load of each component and is output by the simulation process generation module after completing the component association setup. The component dependency information refers to the execution order constraints between functional components. For example, the geometry recognition component must be executed before the coordinate establishment component, and the mesh generation component must be executed before the solution calculation component. This dependency is determined by the input / output interface definitions of each component in the component library and is automatically generated during process setup.

[0034] The system combines the above three types of information to form scheduling status information. Combination refers to encapsulating available supercomputing resources, queues of components to be executed, and component dependencies according to a predefined data structure to form a unified input vector. This vector fully describes the current system state, including which components need to be executed, the sequential constraints between components, and the amount of computing resources available for allocation. Subsequently, the system inputs this scheduling status information into a reinforcement learning model. The reinforcement learning model's decision-making process involves selecting the optimal component scheduling action from all possible options, considering both minimizing computation time and maximizing resource utilization, based on the current state. This generates the target component execution order. The target component execution order refers to the sequential arrangement and parallel allocation scheme of the functional components during actual execution. This order satisfies the dependencies between components and achieves the shortest overall computation time under the current available resource conditions. At this point, the system completes the scheduling decision, providing clear sequential instructions for subsequent component execution.

[0035] Step S106: According to the execution order of the target components, call each of the functional components, generate a dual-axis equivalent loading model based on the three-dimensional model of the wheel, and perform dual-axis performance evaluation processing on the dual-axis equivalent loading model to obtain the dual-axis performance evaluation result of the wheel.

[0036] In one implementation, after generating the execution order of the target components, the system calls various functional components according to this execution order. During the calling process, based on the user-uploaded 3D wheel model, the system first executes the geometry recognition component and the coordinate establishment component to generate a biaxial equivalent loading model. The biaxial equivalent loading model refers to the wheel numerical model after axis recognition, coordinate system establishment, and boundary condition definition. This model includes the wheel's geometric topology information, local polar coordinate system definition, and flange surface fixed constraint settings. It is the basic data object for subsequent biaxial coupling calculations. Specifically, based on the flange mounting surface area, rim edge area, and tire bead contact area information selected by the user on the interactive 3D model, the system automatically identifies the wheel's rotation axis using a least-squares axis fitting algorithm, and establishes a local polar coordinate system with this axis as the Z-axis and the flange surface centroid as the origin, thereby completing the construction of the biaxial equivalent loading model.

[0037] Subsequently, the system performs biaxial performance evaluation on the biaxial equivalent loading model. Biaxial performance evaluation involves simultaneously superimposing and calculating the radial pressure field and bending load within the same time integration step, outputting the equivalent stress field and principal stress distribution, and calculating fatigue life based on these stress field results. Specifically, the system first constructs the radial pressure field and bending load. The radial pressure field is obtained by transforming the rolling contact problem into a time-varying contact pressure field, and the bending load is obtained by applying an equivalent distributed force to the rim edge using a reverse reaction force loading method. Then, the system simultaneously superimposes and calculates the radial pressure field and bending load within the same time integration step, generating the equivalent stress field and principal stress distribution for each time step. Finally, the system performs fatigue life calculation based on these stress field results, generating comprehensive cumulative damage information and predicted wheel fatigue life information. The system outputs the above information as the biaxial performance evaluation result of the wheel. This result can be directly used to determine whether the fatigue life of the wheel under a given working condition meets the design requirements and provides a quantitative basis for wheel structure optimization.

[0038] The numerical calculation-based wheel dual-axle performance evaluation method provided in this invention can significantly improve the efficiency and accuracy of wheel dual-axle performance evaluation.

[0039] In one implementation, a 3D model of the wheel can be rendered to obtain an interactive 3D model. Based on the interaction information between the user and the interactive 3D model, information on the flange mounting surface area, the rim edge area, and the bead contact area is determined. Then, through a geometric recognition component and a coordinate establishment component, based on the flange mounting surface area information, the rim edge area information, the bead contact area information, and the execution order of the target component, the wheel rotation axis is automatically identified and a local polar coordinate system is established to generate a dual-axis equivalent loading model. Finally, by calling a boundary condition construction component, the flange mounting surface area corresponding to the flange mounting surface area information is set as a fixed area, and the translational and rotational degrees of freedom are constrained within the fixed area to generate boundary constraint conditions. The boundary constraint conditions are used to simulate the force state of the wheel connected to the axle through the flange.

[0040] Specifically, in the intelligent scheduling phase of the system, to improve cloud simulation efficiency and reduce computing resource consumption, this invention introduces an agent scheduling strategy based on reinforcement learning. The system uses the currently available supercomputing resources, the queue of components to be executed, and the complete dependencies of the simulation application process as the scheduling state, and the execution order of components as the decision action. The agent continuously learns the computation time and resource consumption under different scheduling strategies, updates the state action value function, and thus obtains the optimal execution strategy.

[0041] In actual execution, the agent prioritizes the parallel execution of components with high computational load but few dependencies (such as the meshing component, which will perform independent meshing first), while sequentially arranging components with data dependencies (such as remote force components that require interactive surface information), so that the entire dual-axis simulation process can be parallelized in the cloud.

[0042] The agent scheduling part adopts a reinforcement learning model, and the system state is defined as follows:

[0043] in, Let t be the system state at time t; The status of computing resources (supercomputing node resources); Queue of components to be executed; This refers to the component relationships.

