A lightweight design method for variable thickness frame structure of heavy vehicles

Abstract

The present disclosure relates to the field of vehicle frame structure design, and specifically to a lightweight design method for a variable-thickness frame structure of a heavy-duty vehicle. The method comprises the following steps: establishing a finite element model of the frame, defining the following eight variables in the dimensional parameters of the finite element model of the frame as optimization parameter design variables, and performing experimental design sampling. A prediction model for lightweight optimization design of the frame is established based on an improved radial basis function neural network method according to the simulation data to obtain an optimal prediction model. The optimization objective functions are to minimize the mass of the frame structure, maximize the fatigue life, and maximize the first-order bending modal frequency of the vehicle body. The maximum stress, maximum deformation, and first-order torsional modal frequency of the frame structure under full-load torsional working conditions are used as constraints. A multi-island genetic algorithm is used to optimize and solve the prediction model that meets the requirements to obtain an optimized design result. The present disclosure can reduce the structural mass while ensuring that the first-order modal frequency response of the frame meets the design requirements, while effectively improving the fatigue life.

Description

Technical Field

[0001] The present disclosure relates to the field of vehicle frame structure design, and in particular to a lightweight design method for a variable-thickness frame structure of a heavy-duty vehicle. Background Art

[0002] The frame of a heavy-duty vehicle, such as a fire truck, is a critical load-bearing component, supporting and connecting the cab, engine, transmission, and rescue equipment assembly. It also withstands various internal and external excitations. Fire and rescue vehicles carry heavy loads, navigate complex road conditions, and withstand significant loads and impacts during operation. Therefore, the frame structure must possess high rigidity, strength, and durability to reduce the risk of structural fatigue.

[0003] Fire truck frames are often subjected to bending, twisting, braking, and cornering during operation. Whether the structural strength and rigidity of the frame meet these requirements is crucial not only for the proper functioning of firefighting and rescue equipment, but also for the integrity and safety of firefighting equipment during operation. While existing frames have simple structures and advanced manufacturing processes, they lack the ability to withstand bending and twisting loads, resulting in high structural weight and raw material consumption, making them inefficient in achieving energy conservation and emission reduction.

[0004] Existing frame structure optimization designs primarily focus on single objectives, such as structural strength, stiffness, natural frequency, fatigue life, or lightweighting. These optimizations are achieved by varying the structure's variable thickness parameters. However, as a critical load-bearing component of the vehicle, a frame structure that only meets a single optimization objective cannot meet the reliability and safety requirements for driving and rescue operations. A search of relevant domestic and international literature has uncovered no similar lightweight design methods for variable thickness frames for heavy vehicles. Summary of the Invention

[0005] This disclosure proposes a lightweight design method for a variable-thickness frame structure for a heavy-duty vehicle, which can achieve the goal of reducing weight while maintaining strength, stiffness, and fatigue life. This disclosure provides a lightweight design method for a variable-thickness frame structure for a heavy-duty vehicle, comprising the following steps:

[0006] Step 1: Obtain a frame structure with known performance that meets the requirements for finite element modeling to obtain a frame finite element model. Based on the load borne by the frame structure under the full-load torsion working condition of the fire truck, calculate the first-order modal frequency, fatigue life, mass, maximum stress and maximum deformation of the frame structure in the free state, and compare and verify with the actual working condition performance test data of the known frame structure;

[0007] Step 2: Define the following eight variables in the dimensional parameters of the frame finite element model as optimization parameter design variables: the thickness of the upper wing panel, the middle panel, and the lower wing panel of the front cross member; the thickness of the upper wing panel, the thickness of the lower wing panel, and the middle panel of the side longitudinal member; and the thickness of the upper end panel and the lower end panel of the middle cross member; thus obtaining the frame finite element model of the variable thickness components;

[0008] Step 3: Performing experimental design sampling on the finite element model of the vehicle frame with variable thickness components, and after obtaining a number of simulation data, establishing a prediction model for the lightweight optimization design of the vehicle frame based on the improved radial basis function neural network method according to the simulation data; the improved radial basis function neural network method comprises: obtaining RBF neural network code; calculating the fitness value of the particles after initializing the particle swarm parameters; finding the historical optimal values ​​of individuals and groups, and updating the particle speed and position according to the historical optimal values; judging whether the parameters are optimal, and if so, obtaining the optimal prediction model; if not, continuing to calculate the fitness value of the particles to find the next historical optimal value;

