Establishment method of body-in-white rigidity target management system based on machine learning

Abstract

The invention relates to the technical field of vehicle engineering, in particular to a method for establishing a body-in-white rigidity target management system based on machine learning, which comprises the following steps of: S1, analyzing energy distribution of 36 types of body-in-white torsional rigidity, and identifying a key design structure area; s2, establishing a stiffness target model based on the vehicle body stiffness database; s3, vehicle body rigidity target decomposition based on clustering analysis; s4, performing benchmarking vehicle screening based on KNN analysis; and S5, optimizing the body-in-white structure: by comparing the structural strain energy difference between the benchmarking vehicle and the development vehicle, converting the comparison between the development vehicle type and the benchmarking vehicle from the macroscopic rigidity comparison into the strain energy comparison, directly obtaining the position generated by the performance difference between the development vehicle type and the benchmarking vehicle, and determining the optimization direction. The method has the advantages that the body-in-white torsional rigidity target management system is constructed based on machine learning, target decomposition is achieved through energy, and the body structure design efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle engineering, and in particular to a method for establishing a white body stiffness target management system based on machine learning. Background Art

[0002] In the concept design stage, the white body stiffness is an important indicator for evaluating the design reliability and overall safety performance of a vehicle, and has very important practical significance for improving the relevant body structure, improving the strength and stiffness of the vehicle, and enhancing the safety and reliability of the vehicle. Moreover, the body torsional stiffness is a key parameter in the evaluation index of the body lightweight coefficient, and its magnitude directly relates to the lightweight level of the body. Therefore, the research on the decomposition and formulation of the body stiffness target, the optimization of the body performance, and the benchmarking has important guiding value.

[0003] Currently, there is relatively little research on the decomposition of the body stiffness target at home and abroad. Machine learning has been widely applied in prediction. In the current development of the white body, the stiffness design ability has been relatively perfect. In the research, the stiffness decomposition cannot solve the practical engineering problems, and the key problems still exist:

[0004] 1. When developing a vehicle model, currently only the overall stiffness of the body can be formulated, and the stiffness cannot be decomposed downward. Although there are rich means for optimizing the body structure (topology optimization / sensitivity analysis / size optimization / multi-objective optimization, etc.), if the optimization positions cannot be accurately identified, it is still possible to fall into a misunderstanding, resulting in a waste of time and resources.

[0005] 2. When benchmarking with a benchmark vehicle model, only the overall stiffness can be compared, and the specific difference points between the developed vehicle model and the benchmark vehicle model cannot be quantified.

[0006] 3. For the same vehicle model, the performance differences that may be caused after the structure is updated require spending a lot of time to verify each updated structure one by one, and a quantitative explanation cannot be quickly given.

[0007] Based on the energy method, this article constructs a body stiffness target management system to guide the body stiffness design and solve the above three key problems. First, for... Summary of the Invention

[0008] To solve the above problems, the present invention provides a method for establishing a white body stiffness target management system based on machine learning. A white body torsional stiffness target management system is constructed based on machine learning, and target decomposition is realized through energy to improve the efficiency of vehicle body structure design. First, the energy distribution of the torsional stiffness of 36 white bodies is analyzed to identify key structural areas; secondly, based on data analysis, the rules are found, and a general standard for white body stiffness distribution is formulated; then, using the calculated data, a vehicle body stiffness target management system is constructed, and the application method of the system is initially defined. Finally, taking a certain vehicle model under development as an example, aiming at the problem of insufficient stiffness, through precise comparison with the benchmark vehicle, the structural positions to be optimized are located, and the stiffness target is achieved.