[0044] The Q-learning update strategy is used to optimize the component execution order and resource allocation.

[0045] in, Let the state-action value function be... The scheduling function that specifies the component to be executed at the current moment; The learning rate; Discount factor; The reward function consists of computation time parameters and grid quality parameters with different weights.

[0046] The intelligent system for evaluating the performance of a wheel on two axes needs to take into account a series of radial load coefficients, lateral load coefficients and test distances in the wheel load spectrum conditions. First, a unified algorithm model for dual-axis simulation under a single load spectrum condition is constructed.

[0047] After developing a cloud simulation application for biaxial performance evaluation, the evaluation system can be displayed on a webpage via a backend call and interactive frontend. After rendering a 3D model of the wheel on the webpage, users can select the flange mounting surface area, rim edge area, and bead contact area through basic 3D interaction. Based on the selected set of flange mounting surface nodes, the system automatically identifies the wheel's axis of rotation using a least-squares axis fitting algorithm.

[0048] Let the user select the set of flange mounting surface nodes. for:

[0049] Where n is the total number of selected nodes; Let be the coordinates of the i-th node in the global coordinate system.

[0050] Calculate the centroid of the node :

[0051] Specifically, this is achieved by calculating the geometric centroid of the selected nodes and constructing the covariance matrix:

[0052] in, It is a 3×3 covariance matrix used to describe the directionality of point cloud distribution.

[0053] This method incorporates road spectrum information from biaxial tests, taking into account the deflection angle between the wheel and drive shaft during drifting or turning. The simulation method also incorporates this deflection under wheel load conditions. By solving for its eigenvalues ​​and eigenvectors, the eigenvector corresponding to the smallest eigenvalue is taken as the rotation axis direction, thus obtaining a stable and noise-resistant axis identification result. A local polar coordinate system is established with this rotation axis as the Z-axis and the flange centroid as the origin, achieving a unified expression of subsequent radial load and bending moment distribution. This automatic identification mechanism avoids errors caused by manually establishing a reference coordinate system, improving modeling efficiency and computational consistency.

[0054] This invention also provides an implementation method for evaluating the dual-axle performance of wheels, as detailed in (1) to (3) below: (1) Obtaining radial pressure field and bending load: By using the reverse reaction force loading model, a bending moment that varies with time is applied to the edge region of the rim. Then, based on the bending moment that varies with time and the rim radius information, the bending moment is converted into an equivalent distributed force, and the equivalent distributed force is applied to the edge node of the rim to generate a bending load that is uniformly distributed along the circumference of the rim.

[0055] In one implementation, see Figure 2 The diagram shows a local coordinate system where the circumferential angles directly reflect the wheel's rotational state under different operating conditions, thus corresponding to the road spectrum. Figure 2 This diagram shows the XZ plane (with the axis of the wheel's A-side as the Z-axis). A simplified diagram of the wheel is shown on the left, with the origin at... This refers to the center point of the wheel coordinates. The horizontal black line represents the driveshaft connecting the wheel. Based on the actual vehicle conditions requiring wheel deflection, it shows the state of the driveshaft when deflected by an angle θ. The dashed line of length L of the driveshaft represents the new driveshaft position. The x and z coordinates of the driveshaft endpoint after rotation can be determined using the following formula. The driveshaft length L and deflection angle θ are known quantities. Here, the radial distance r is the absolute value of the x-coordinate, so the x and z axes of the driveshaft under the new operating condition can be calculated. Combining this with the principle that the driveshaft length remains unchanged, its y-coordinate can also be solved. The orientation of the driveshaft coupling point conforming to the physical deflection angle can be standardized to reconstruct the road spectrum state.

[0056]

[0057] Where r is the radial distance and θ is the circumferential angle.

[0058] Regarding boundary condition construction, to make the simulation conditions more consistent with the actual vehicle installation state, this invention changes the traditional passenger car bending life assessment method of clamping the wheel rim edge and applying bending moment through bolt holes. Instead, it defines the wheel flange mounting surface area as a new fixed area. The system constrains the translational and rotational degrees of freedom in this area to simulate the actual stress state of the wheel connected to the axle via the flange and the caliper. This boundary reconstruction method makes the load transfer path more consistent with actual vehicle assembly conditions, avoids the problem of excessive local stiffness caused by traditional clamping models, and also provides a boundary basis for simultaneous step loading in dual-axle conditions.

[0059] In constructing the axial bending condition, this invention employs a reverse reaction force loading method, that is, applying a time-varying bending moment to the rim edge region, rather than directly applying a concentrated moment through the bolt hole region. The bending moment expression is set as follows:

[0060] in, The bending moment amplitude, The loading frequency (which can be synchronized with the rolling angular velocity). This is the phase difference parameter.

[0061] The bending moment is applied to the edge node of the rim in the form of an equivalent distributed force, and its equivalent force value satisfies:

[0062] in Let be the rim radius. This method achieves a uniform distribution of bending load along the circumference of the rim, enabling a real coupling relationship between the bending stress field and the radial rolling stress field in physical space.

[0063] Further, see Figure 3 The diagram shows a biaxial simulation effect. The present invention adopts a unified time step synchronous loading method to superimpose the radial pressure field and bending load in the same time integration step, instead of using the traditional method of linear superposition of stress after step loading.