[0009] Step 4: If the accuracy of the prediction model is verified to meet the requirements, then the minimum frame structure mass, the maximum fatigue life and the maximum first-order bending modal frequency of the car body are used as optimization objective functions, the maximum stress, the maximum deformation and the first-order torsional modal frequency of the frame structure under full-load torsional working conditions are used as constraints, and the eight variables are used as optimization parameter design variables. A multi-island genetic algorithm is used to optimize and solve the prediction model that meets the requirements, and a frame structure lightweight optimization design analysis model is established. After optimization iterative calculation of the frame structure lightweight optimization design analysis model, a Pareto frontier solution set of the frame lightweight optimization design is obtained, and the optimal solution in the Pareto frontier solution set is selected as the optimization design result;

[0010] Step 5: Perform engineering rounding on the lightweight optimization data in the optimization design result to obtain a rounded value, and assign the rounded value to the frame performance after the frame finite element model analysis and calculation optimization.

[0011] Preferably, the experimental design sampling is performed on the finite element model of the vehicle frame of the variable thickness component, and the step of obtaining a plurality of sample points specifically includes the following steps: determining the variation range of the optimization parameter design variable, setting the constraint conditions of the maximum stress and the maximum deformation, and taking the mass M as the minimum, the fatigue life F as the maximum and the first-order bending modal frequency f as the maximum. b The maximum is the optimization goal, and the Hammersley sampling method is used to carry out experimental design in the design space to obtain the experimental design matrix of 8 design variables and several sample points. Finite element calculation and analysis are performed on each group of sample points according to the modified size to obtain simulation data.

[0012] Preferably, after reassigning the rounded value to the frame performance after the frame finite element model analysis and calculation optimization, the following step is also included: judging whether the frame performance meets the performance requirements; if it does not meet the performance requirements, re-determining the range of change of the optimization parameter design variable and then executing step three.

[0013] Preferably, if the frame performance meets the performance requirements, a lightweight design scheme is determined, and a lightweight frame structure after lightweight optimization is trial-produced according to the parameter combination, and the lightweight frame structure is tested and verified in accordance with the requirements of full-load torsional working conditions, first-order modal and fatigue life analysis.

[0014] Preferably, the optimization objective function of the optimization objective is expressed as follows:

[0015] minM(x1,x2,...,x8);maxF(x1,x2,...,x8);maxf b (x1,x2,...,x8);

[0016] Where M is the frame mass, F is the frame fatigue life, and f b is the first-order bending mode frequency.

[0017] Preferably, the constraints for setting the maximum stress and maximum deformation are specifically as follows:

[0018] The maximum stress does not exceed 475MPa and the maximum deformation does not exceed 10mm. The constraint conditions are described as stσ≤475MPa; D≤10mm; f t ≥9.5Hz, where σ is the yield strength of the frame structure, D is the maximum allowable deformation of the frame, and f t is the first-order torsional mode frequency.

[0019] Preferably, the range of variation of the optimization parameter design variables is determined as follows: for the design variables "front crossbeam upper wing panel thickness x1, middle panel thickness x2, lower wing panel thickness x3, side longitudinal beam upper wing panel thickness x4, lower wing panel thickness x5 and middle panel thickness x6, middle crossbeam upper end panel thickness x7 and lower end panel thickness x8", they are specifically: x1∈[2.2,3], x2∈[1.8,2.4], x3∈[2.5,3], x4∈[1,2], x5∈[2.4,3.2], x6∈[2,3], x7∈[1.5,3], x8∈[3,4].

[0020] Preferably, the multi-island genetic algorithm is used to optimize and solve the prediction model that meets the requirements, specifically including: importing the above-mentioned prediction model after establishment, and then determining the subgroup size to be 10 and the total group size to be 200; the number of subgroups to be 20; the total number of generations of evolution to be 20; the crossover probability to be 0.9; the mutation probability to be 0.009; the inter-island migration rate to be 0.4; the migration interval generation number to be 5; and then setting the above parameters and performing iterative optimization calculations.