[0009] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0010] A method for establishing a white body stiffness target management system based on machine learning, comprising the following steps:

[0011] S1. Analyze the energy distribution of the torsional stiffness of 36 white bodies and identify key design structural areas;

[0012] S2. Establish a stiffness target model based on the vehicle body stiffness database: Analyze the relationship between torsional stiffness and the mass and lightweight coefficient of the white body, and establish linear regression models respectively to obtain its torsional stiffness target K,

[0013] K = -2230.6L + 27551 (4)

[0014] In the formula, K is the static torsional stiffness of the white body, N·m / °; L is the lightweight coefficient of the white body

[0015] S3. Body stiffness target decomposition based on cluster analysis: Determine similar vehicle models according to the characteristics of the vehicle model, and then determine the proportion of strain energy of the white body stiffness to achieve stiffness decomposition;

[0016] S4. Selection of benchmark vehicles based on KNN analysis: Through classification analysis of the vehicle body stiffness library, select benchmark vehicles and output benchmark vehicle data for benchmark analysis with the development vehicle;

[0017] S5. Optimization of the white body structure: By comparing the differences in structural strain energy between the benchmark vehicle and the development vehicle, the comparison between the development vehicle and the benchmark vehicle changes from a macroscopic stiffness comparison to a comparison of strain energy, directly obtaining the positions where the performance differences between the two occur and determining the optimization direction.

[0018] Further explanation, in step S1, among the 36 white vehicles, there are 9 sedans (CAR), 11 MPVs, and 18 SUVs, namely Fengxing T5, Avita Alita, Dongfeng Peugeot 3008, Baojun 310, Jiangling Yushang, Volkswagen Golf, Toyota Highlander, Jianghuai Refine, Jianghuai Siegel A5, BYD Song MAX, Baojun 360, Baojun 510, Honda Jade, New Baojun RS5, Fengxing Yacht, Fengxing Yacht (non - sunroof version), Tesla Model 3, XPeng P7, Haval Euler, Peugeot 308, Jaguar - Land Rover, Ford Edge, GAC Trumpchi GA8, Jianghuai Siegel X8, Fengxing T5 (non - sunroof version), Volkswagen Tiguan L, Volkswagen Touran L, Volkswagen Tiguan L (non - sunroof version), Fengxing S50ev, New Baojun RS7, Baojun 730, Volkswagen Touran, Volkswagen ID4, Volkswagen ID6, Volvo XC60.

[0019] Further explanation, in step S1, the energy of the body - in - white torsional stiffness refers to the total strain energy W.

[0020] The total strain energy W is equal to the work done by the body when it generates a torsional angle θ under the action of a torque T, as shown in Equation (1). Furthermore, the relationship between the torsional stiffness K and the strain energy W can be converted, as shown in Equation (2).

[0021]

[0022]

[0023] The definition of the body - in - white lightweight coefficient L is as follows:

[0024]

[0025] In the formula, m is the mass of the body - in - white structure, Kg; K is the static torsional stiffness of the body - in - white, N·m / °; A is the footprint area of the body - in - white, m2, and A = wheelbase × (front track + rear track) / 2.

[0026] Further explanation, in step S3, the characteristics of the vehicle models include sunroof type, lightweight coefficient, torsional stiffness, glass, roof panel, front compartment, side panel, front floor, and rear floor.

[0027] Further explanation, in step S3, the stiffness decomposition includes the rear floor and side panel of the body - in - white; the strain energy of the rear floor is decomposed downward into joint strain energy and beam structure strain energy; the strain energy of the side panel is decomposed downward into joint strain energy and beam structure strain energy.

[0028] The beneficial effects of the present invention are:

[0029] Construct a body stiffness target management system based on the energy method to guide the body stiffness design and solve the three key problems pointed out in the background technology. First, conduct a torsional stiffness analysis on 36 various types of vehicles to determine the strain energy distribution of each vehicle, and perform data analysis to find the patterns among them, and formulate a general standard for the body-in-white stiffness distribution. Then, use the calculated data to construct a body stiffness target management system, and initially define the application method of the system: First, determine the overall stiffness target according to the vehicle model positioning, obtain the corresponding strain energy, and distribute it to each assembly according to the energy distribution ratio; then conduct an energy comparison with the benchmark vehicle to determine the structure update ratio; finally, take a vehicle under development as an example, and for the problem of insufficient stiffness, through an accurate comparison with the benchmark vehicle, locate the structural positions to be optimized, and achieve the compliance of the stiffness target. Brief Description of the Drawings

[0030] Figure 1 It is a schematic diagram of the distribution of the body modules of the invention.