[0064] (2) Within the same time integration step, the radial pressure field and bending load are synchronously superimposed and calculated to generate the equivalent stress field and principal stress distribution. In a real-time mode, the body structure information of the wheel bolt rod can be obtained, and the body structure information is geometrically classified and identified to obtain the cylinder identification result. Then, based on the cylinder identification result, the cylinder side information is extracted from the body structure information, and the cylinder side information is fitted with the axis to obtain the bolt rod axis direction information. Finally, the bolt rod axis direction information is used as the preload application direction, and the bolt preload is constructed based on the preload application direction. The bolt preload is used as the load boundary condition to participate in the synchronous superimposed calculation and generate the equivalent stress field and principal stress distribution.

[0065] In one implementation, see Figure 4 The diagram illustrates the axis of a bolt shank. In biaxial testing, changes in the boundary relationship driving state are crucial for accurately simulating the bending process. Due to the introduction of the driving mode, bolt preload is no longer a negligible factor. Therefore, this invention adds a preload setting component and improves upon the traditional interaction method. Traditional methods typically define preload by selecting the cylindrical surface of the bolt shank. However, in actual models, due to differences in CAD modeling formats or geometric segmentation issues, a complete cylindrical surface may be split into multiple sub-surfaces, resulting in inconsistent cylindrical surface features of the bolt shank. In this case, the number of faces selected by the user in the interaction layer no longer corresponds one-to-one with the actual number of bolts. The backend struggles to accurately create preload objects based on the number of faces, easily leading to mapping errors or preload creation failures.

[0066] To address the aforementioned issues, this invention starts from the essence of bolt preload setting, namely, obtaining the axial direction of the bolt shank. It innovatively introduces a body selection interaction method, allowing users to directly select the solid structure of the bolt shank entity instead of a cylindrical surface. Because this solid structure is topologically unique and unaffected by surface splitting, the number of bodies selected by the user always matches the number of bolts, thus ensuring a stable mapping between the interaction layer and the backend logic. The backend system can automatically generate a corresponding number of preload objects based on the number of selected bolt shank bodies, avoiding recognition errors caused by inconsistent geometric features and improving the robustness and reliability of the interaction. In terms of algorithm implementation, the process of obtaining the axis by selecting a cylinder can be implemented using the following steps: (2-1) Entity type recognition: Perform geometric classification on the user-selected entity and determine whether the entity is a cylindrical (or approximately cylindrical) entity by parsing B-Rep data. Preliminary recognition can be achieved by detecting whether the side surface is a surface of revolution with equal radius.

[0067] (2-2) Side surface extraction: Extract all sides from the solid topology and filter out the set of surfaces with constant curvature and normal variation that conform to the characteristics of a cylinder. If there are multiple split surfaces, merge them.

[0068] (2-3) Axis Fitting: Solve for the axis of the selected cylindrical side surface. Based on analytical geometry: Directly read the cylindrical surface parameters (center point, axial vector) provided by the geometric kernel.

[0069] (2-4) Axial consistency verification: If an entity contains multiple cylindrical surfaces (such as chamfers or steps), the candidate axes need to be judged for directional consistency, and the axis corresponding to the cylindrical surface with the largest length or area is selected as the bolt rod axis.

[0070] (2-5) Preload creation: The obtained axial direction is used as the direction of preload application, and the bolt preload object is generated with the center of the solid end face or the midpoint of the axis as the reference position.

[0071] Using the above method, the system can reliably extract the bolt rod axis from the body selection interaction, which not only avoids the identification problem caused by the split cylindrical surface, but also improves the automation and accuracy of preload setting under complex model conditions.

[0072] (3) The fatigue life calculation is performed based on the equivalent stress field and principal stress distribution to obtain the wheel dual-axle performance evaluation result. In a real-time mode, the standardized road spectrum input parameter information can be obtained and the standardized road spectrum input parameter information is decomposed to generate various sub-working condition information. Then, the equivalent stress field and principal stress distribution corresponding to each time step in the various sub-working condition information are arranged in time order to obtain the stress time history information corresponding to each sub-working condition information. Finally, the stress time history information is used for cyclic identification and damage accumulation processing to generate comprehensive accumulated damage information and wheel predicted fatigue life information. The comprehensive accumulated damage information and wheel predicted fatigue life information are used as the wheel dual-axle performance evaluation result.

[0073] In one implementation, the interactive phase of the entire biaxial testing system includes the wheel flange mounting contact surface, the rim edge surface, and the tire bead contact compression area. The system involves multiple load boundary components, some of which need to share an interactive surface to avoid duplicate selections in the user interface. Here, shared parameter and shared surface setting components are designed, and a mapping and association mechanism between load components and common components is established, allowing user interface settings to be synchronized to various load or boundary setting components through a unified common component. To meet biaxial testing conditions, a remote force component with built-in periodic functions is designed, and an offset function is provided for setting the remote point of the basic conditions for the remote force. To simulate the real state of the drive shaft, a component setting function is provided to offset the remote point according to the shaft length. After the remote point is fitted according to the geometric center of the reference interactive surface information, it is simultaneously offset according to the offset value to simulate the bending moment application effect of the drive shaft.