[0021] Preferably, the steps for obtaining the Pareto frontier solution set are as follows: in the steps of the multi-island genetic algorithm, the subgroup size is set to 20, the evolutionary generations are set to 60, the crossover probability is set to 0.9, and the mutation probability is set to 0.09; after optimized iterative calculation, the Pareto frontier solution set of the lightweight optimization design of the frame is obtained.

[0022] Preferably, the plurality of sample points is 40 sample points.

[0023] Compared with the prior art, the present disclosure has the following beneficial effects:

[0024] 1. The present invention is simple to operate and has a reasonable process. Compared with the existing frame structure, the frame structure after lightweight optimization design has a greater weight reduction while meeting the performance requirements, thereby maximizing the material utilization rate. It is suitable for the development direction of lightweight heavy vehicle frames such as fire truck frames;

[0025] 2. Comprehensive consideration is given to the requirements of full-load bending, torsion, braking and cornering conditions, as well as objective functions such as fatigue life, mass and first-order bending modal frequency response. This can not only reduce structural mass but also ensure that the first-order modal frequency response of the frame meets the design requirements, while effectively improving fatigue life. Therefore, the consideration is more comprehensive. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 Flowchart of lightweight design method for variable thickness frame structure of heavy vehicles;

[0027] Figure 2 It is a three-dimensional solid model diagram of the frame structure;

[0028] Figure 3 It is an isometric view of the front crossbeam structure of the vehicle frame structure;

[0029] Figure 4 It is an isometric view of the side longitudinal beam structure of the frame structure;

[0030] Figure 5 It is an isometric view of the middle crossbeam of the frame structure;

[0031] Figure 6 It is the isometric view of the finite element model of the frame structure;

[0032] Figure 7 This is the flow chart of the improved radial basis neural network method;

[0033] Figure 8 This is the accuracy test result diagram of the quality response prediction model;

[0034] Figure 9 The accuracy test results of the fatigue life response prediction model

[0035] Figure 10 This is the test result of the first-order modal frequency response prediction model.

[0036] Figure 11 This is the test result of the maximum deformation response prediction model.

[0037] Figure 12 This is the accuracy test result of the maximum stress response prediction model.

[0038] Figure 13 is the Pareto frontier solution set of the frame structure.

[0039] Figure 14 Isometric view of the model after lightweight optimization of the frame structure. DETAILED DESCRIPTION

[0040] The following will be combined with the accompanying drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Obviously, the embodiments described are only part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present disclosure.

[0041] The present disclosure provides a lightweight design method for a variable thickness frame structure of a heavy vehicle. The following description is given using a fire truck frame as an example, but those skilled in the art will not limit it to a fire truck frame. Figure 1 As shown, the following steps are included:

[0042] Step 1: Obtain a frame structure with known performance that meets the requirements for finite element modeling to obtain a frame finite element model. Based on the load borne by the frame structure under the full-load torsion working condition of the fire truck, calculate the first-order modal frequency, fatigue life, mass, maximum stress and maximum deformation of the frame structure in the free state, and compare and verify with the actual working condition performance test data of the known frame structure;

[0043] In this embodiment, Hypermesh is used as the software platform, and fatigue in Nastran and Hypermesh is used as the solver. Figure 2As shown in the figure, based on the known geometric dimensions of the frame structure, CATIA is used to establish a geometric model, and the geometric model is imported into the finite element software Hypermesh. The three-dimensional solid model of the frame structure is structurally simplified, and the fillets, holes and small parts that have little effect on the performance of the frame are ignored. The fine lines and surfaces of the initial model that have little effect on the performance are removed. The unit parameters such as mm, S, t, MPa, etc. are set in Hypermesh, and the mid-surfaces of the parts divided by shell elements are extracted. The entire model is divided into shell element meshes, and the mesh size of the entire model is 20 mm, thereby establishing a model suitable for finite element analysis.

[0044] like Figure 3 、 4 As shown in Figures 5 and 6, a finite element model of a fire truck frame is established. The model includes 203,185 units and 195,320 nodes, including 14,900 rigid units. Among them, there are 3,460 triangular units in the shell units (accounting for 1.7% of the number of model units). Then, a mesh quality check is performed. The failure units of the Jacobian, distortion, and warping angle in the model are all less than 0.2%, proving that the model accuracy is reliable.