[0031] Figure 2 It is a flowchart of the establishment and application of the body-in-white stiffness target management system of the invention.

[0032] Figure 3 It is a diagram of the downward decomposition of the body-in-white strain energy of the invention.

[0033] Figure 4 It is a diagram of the comparative analysis of the body-in-white strain energy of the invention.

[0034] Figure 5 It is a model diagram of the body-in-white stiffness target definition of the invention.

[0035] Figure 6 It is a diagram of the analysis of the change trend of the proportion of the strain energy in the body of the invention.

[0036] Figure 7 It is a diagram of the distribution of the proportion and quantity of the strain energy of each assembly of the body-in-white of the invention.

[0037] Figure 8 It is a diagram of the comparison of the torsional stiffness between the benchmark vehicle and the vehicle under development in an embodiment of the invention.

[0038] Figure 9 It is a diagram of the comparison of the strain energy of the assemblies between the benchmark vehicle and the vehicle under development in an embodiment of the invention.

[0039] Figure 10 It is a structural design diagram of the roof panel strain energy in an embodiment of the invention.

[0040] Figure 11 It is a diagram of the distribution of the body modules in an embodiment of the invention.

[0041] Figure 12 It is a structural design diagram of the J1 joint benchmark in an embodiment of the invention.

[0042] Figure 13 It is the design drawing of the J8 joint alignment structure in an embodiment of the invention.

[0043] Figure 14 It is the design drawing of the B4 sill topology structure in an embodiment of the invention. Specific Embodiments

[0044] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0045] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this invention belongs. The terms used in the specification of this invention herein are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0046] Please also refer to Figure 1 and Figure 7 , a method for establishing a white body stiffness target management system based on machine learning in a preferred embodiment of the present invention, includes the following steps:

[0047] S1. Strain energy analysis of the white body torsion stiffness condition.

[0048] The specific conditions in the body torsion stiffness simulation analysis are as follows: Constraining the degrees of freedom 123 of the center point of the installation hole of the left rear shock absorber mounting seat, the degrees of freedom 13 of the center point of the right rear shock absorber mounting hole, and the degree of freedom 3 of the midpoint at the bottom of the front bumper beam. Apply a pair of forces F in the opposite Z directions at the center points of the installation holes of the left and right front shock absorber mounting seats of the white body.

[0049] According to the law of conservation of energy, the total strain energy W is equal to the work done by the body when generating a torsional angle θ under the action of torque T, as shown in Equation (1). Furthermore, the relationship between the torsional stiffness K and the strain energy W can be converted, as shown in Equation 2.

[0050]

[0051]

[0052] At the same time, the Light Weight Index of BIW (L) proposed by BMW is an index that is currently widely accepted in the automotive industry to evaluate the lightweight level of the white body. The definition of the white body lightweight coefficient L is:

[0053]

[0054] In the formula, m is the body-in-white structure mass, in Kg; K is the static torsional stiffness of the body-in-white, in N·m / °; A is the footprint area of the body-in-white, in m2, and A = wheelbase × (front track + rear track) / 2.

[0055] In view of the cumulative characteristic of strain energy, the vehicle body is divided into five modules: roof panel, front compartment, side panel, front floor, and rear floor. Analyze the strain energy distribution of each module under the torsional condition, as Figure 1 shown. A total of 36 vehicle models were analyzed, including 9 sedans (CAR), 11 MPVs, and 18 SUVs. The specific data are shown in Table 1.

[0056] It can be seen from Equation (2) and Table 1 that:

[0057] 1. The magnitude of strain energy can directly reflect the level of torsional stiffness of the body-in-white. In terms of stiffness, the overall stiffness is difficult to decompose into each module, while strain energy can be freely added.