[0074] The radial rolling condition is simulated by transforming the real rolling contact problem into a time-varying contact pressure field. A periodic radial load function is constructed in the time domain, causing the pressure in the contact area to move circumferentially over time, thus equivalently reflecting the force characteristics of the wheel under rolling conditions. After completing the construction of a single biaxial coupling model, a road spectrum driving mechanism is further established. The system defines standardized road spectrum input parameters, including load amplitude distribution, frequency characteristics, number of cycles, impact amplification factor, and road condition scaling factor. The system decomposes the comprehensive road spectrum into several representative sub-conditions, each corresponding to a specific load amplitude range and number of cycles, and automatically calls a unified biaxial simulation algorithm for batch solving. The solution output includes the equivalent stress field and principal stress distribution, thereby realizing automated post-processing output under multiple conditions. Through a unified data interface, all single-condition calculation results are stored in a structured manner, providing standard input for subsequent fatigue life calculations.

[0075] In the life assessment phase, this invention constructs a fatigue analysis calculation process based on stress time history. The engineering fatigue analysis system adopted achieves automated calculation on the web end through a built-in modular algorithm. This phase uses the stress time history of critical nodes obtained from biaxial coupling simulation as input, and completes life prediction through three levels: cyclic identification, damage calculation, and multi-condition comprehensive accumulation.

[0076] First, the equivalent stress or principal stress time history obtained from the finite element method is... Preprocessing is performed. The system extracts peak and valley values ​​from discrete time points to form a stress inflection point sequence. To reduce computational redundancy and maintain the integrity of the stress history, a rainflow counting algorithm was then used to identify the inflection point sequence in cycles. The basic idea of ​​rainflow counting is to decompose the complex stress waveform into several closed stress cycles, each cycle containing stress amplitude values. with the mean Its definition is:

[0077] in, This represents the maximum stress value in a single cycle. This represents the minimum stress value in a single cycle. This refers to the stress amplitude. This represents the average stress value.

[0078] To address the influence of mean stress, this invention introduces a mean correction model. Using the Goodman correction formula, the actual stress amplitude is converted into an equivalent fully symmetric cyclic stress amplitude:

[0079] in, This is the equivalent stress amplitude; The tensile strength limit of the material; This represents the average value of cyclic stress.

[0080] After mean correction, the system calculates the corresponding fatigue life based on the material's SN curve (stress-life curve). The SN curve is typically represented using the Basquin expression as follows:

[0081] After adjustment, the cycle life is obtained:

[0082] in, This represents the number of cycles that the system can withstand under this stress amplitude. is the fatigue strength coefficient; b is the fatigue strength index.

[0083] The fatigue strength coefficient and fatigue strength index here are fitted using material fatigue test data: multiple fatigue tests with different stress amplitudes are conducted to obtain the corresponding life. Perform linear fitting in a log-log coordinate system; the slope is b; the intercept is obtained by inverse calculation. .

[0084]

[0085] For low-cycle fatigue conditions involving ductile materials, the Coffin-Manson strain-life model is adopted:

[0086] in, This is the equivalent amplitude; is the fatigue ductility coefficient; c: fatigue ductility index (negative value).

[0087] In multi-path, multi-condition biaxial wheel tests, the load task corresponding to the target mileage of the vehicle is decomposed into several condition parameter sets. The fatigue damage to the wheel is calculated for each condition parameter set, and the results are weighted and accumulated based on the test mileage of each condition to obtain the predicted fatigue life of the wheel's critical parts. The condition parameter set is used to uniquely determine the boundary conditions and load input for a single biaxial simulation calculation. It includes at least the lateral load coefficient, axial load coefficient, road friction coefficient, and the test mileage corresponding to that condition. The lateral load coefficient and axial load coefficient are used to calibrate the amplitude of the biaxial load input. The road friction coefficient is used to determine the contact conditions and affect the tangential force and constraint state under the tire-road or equivalent contact boundary. The test mileage is used to characterize the occurrence length of that condition in the total target mileage and serves as the weight for damage aggregation. Each set of operating parameters has reproducible parameter values ​​and corresponding mileages, and the sum of all operating mileages equals the total target mileage; at the same time, each set of operating parameters is bound to its dual-channel load input data, which includes the time history and phase relationship of radial and axial loads, or equivalent synchronous load sequences, thereby ensuring that the biaxial coupling characteristics remain consistent in simulation calculation and fatigue assessment.

[0088] This invention transforms the full-condition problem into a single-condition traversal calculation and unified post-processing summary: For each condition parameter set in the condition list, a single-condition simulation task is constructed and submitted. The input of the simulation task only includes the lateral load coefficient, axial load coefficient, road friction coefficient, and their associated dual-channel load input data corresponding to that condition parameter set. The solution is executed under the same wheel finite element model and the same contact / constraint modeling scheme, and the response time history of the key areas of the wheel under that condition is output. In the fatigue assessment stage, the fatigue evaluation method, cycle counting method, material fatigue curve, and damage accumulation criterion are fixed, so that the damage results of different conditions can be directly accumulated under the same criterion system.

[0089] In terms of specific algorithm implementation, the main tasks included nominal stress (with direction) history extraction, stress history inflection point extraction, Rainflow loop counting, loop feature definition, stress correction, single-cycle lifetime calculation, and Miner cumulative damage calculation.

[0090] Algorithm input: Signed stress of the element Process (multi-frame):

[0091] Step 1: Inflection Point Extraction. Objective: To extract local peaks and troughs from the original stress history and remove monotonic points that are meaningless for fatigue cycle counting. Input: Original stress sequence: Output the inflection point sequence: .