[0045] Directly input the parameters of the materials and properties involved in the frame structure into this effective and reasonable finite element model, and set the parameters such as elastic modulus, Poisson's ratio, density and thickness in the model. Note that the dimensions must be consistent with the initial solid model. The material is high-strength steel; as a preferred method, the elastic modulus is set to 2.1×10 5 MPa, density is 7.85×10 -9 t / mm 3 , Poisson's ratio is 0.3. The connections between different parts of the model are simulated by common nodes or spot welding, seam welding and bolt connection according to actual conditions.

[0046] The finite element model of the vehicle frame was used as the research object. Constraints and loading were set according to the actual operating conditions of the vehicle frame. Since the fire truck frame is constantly subjected to random loads such as road impact excitation during driving, the dynamic load acting on the frame under different operating conditions must be considered. The dynamic load value is usually calculated by multiplying the static load and the dynamic load coefficient. When the vehicle is in a fully loaded torsional condition, the dynamic load coefficient is 1.2.

[0047] When calculating the first-order modal frequency of the frame in the free state, the card parameter PARAM is set, and EIGRL is set according to the requirements of the calculated modal frequency, and the first-order modal frequency except the rigid mode is obtained.

[0048] The frame fatigue life analysis was performed using the nominal stress method of high-cycle fatigue analysis. In Hypermesh, a vertical upward unit load was applied to each frame connection point. Sixteen load steps were set to correspond to the load-time history of each frame connection point. Virtual constraints were added, and the resulting calculation file was imported into fatigue software for fatigue life analysis. The material's SN curve was also determined based on the material properties. After setting the aforementioned parameters, the Goodman method's mean stress correction method was used, with a survival rate set at 96%. The final result represents the cumulative damage to the frame under the entire load spectrum, the reciprocal of which represents the minimum fatigue life of the frame.

[0049] The final performance indicators to be extracted include: maximum stress and maximum deformation under full-load torsion conditions, first-order modal frequency in the free state, fatigue life, and mass. The calculated first-order modal frequency performance is compared with an actual frame structure model that is known to meet the performance requirements and is currently in use in engineering. This actual structure is then subjected to experimental analysis, using the hammer method to obtain the first-order modal frequency in the free state of the frame structure. If the simulation data is consistent with the experimental data, thus verifying the model's rationality, the next step can be carried out. The initial model simulation results are shown in Table 1. Otherwise, the finite element model needs to be re-examined and revised.

[0050] Table 1 Initial model simulation results

[0051]

[0052] Step 2: Define the following eight variables in the dimensional parameters of the frame finite element model as optimization parameter design variables: the thickness of the upper wing panel of the front cross member x1, the thickness of the middle panel x2, and the thickness of the lower wing panel x3; the thickness of the upper wing panel of the side longitudinal member x4, the thickness of the lower wing panel x5, and the thickness of the middle panel x6; the thickness of the upper end panel x7 and the thickness of the lower end panel x8 of the middle cross member; and obtain the finite element model of the frame with variable thickness components.

[0053] Step 3: Performing experimental design sampling on the finite element model of the vehicle frame with variable thickness components, and after obtaining a number of simulation data, establishing a prediction model for the lightweight optimization design of the vehicle frame based on the improved radial basis function neural network method according to the simulation data; Figure 7 The figure shows the flow chart of the improved radial basis function neural network method: obtaining the RBF neural network code; calculating the fitness value of the particle after initializing the particle swarm parameters; finding the historical optimal values ​​of the individual and group, and updating the particle speed and position according to the historical optimal values; judging whether the parameters are the optimal values, if so, obtaining the optimal prediction model; if not, continuing to calculate the fitness value of the particle to find the next historical optimal value;

[0054] In step three, a design space is defined for the parameters of the design variables "front crossbeam upper wing panel thickness x1, middle panel thickness x2, lower wing panel thickness x3; side longitudinal beam upper wing panel thickness x4, lower wing panel thickness x5 and middle panel thickness x6; middle crossbeam upper end panel thickness x7 and lower end panel thickness x8". In this embodiment, as a preferred embodiment, the range of variation of the optimized parameter design variables is determined to be x1∈[2.2,3], x2∈[1.8,2.4], x3∈[2.5,3], x4∈[1,2], x5∈[2.4,3.2], x6∈[2,3], x7∈[1.5,3], x8∈[3,4].