[0058] 2. Strain energy is mainly concentrated in the side panel, rear floor, and front compartment areas, which are the key design areas.

[0059] Table 1 Analysis Results of Strain Energy of 36 Vehicle Models (Supplementary)

[0060]

[0061]

[0062]

[0063]

[0064] S2 Establish a stiffness target model based on the vehicle body stiffness database: On the basis of existing data, analyze the data law, establish a simple and practical regression model, and reasonably define parameters such as vehicle body weight and torsional stiffness according to the current vehicle model positioning. The vehicle body stiffness database will also be continuously improved during the vehicle model development and benchmarking process of the vehicle manufacturer to meet the increasing demand for lightweight in the current automotive industry.

[0065] S3. Decompose the vehicle body stiffness target based on cluster analysis. Since the vehicle body stiffness performance is directly related to the magnitude of strain energy, there is a large error through strain energy decomposition, while the strain energy ratio of each vehicle model under torsional stiffness has a large concentration characteristic. Therefore, according to the characteristics of the vehicle model, such as sunroof structure, floor structure, vehicle type (SUV, MPV, SEDAN), and vehicle body torsional stiffness, determine similar vehicle models, and then determine the strain energy ratio of the body-in-white stiffness to achieve stiffness decomposition, as Figure 3 shown.

[0066] S4. Benchmark vehicle screening based on KNN analysis. By classifying and analyzing the body stiffness library, scientifically select the benchmark vehicle and output the benchmark vehicle data for benchmark analysis with the development vehicle.

[0067] S5. White body structure optimization. By comparing the differences in structural strain energy between the benchmark vehicle and the development vehicle, the comparison between the development model and the benchmark vehicle changes from a macroscopic stiffness comparison to a comparison of strain energy, directly obtaining the location where the performance differences between the two occur and determining the optimization direction. As Figure 4 shown, W* represents the strain energy of a certain structure of the body. Subsequently, various means such as strain energy distribution analysis, structural benchmarking, and topology optimization are used to optimize the design of key positions until the requirements are met.

[0068] Among them, the definition of the initial body stiffness target: According to the data in Table 1, analyze the relationship between the torsional stiffness and the body-in-white mass and lightweight coefficient, and establish a linear regression model respectively, as Figure 5 shown. Compared with the body mass, the lightweight coefficient has a higher correlation with the torsional stiffness, and the lightweight coefficient includes the basic body size information. Therefore, Equation (4) is selected as the basis for the preliminary definition of the new vehicle development target. In the development of a new vehicle, after determining the dimension A, weight target, and lightweight level, the torsional stiffness target can be obtained.

[0069] K = -2230.6L + 27551 (4)

[0070] Analysis of the correlation between strain energy and the stiffness of the body assembly: It is proposed that under a given load condition, after the body structure deforms, the total strain energy is W, and the proportion η i of the strain energy stored in module W i in the total strain energy of the body is called the stiffness contribution of this module. This inference has obvious limitations. For example, when a certain module is rigid, the stiffness is infinite, and at this time, the module strain energy W i and the proportion η i are both 0; when the stiffness of a certain module is very small, at this time, the module strain energy W i and the proportion η i are both relatively high, which is significantly inconsistent with the actual situation.

[0071] Use the elastic modulus of each module's material to simulate the equivalent stiffness of the module. The elastic modulus of the actual material is 2.1E5 MPa for steel. Analyze the strain energy distribution of each module as the equivalent stiffness of the module gradually increases. The analysis results are as Figure 6 shown.

[0072] 1. From Figure 6Analysis in (a) shows that when the equivalent stiffness of each module changes simultaneously, the proportion of strain energy of each module remains basically unchanged, especially for the side panel, rear floor, and front compartment modules. Therefore, the stiffness contribution η i The establishment of the condition requires that there are no significant stiffness changes in each module.