[0092] Each of them (The horizontal axis of the stress amplitude graph represents the frame number, which is a dimensionless parameter) satisfies one of the following: the first and last boundary points, i.e. , Local peak values: Local valley value: .

[0093] For example, suppose the original stress history of a node is as follows: [210, 50, -170, 40, 110, 10, -90] After extracting the inflection point, we get:

[0094] Step 2: Rainflow Cycle Counting. Objective: To decompose the inflection point sequence into several fatigue cycles. Each cycle includes: maximum value; minimum value; stress range; stress amplitude; mean stress; cycle weight (full cycle / half cycle); and full cycle definition.

[0095] During the Rainflow stack judgment process, if the current four consecutive turning points are: , , , .

[0096] Define three ranges:

[0097] If the following conditions are met: and

[0098] Then, the intermediate fluctuations: ( , This can be considered as a closed loop, and the value is removed from the stack after being output.

[0099] Semi-loop definition: Any remaining loop in the Rainflow stack that cannot be closed into a full loop is defined as a semi-loop.

[0100] Loop parameter definition: For the i-th identified loop, let its maximum and minimum values ​​be: ,

[0101] Therefore, the stress range is:

[0102] Stress amplitude:

[0103] Mean stress:

[0104] The number of iterations is weighted into full loop and semi loop.

[0105] Full cycle:

[0106] Half-cycle:

[0107] In one implementation, the inflection point of step one is continued:

[0108] Rainflow results can be written as: Loop 1, full loop: (-170, 110)

[0109]

[0110]

[0111]

[0112] Loop 2, half loop: (210, -170)

[0113] Loop 3, half loop: (110, -90)

[0114] Step 3: Goodman Stress Correction. Objective: To consider the effect of mean stress on fatigue life and correct the original stress amplitude for each cycle to the equivalent stress amplitude. Input: For the i-th cycle, known: Stress amplitude: Mean stress: Material tensile strength: .

[0115] Goodman's modified formula. The modified equivalent stress amplitude is defined as:

[0116] Step 4: SN calculates single-cycle lifetime.

[0117] Step 5: Miner Accumulates Damage. Objective: To accumulate multiple cyclic damages of different magnitudes to obtain the total damage and corresponding lifetime of the process block. Single Cycle Damage: For the i-th cycle, its damage is defined as:

[0118] The total damage is:

[0119] Total lifetime: If the current process block repeats, then the block lifetime is:

[0120] After each single-condition simulation is completed, for each evaluation location under that condition, an evaluation quantity time history for fatigue calculation is constructed from the biaxial response time history. Cyclic identification and counting are performed on the evaluation quantity time history to form a cyclic spectrum for that condition at that evaluation location. Subsequently, the cyclic spectrum is mapped to damage values ​​based on material fatigue performance data, and the damage values ​​are normalized according to the mileage of the condition to obtain the unit mileage damage for that condition at that evaluation location. The unit mileage damage is defined as: the amount of fatigue damage consumed at that evaluation location per unit mileage traveled by the vehicle under the load coefficient and friction coefficient specified in the parameter set of that condition. After completing the unit mileage damage calculation for all condition parameter sets, a deterministic weighted summary is performed on each condition according to the mileage allocation table: for each evaluation location, the unit mileage damage of each condition is multiplied by the test mileage of that condition to obtain the damage contribution for that mileage segment; then, the damage contributions of all conditions are summed to obtain the cumulative damage at that evaluation location under the total target mileage. Using the cumulative damage reaching a preset failure threshold as the life criterion, the fatigue life of the evaluation location is obtained by back-calculating the total target mileage or the number of times the bench test program can be repeated. The wheel with the shortest life is selected from the set of evaluation locations as the predicted fatigue life of the wheel, and the set of working condition parameters with the greatest damage contribution is output at the same time to clarify the combination of lateral load coefficient, axial load coefficient and road friction coefficient that dominate the wheel life.

[0121] The overall approach technically comprises a three-stage closed-loop process: unified dual-axis algorithm construction, road spectrum-driven multi-condition decomposition, and comprehensive fatigue damage assessment. This method requires no complex pre-processing software; it relies solely on basic web-based 3D interaction to complete dual-axis coupled modeling and solving. Built-in algorithms ensure computational consistency and engineering reliability, effectively replacing traditional high-cost physical dual-axis testing equipment and enabling rapid verification and life prediction of wheel products during the design phase.

[0122] In practical applications, the intelligent system for dual-axle performance evaluation of wheels first automatically calls the component orchestration algorithm according to the requirements of the dual-axle performance evaluation task. It then associates and builds the geometric recognition component, coordinate establishment component, material library component, remote force component, pressure component, road spectrum processing component, mesh generation component, shared parameter component, solution calculation component, report component, and fatigue life evaluation component to form a complete cloud simulation application for dual-axle wheels, and generates the corresponding interactive interface on the web page.