[0055] The maximum stress does not exceed 475MPa and the maximum deformation does not exceed 10mm. The constraint conditions are described as stσ≤475MPa; D≤10mm; f t ≥9.5Hz; with the minimum mass M, the maximum fatigue life F and the first-order bending mode frequency f b The maximum is the optimization goal, and the optimization objective function is described as: minM(x1,x2,...,x8); maxF(x1,x2,...,x8); maxf b (x1,x2,...,x8); where M is the mass of the frame, F is the fatigue life of the frame, and f b is the first-order bending modal frequency, σ is the yield strength of the frame structure, D is the maximum allowable deformation of the frame, f t is the first-order torsional mode frequency.

[0056] The Hammersley sampling method was used to design the experiment within the design space, resulting in an experimental design matrix with eight design variables and a total of 40 sample points. The experimental design matrix is ​​shown in Table 2.

[0057] Table 2 Experimental design

[0058]

[0059]

[0060]

[0061]

[0062] According to the sampling data shown in Table 2, the geometric dimensions of the frame structure were redesigned one by one to construct the model calculation: the dimensional parameters of the variables with dimensional changes were directly redefined, and then each set of sample points was subjected to finite element calculation and analysis according to the modified dimensions to obtain simulation data. Based on these 40 sets of result data, a prediction model for performance responses such as construction quality, fatigue life, maximum deformation, maximum stress, and first-order modal frequency was constructed based on the improved radial basis function neural network method.

[0063] In the design space, 15 additional sample points are selected to test the accuracy of the prediction model. Common error analysis evaluation indicators include the coefficient of determination (R 2 ), root mean square error (RMSE), R 2 The closer it is to 1, the closer the RMSE value is to 0, indicating that the accuracy of the prediction model is higher. The accuracy test of the prediction model of each response is as follows: Figure 8-12 The approximate error analysis results are shown in Table 3.

[0064] Table 3 Approximate error analysis results

[0065]

[0066] After the accuracy of the prediction model is tested, if the accuracy meets the requirements of simulation analysis, the frame structure can be lightweight optimized according to the prediction model. Otherwise, it is necessary to resample and construct the prediction model using the Hammersley method in the design space.

[0067] Step 4: If the accuracy of the prediction model is verified to meet the requirements, then the minimum frame structure mass, the maximum fatigue life and the maximum first-order bending modal frequency of the car body are used as optimization objective functions, the maximum stress, the maximum deformation and the first-order torsional modal frequency of the frame structure under full-load torsional working conditions are used as constraints, and the eight variables are used as optimization parameter design variables. A multi-island genetic algorithm is used to optimize and solve the prediction model that meets the requirements, and a frame structure lightweight optimization design analysis model is established. After optimization iterative calculation of the frame structure lightweight optimization design analysis model, a Pareto frontier solution set of the frame lightweight optimization design is obtained, and the optimal solution in the Pareto frontier solution set is selected as the optimization design result;

[0068] In step 4, the minimum frame structure mass, maximum fatigue life and maximum first-order bending modal frequency of the car body are used as optimization objective functions, the maximum stress, maximum deformation and first-order torsional modal frequency of the frame structure under full-load torsional working conditions are used as constraints, and the structural dimensions of the main components of the frame are used as optimization design variables. The multi-island genetic optimization algorithm is used to optimize and solve the prediction model that meets the requirements, and a frame structure lightweight optimization design analysis model based on the prediction model established by the improved radial basis neural network method is established. In the optimization component, the subgroup size is set to 20, the evolutionary generation is set to 60, the crossover probability is set to 0.9, and the mutation probability is set to 0.09; after optimization iterative calculation, the Pareto frontier solution set of the frame lightweight optimization design is obtained, as shown in FIG. Figure 14As shown in the figure; since this is a lightweight design study for the frame, while ensuring that the first-order bending modal frequency of the frame is not reduced and the fatigue life is improved, the mass of the frame is minimized as much as possible, and the optimal solution in the Pareto frontier is selected as the final result of the optimized design.

[0069] The optimization results of the design variables are rounded according to the actual requirements of the project, and the specific values ​​of the optimized design variables are shown in Table 4.

[0070] Table 4 Design variables after lightweight optimization

[0071]

[0072] Step 5: Perform engineering rounding on the lightweight optimization data in the optimization design result to obtain a rounded value, and assign the rounded value to the frame performance after the frame finite element model analysis and calculation optimization.

[0073] In step five, by modifying the dimensional parameters in Hypermesh, and then based on the full-load torsion condition, first-order mode, and fatigue life settings, the output file was submitted to Nastran for calculation of the maximum deformation, maximum stress, and first-order modal frequency under the full-load torsion condition, and fatigue life calculation using the Hypermesh fatigue solver. The optimized frame performance was analyzed and calculated, and the simulation results verified the feasibility of the optimized design. The optimized design results are shown in Table 5. The results show that the lightweight optimization design of the frame structure meets the design requirements. If it does not meet the requirements, the value range of the finite element model parameter size needs to be redefined and revalued until the optimized dimensions meet the requirements. The calculated weight reduction before and after lightweighting reaches 84 kg, a weight reduction ratio of 10.2% of the frame mass, fully improving material utilization.

[0074] Table 5 Optimization design simulation results

[0075]

[0076] Step 6. Based on the simulation analysis, a lightweight design scheme that meets the performance requirements can be obtained. The lightweight and optimized frame structure is trial-produced according to the parameter combination. The lightweight frame structure is tested and verified according to the requirements of full-load torsional conditions, first-order modal and fatigue life analysis. The verification results show that the lightweight design scheme is effective.

[0077] In this embodiment, three-dimensional modeling software Pro / E, simulation software HyperMesh14.0, Patran2010 & Nastran2010, optimization platform Hyperstudy, and operating system Windows are used.

[0078] Although the embodiments of the present disclosure have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and alterations may be made to the embodiments without departing from the principles and spirit of the present disclosure, and the scope of the present disclosure is defined by the appended claims and their equivalents.

Claims

1. A lightweight design method for a variable thickness frame structure of a heavy vehicle, characterized in that: The steps include: Step 1: Obtain a frame structure with known performance that meets the requirements for finite element modeling to obtain a frame finite element model. Based on the load borne by the frame structure under the full-load torsion working condition of the fire truck, calculate the first-order modal frequency, fatigue life, mass, maximum stress and maximum deformation of the frame structure in the free state, and compare and verify with the actual working condition performance test data of the known frame structure; Step 2: Define the following eight variables in the dimensional parameters of the frame finite element model as optimization parameter design variables: the thickness of the upper wing panel, the middle panel, and the lower wing panel of the front cross member; the thickness of the upper wing panel, the thickness of the lower wing panel, and the middle panel of the side longitudinal member; and the thickness of the upper end panel and the lower end panel of the middle cross member; thus obtaining the frame finite element model of the variable thickness components; Step 3: Performing experimental design sampling on the finite element model of the vehicle frame with variable thickness components, and after obtaining a number of simulation data, establishing a prediction model for the lightweight optimization design of the vehicle frame based on an improved radial basis function neural network method according to the simulation data; the improved radial basis function neural network method comprises: obtaining an RBF neural network code; After the particle swarm parameters are initialized, the fitness value of the particles is calculated; the historical optimal values ​​of individuals and groups are found, and the particle speed and position are updated according to the historical optimal values; it is determined whether the parameters are optimal. If so, the optimal prediction model is obtained; if not, the fitness value of the particles is continued to be calculated to find the next historical optimal value; Step 4: If the accuracy of the prediction model is verified to meet the requirements, then the minimum frame structure mass, the maximum fatigue life and the maximum first-order bending modal frequency of the car body are used as optimization objective functions, the maximum stress, the maximum deformation and the first-order torsional modal frequency of the frame structure under full-load torsional working conditions are used as constraints, and the eight variables are used as optimization parameter design variables. A multi-island genetic algorithm is used to optimize and solve the prediction model that meets the requirements, and a frame structure lightweight optimization design analysis model is established. After optimization iterative calculation of the frame structure lightweight optimization design analysis model, a Pareto frontier solution set of the frame lightweight optimization design is obtained, and the optimal solution in the Pareto frontier solution set is selected as the optimization design result; Step 5: Perform engineering rounding on the lightweight optimization data in the optimization design result to obtain a rounded value, and assign the rounded value to the frame performance after the frame finite element model analysis and calculation optimization.