[0123] Users conducted a biaxial performance evaluation based on an aluminum alloy passenger car wheel. First, the user uploaded the wheel's 3D model to the intelligent system for biaxial performance evaluation via a web interface. The system automatically parsed the model format and performed lightweight rendering, displaying interactive 3D geometry in the browser. The user then selected the flange mounting surface area, radial load area, rim edge area, and tire pressure area through basic 3D interactive operations. The system automatically identified the wheel's axis of rotation based on the selected areas and established a local polar coordinate system, simultaneously generating a biaxial equivalent loading model coupling radial rolling and axial bending. Subsequently, the user input basic parameters such as radial load amplitude, bending moment range, and loading frequency in the simulation condition settings interface. The system automatically generated a uniform time-step load time history based on a preset biaxial test algorithm.

[0124] During the road spectrum input phase, users upload load spectrum data, including deflection angle, radial load coefficient, lateral load coefficient, and test distance. The system standardizes each road spectrum and automatically maps it to a combined radial load and bending moment time history. The backend adjusts the remote point position of the bending moment corresponding to the deflection angle based on the algorithm, and sets the usage ratio for each road condition. After completing the load definition, users select the corresponding aluminum alloy material from the material library and assign it to the wheel model. The system automatically reads the material's elastic parameters and fatigue parameters for subsequent life calculations.

[0125] During the background computation phase, the intelligent agent system automatically invokes the mesh generation component to perform adaptive finite element discretization on the wheel model, focusing on local refinement of the spoke transition area and the rim root, and generating a computational mesh that meets the accuracy requirements. Subsequently, the system executes a coupled solution algorithm to simultaneously calculate radial contact loads and bending loads in a unified time domain, obtaining stress and strain time history results for key areas. After the calculation is completed, the system automatically generates stress contour maps, deformation contour maps, and single-condition fatigue life data, and provides interactive post-processing analysis functions on the web interface.

[0126] During the fatigue life assessment stage, the system performs cycle counting and single-cycle damage calculation on the stress time history under different road spectrum conditions, and performs weighted accumulation according to the usage ratio of each road condition, outputting the life of each single road condition and the comprehensive cumulative damage results, and finally obtaining the equivalent design life of the wheel under comprehensive road spectrum conditions. At the same time, it provides the most dangerous area of ​​life and optimization suggestions, thus completing the complete dual-axle performance assessment process of the wheel.

[0127] In summary, this invention, while achieving automation of the simulation process and consistency of results, ensures the authenticity of dual-axis coupling through synchronous superposition calculation at a unified time step, and realizes comprehensive fatigue life prediction under multiple working conditions by combining road spectrum driving, effectively overcoming many defects existing in the prior art.

[0128] Specifically, this invention enables an intelligent agent to automatically call components from a functional component library, associate and build a simulation application process, and obtain available supercomputing resources, a queue of components to be executed, and component dependency information to form a scheduling state. By using a reinforcement learning model to generate the execution order of target components, the invention achieves automated construction and intelligent scheduling of the simulation process, avoids the uncertainty caused by manual operation, and significantly improves simulation efficiency and result consistency.

[0129] Furthermore, this invention performs synchronous superposition calculations on the radial pressure field and bending load within the same time integration step to generate an equivalent stress field and principal stress distribution, thereby achieving true coupling of biaxial loads. The output stress field can accurately reflect the stress change process under actual biaxial test conditions, providing accurate input for fatigue life prediction.

[0130] Finally, this invention obtains standardized road spectrum input parameter information and decomposes it into multiple sub-working condition information. The equivalent stress field and principal stress distribution corresponding to each time step in each sub-working condition are arranged in chronological order to obtain stress time history information. Then, cyclic identification and damage accumulation processing are performed to generate comprehensive accumulated damage information and wheel predicted fatigue life information, thereby realizing comprehensive fatigue life prediction under multiple working conditions, which can more accurately evaluate the durability performance of wheels under actual road use conditions.

[0131] Regarding the numerical calculation-based dual-axle performance evaluation method for wheels provided in the foregoing embodiments, this invention provides a numerical calculation-based dual-axle performance evaluation device for wheels, see [link to previous document]. Figure 5 The diagram shows a structural schematic of a wheel dual-axle performance evaluation device based on numerical calculation. The device includes the following parts: The component association module 502 obtains the three-dimensional model of the wheel and the dual-axle performance evaluation requirements of the wheel. Based on the dual-axle performance evaluation requirements of the wheel, it calls the corresponding functional components from the preset functional component library to associate and build the various functional components, and generates the dual-axle simulation application process of the wheel corresponding to the dual-axle performance evaluation requirements. The reinforcement learning module 504 obtains the available supercomputing resources at the current moment, and combines the available supercomputing resources with the queue information of the components to be executed and the component dependency information corresponding to the dual-axle wheel simulation application process to form scheduling state information. The reinforcement learning model then processes the scheduling state information to generate the execution order of the target components. The dual-axis performance evaluation module 506 calls various functional components according to the execution order of the target components, generates a dual-axis equivalent loading model based on the wheel 3D model, and performs dual-axis performance evaluation processing on the dual-axis equivalent loading model to obtain the wheel dual-axis performance evaluation results.

[0132] The numerical calculation-based dual-axle performance evaluation device for wheels provided in this application embodiment can significantly improve the efficiency and accuracy of dual-axle performance evaluation.

[0133] In one embodiment, during the step of calling various functional components according to the execution order of the target components and generating a biaxial equivalent loading model based on the wheel 3D model, the aforementioned biaxial performance evaluation module 506 is further configured to: render the wheel 3D model to obtain an interactive 3D model, and determine the flange mounting surface area information, rim edge area information, and bead contact area information based on the user's interaction information with the interactive 3D model; and through the geometric recognition component and coordinate establishment component, automatically identify the wheel rotation axis based on the flange mounting surface area information, rim edge area information, bead contact area information, and the execution order of the target components, and establish a local polar coordinate system to generate a biaxial equivalent loading model.

[0134] In one embodiment, after performing the step of generating a biaxial equivalent loading model, the biaxial performance evaluation module 506 is further configured to: set the flange mounting surface area corresponding to the flange mounting surface area information as a fixed area by calling the boundary condition construction component, and constrain the translational and rotational degrees of freedom within the fixed area to generate boundary constraint conditions, wherein the boundary constraint conditions are used to simulate the force state of the wheel connected to the axle through the flange.

[0135] In one embodiment, when performing biaxial performance evaluation processing on the biaxial equivalent loading model to obtain the biaxial performance evaluation result of the wheel, the aforementioned biaxial performance evaluation module 506 is further configured to: acquire the radial pressure field and bending load, and simultaneously perform superposition calculation processing on the radial pressure field and bending load within the same time integration step to generate the equivalent stress field and principal stress distribution; perform fatigue life calculation processing based on the equivalent stress field and principal stress distribution to obtain the biaxial performance evaluation result of the wheel.

[0136] In one embodiment, when performing the step of obtaining bending load, the biaxial performance evaluation module 506 is further configured to: apply a bending moment that varies with time to the rim edge region using a reverse reaction force loading model; convert the bending moment into an equivalent distributed force based on the bending moment that varies with time and the rim radius information, and apply the equivalent distributed force to the rim edge node to generate a bending load that is uniformly distributed along the rim circumference.

[0137] In one embodiment, before performing the synchronous superposition calculation of the radial pressure field and bending load to generate the equivalent stress field and principal stress distribution, the biaxial performance evaluation module 506 is further configured to: acquire the body structure information of the wheel bolt rod, and perform geometric classification and recognition processing on the body structure information to obtain the cylinder recognition result; extract the cylinder side surface information from the body structure information based on the cylinder recognition result, and perform axis fitting processing on the cylinder side surface information to obtain the bolt rod axis direction information; use the bolt rod axis direction information as the preload application direction, and construct the bolt preload based on the preload application direction, so as to use the bolt preload as a load boundary condition to participate in the synchronous superposition calculation processing to generate the equivalent stress field and principal stress distribution.

[0138] In one embodiment, when performing fatigue life calculation based on the equivalent stress field and principal stress distribution to obtain the dual-axle performance evaluation result of the wheel, the dual-axle performance evaluation module 506 is further configured to: acquire standardized road spectrum input parameter information, decompose the standardized road spectrum input parameter information to generate various sub-working condition information; arrange the equivalent stress field and principal stress distribution corresponding to each time step in the various sub-working condition information in chronological order to obtain the stress time history information corresponding to each sub-working condition information; perform cyclic identification and damage accumulation processing based on the stress time history information to generate comprehensive accumulated damage information and wheel predicted fatigue life information, and use the comprehensive accumulated damage information and wheel predicted fatigue life information as the dual-axle performance evaluation result of the wheel.

[0139] The device provided in this embodiment of the invention has the same implementation principle and technical effect as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the device embodiment can be referred to the corresponding content in the aforementioned method embodiment.

[0140] This invention provides a server, specifically, the server includes a processor and a storage device; the storage device stores a computer program, which, when run by the processor, executes the method described in any of the above embodiments.

[0141] Figure 6 This is a schematic diagram of the structure of a server provided in an embodiment of the present invention. The server 100 includes: a processor 60, a memory 61, a bus 62, and a communication interface 63. The processor 60, the communication interface 63, and the memory 61 are connected through the bus 62. The processor 60 is used to execute executable modules, such as computer programs, stored in the memory 61.

[0142] The memory 61 may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 63 (which can be wired or wireless), such as the Internet, wide area network, local area network, metropolitan area network, etc.

[0143] Bus 62 can be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 6 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0144] The memory 61 is used to store programs. After receiving an execution instruction, the processor 60 executes the program. The method executed by the device for defining the flow process disclosed in any of the foregoing embodiments of the present invention can be applied to the processor 60 or implemented by the processor 60.

[0145] Processor 60 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of processor 60 or by instructions in software form. Processor 60 can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 61. Processor 60 reads the information in memory 61 and, in conjunction with its hardware, completes the steps of the above method.

[0146] The computer program product of the readable storage medium provided in the embodiments of the present invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the foregoing method embodiments. For specific implementation, please refer to the foregoing method embodiments, which will not be repeated here.

[0147] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0148] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, 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, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for evaluating the dual-axle performance of wheels based on numerical calculation, characterized in that, The method includes: Obtain the 3D model of the wheel and the dual-axle performance evaluation requirements of the wheel. Based on the dual-axle performance evaluation requirements of the wheel, call the corresponding functional components from the preset functional component library to associate and build each of the functional components, and generate the dual-axle simulation application process of the wheel corresponding to the dual-axle performance evaluation requirements of the wheel. Obtain the available supercomputing resources at the current moment, and combine the available supercomputing resources with the queue information of the components to be executed and the component dependency information corresponding to the dual-axle wheel simulation application process to form scheduling status information. Then, use a reinforcement learning model to make decisions on the scheduling status information and generate the execution order of the target components. According to the execution order of the target component, each of the functional components is called. Based on the three-dimensional model of the wheel, a dual-axis equivalent loading model is generated, and the dual-axis equivalent loading model is subjected to dual-axis performance evaluation processing to obtain the dual-axis performance evaluation result of the wheel.