2. The lightweight design method for a variable thickness frame structure of a heavy vehicle according to claim 1, characterized in that: The experimental design sampling of the finite element model of the vehicle frame with variable thickness components is carried out. The steps of obtaining a number of sample points specifically include the following steps: determining the range of variation of the optimization parameter design variables, setting the constraints of the maximum stress and maximum deformation, and taking the minimum mass M, the maximum fatigue life F and the first-order bending modal frequency f as the optimal parameter design variables, setting the maximum stress and maximum deformation constraints, and taking the maximum stress and maximum deformation constraints as the optimal parameter design variables, setting the maximum stress and maximum deformation constraints, and setting ..., setting the maximum stress and maximum deformation constraints, and setting the maximum stress and maximum deformation constraints, setting the b The maximum is the optimization goal, and the Hammersley sampling method is used to carry out experimental design in the design space to obtain the experimental design matrix of 8 design variables and several sample points. Finite element calculation and analysis are performed on each group of sample points according to the modified size to obtain simulation data.

3. The lightweight design method for a variable thickness frame structure of a heavy vehicle according to claim 2, characterized in that: After the rounded value is reassigned to the frame performance after the frame finite element model analysis and calculation is optimized, the following step is also included: judging whether the frame performance meets the performance requirements; if it does not meet the performance requirements, re-determining the range of change of the optimization parameter design variable and then executing step three.

4. The lightweight design method for a variable thickness frame structure of a heavy vehicle according to claim 3, characterized in that: If the frame performance meets the performance requirements, a lightweight design scheme is determined, and a lightweight frame structure after lightweight optimization is trial-produced according to the parameter combination. The lightweight frame structure is then tested and verified in accordance with the requirements of full-load torsional working conditions, first-order modal and fatigue life analysis.

5. The lightweight design method for a variable thickness frame structure of a heavy vehicle according to claim 2, characterized in that: The optimization objective function of the optimization objective is expressed as follows: min M(x1,x2,…,x8);max F(x1,x2,…,x8);max f b (x1,x2,…,x8)? Where M is the frame mass, F is the frame fatigue life, and f b is the first-order bending mode frequency.

6. The lightweight design method for a variable thickness frame structure of a heavy vehicle according to claim 2, characterized in that: The specific constraints for setting the maximum stress and maximum deformation are: The maximum stress does not exceed 475MPa and the maximum deformation does not exceed 10mm. The constraint conditions are described as stσ≤475MPa; D≤10mm; f t ≥9.5Hz, where σ is the yield strength of the frame structure, D is the maximum allowable deformation of the frame, and f t is the first-order torsional mode frequency.

7. The lightweight design method for a variable thickness frame structure of a heavy vehicle according to claim 2, characterized in that: The variation range of the optimization parameter design variables is determined as follows: for the design variables "front crossbeam upper wing panel thickness x1, middle panel thickness x2, lower wing panel thickness x3, side longitudinal beam upper wing panel thickness x4, lower wing panel thickness x5 and middle panel thickness x6, middle crossbeam upper end panel thickness x7 and lower end panel thickness x8", they are specifically: x1∈[2.2,3], x2∈[1.8,2.4], x3∈[2.5,3], x4∈[1,2], x5∈[2.4,3.2], x6∈[2,3], x7∈[1.5,3], x8∈[3,4].

8. The lightweight design method for a variable thickness frame structure of a heavy vehicle according to claim 1, characterized in that: The multi-island genetic algorithm is used to optimize the prediction model that meets the requirements. Specifically, the above prediction model is imported after establishment, and then the subgroup size is determined to be 10 and the total group size is 200; the number of subgroups is 20; the total number of generations of evolution is 20; the crossover probability is 0.9; the mutation probability is 0.009; the inter-island migration rate is 0.4; the migration interval generation number is 5; then after setting the above parameters, iterative optimization calculation is performed.

9. The lightweight design method for a variable thickness frame structure of a heavy vehicle according to claim 1, characterized in that: The steps for obtaining the Pareto frontier solution set are specifically as follows: in the steps of the multi-island genetic algorithm, the subgroup size is set to 20, the evolutionary generations are set to 60, the crossover probability is set to 0.9, and the mutation probability is set to 0.09; after optimized iterative calculation, the Pareto frontier solution set of the lightweight optimization design of the frame is obtained.

10. The lightweight design method for a variable thickness frame structure of a heavy vehicle according to claim 2, characterized in that: The number of sample points is 40 sample points.

Images