2. The method for evaluating the dual-axle performance of wheels based on numerical calculation according to claim 1, characterized in that, The step of calling each of the functional components according to the execution order of the target component, and generating a dual-axis equivalent loading model based on the wheel 3D model, includes: The wheel 3D model is rendered to obtain an interactive 3D model. Based on the user's interaction information with the interactive 3D model, the flange mounting surface area information, rim edge area information, and tire bead contact area information are determined. Using a geometric recognition component and a coordinate establishment component, based on the flange mounting surface area information, the rim edge area information, the tire bead contact area information, and the execution sequence of the target component, the wheel rotation axis is automatically identified and a local polar coordinate system is established to generate the dual-axis equivalent loading model.

3. The method for evaluating the dual-axle performance of wheels based on numerical calculation according to claim 2, characterized in that, Following the step of generating the biaxial equivalent loading model, the following is included: By calling the boundary condition construction component, the flange mounting surface area corresponding to the flange mounting surface area information is set as a fixed area, and the translational and rotational degrees of freedom are constrained within the fixed area to generate boundary constraint conditions. The boundary constraint conditions are used to simulate the force state of the wheel connected to the axle through the flange.

4. The method for evaluating the dual-axle performance of wheels based on numerical calculation according to claim 1, characterized in that, The step of performing biaxial performance evaluation processing on the biaxial equivalent loading model to obtain the biaxial performance evaluation results of the wheel includes: The radial pressure field and bending load are obtained, and the radial pressure field and bending load are simultaneously superimposed and calculated within the same time integration step to generate the equivalent stress field and principal stress distribution. The fatigue life calculation is performed based on the equivalent stress field and the principal stress distribution to obtain the performance evaluation results of the dual axle of the wheel.

5. The method for evaluating the dual-axle performance of wheels based on numerical calculation according to claim 4, characterized in that, The steps for obtaining bending loads include: By using a reverse reaction force loading model, a time-varying bending moment is applied to the edge region of the wheel rim; Based on the time-varying bending moment and rim radius information, the bending moment is converted into an equivalent distributed force, and the equivalent distributed force is applied to the rim edge node to generate the bending load uniformly distributed along the rim circumference.

6. The method for evaluating the dual-axle performance of wheels based on numerical calculation according to claim 4, characterized in that, Before the step of simultaneously superimposing the radial pressure field and the bending load to generate the equivalent stress field and principal stress distribution, the following steps are included: Obtain the body structure information of the wheel bolt rod, and perform geometric classification and recognition processing on the body structure information to obtain the cylinder recognition result; Based on the cylinder recognition result, the cylinder side surface information is extracted from the body structure information, and the cylinder side surface information is subjected to axis fitting processing to obtain the bolt rod axis direction information; The bolt shank axis direction information is used as the preload application direction, and the bolt preload is constructed based on the preload application direction. The bolt preload is then used as a load boundary condition in the synchronous superposition calculation process to generate the equivalent stress field and the principal stress distribution.

7. The method for evaluating the dual-axle performance of wheels based on numerical calculation according to claim 4, characterized in that, The step of calculating fatigue life based on the equivalent stress field and the principal stress distribution to obtain the dual-axle performance evaluation result of the wheel includes: Obtain standardized road spectrum input parameter information, and decompose the standardized road spectrum input parameter information to generate various sub-working condition information; The equivalent stress field and principal stress distribution corresponding to each time step in each of the sub-working condition information are arranged in chronological order to obtain the stress time history information corresponding to each of the sub-working condition information. Based on the stress time history information, cyclic identification and damage accumulation processing are performed to generate comprehensive cumulative damage information and wheel predicted fatigue life information. The comprehensive cumulative damage information and wheel predicted fatigue life information are then used as the wheel dual-axle performance evaluation results.

8. A wheel dual-axle performance evaluation device based on numerical calculation, characterized in that, The device includes: The component association module obtains the 3D model of the wheel and the dual-axle performance evaluation requirements of the wheel. Based on the dual-axle performance evaluation requirements of the wheel, it calls the corresponding functional components from the preset functional component library to associate and build each of the functional components, and generates the dual-axle simulation application process of the wheel corresponding to the dual-axle performance evaluation requirements of the wheel. The reinforcement learning module acquires the available supercomputing resources at the current moment, and combines the available supercomputing resources with the queue information of the components to be executed and the component dependency information corresponding to the dual-axle wheel simulation application process to form scheduling state information. The reinforcement learning model then performs decision processing on the scheduling state information to generate the execution order of the target components. The dual-axis performance evaluation module calls each of the functional components according to the execution order of the target components, generates a dual-axis equivalent loading model based on the wheel 3D model, and performs dual-axis performance evaluation processing on the dual-axis equivalent loading model to obtain the wheel dual-axis performance evaluation result.

9. A server, characterized in that, The method includes a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when invoked and executed by a processor, cause the processor to perform the method according to any one of claims 1 to 7.

Images