[0073] 2. From Figure 6 Analysis in (b) and (c) shows that when changing the equivalent stiffness of the roof or front floor module, it only affects other modules when the equivalent stiffness is relatively low. After the equivalent stiffness increases, the influence on other modules can be ignored. In vehicle body design, the situation of extremely low stiffness basically does not occur, so the coupling effect between the roof and front floor and the other three modules can be ignored.

[0074] 3. From Figure 6 Analysis in (d), (e), and (f) shows that there is obvious energy exchange between the side panel, rear floor, and front compartment modules, and there is an obvious coupling relationship among the three.

[0075] Analysis of the strain energy distribution law of the body-in-white assembly:

[0076] According to the data in Table 1, analyze the proportion distribution of the strain energy of each assembly of the body-in-white to evaluate the distribution range of the strain energy proportion, as Figure 7 shown. The proportion of the strain energy of the roof mainly concentrates between [1.9%, 4.1%]. There are 9 vehicle models with a relatively high proportion of the strain energy of the roof. After analysis, it is a structure with a skylight as Figure 7 (a) shown; the proportion of the strain energy of the front floor mainly concentrates between [1.9%, 4.0%]. There are 11 vehicle models with a relatively high proportion of the strain energy of the front floor in the front compartment. After analysis, the connection method between the front and rear floors is a through type, as Figure 7 (b) and Figure 8 shown. The proportion of the strain energy of the front compartment mainly concentrates between [10.7%, 17.4%]. The proportion of the side panel and rear floor basically follows a normal distribution, and the proportion of the strain energy is the highest.

[0077] Taking a certain vehicle model development project as an example, introduce in detail the application of the body-in-white stiffness target management system

[0078] Step 1: Determine the structure optimization location

[0079] For a certain vehicle model development project, the main parameters of the vehicle model are. Use the KNN algorithm to locate the benchmark vehicle of a certain vehicle model. First, compare the torsional stiffness difference between the two vehicle models, which is 3660 N·m / °.

[0080] The quantitative comparison of the strain energy of each assembly of the benchmark vehicle and the development vehicle is as Figure 8-9 shown. It can be seen that the weak positions are concentrated in the roof and side panel areas, and these two areas need to be optimized to achieve the stiffness target.

[0081] Step 2 Optimization of the roof cover area structure

[0082] Since the structure of the roof cover area is relatively simple, the weak positions are confirmed through the analysis of the strain energy nephogram ( Figure 9 ). It can be seen from the nephogram that the weak positions are concentrated in the joint area of the roof cover frame. By adding joint reinforcement plates, the stiffness is increased by 1228 N·m / °.

[0083] Step 3 Optimization of the side wall area structure

[0084] Judging from Figure 10 , the structure of the side wall area is complex. By refining the side wall into joint and beam structures, the strain energy differences between the benchmark vehicle and the development vehicle are quantitatively compared. The side wall is divided into 8 joints and 6 beam structures, namely J1 (lower joint of A-pillar), J2 (upper joint of A-pillar), J3 (lower joint of B-pillar), J4 (upper joint of B-pillar), J5 (upper joint of C-pillar), J6 (lower joint of C-pillar), J7 (lower joint of D-pillar), J8 (upper joint of D-pillar), B1 (lower part of A-pillar), B2 (A-pillar), B3 (B-pillar), B4 (door sill), B5 (upper side beam), B6 (rear side wall). Compared with the benchmark vehicle, the strain energy of 4 joints, namely J1, J5, J6 and J8, and 3 beams, namely B2, B4 and B6, of the development vehicle is relatively high. Here, only the structures of J1, J8 and B4 are selected for structural optimization.

[0085] (1.) By benchmarking the structure of the lower joint of the A-pillar, it is found that the thicknesses of the two vehicle models are basically the same, but the structures are quite different. The reinforcement plate of the benchmark vehicle is of the upper and lower through type, as shown by the purple part in Figure 11 , while the concept data of the development vehicle connects the reinforcement plate to the welding edge of the inner panel of the door sill. By referring to the structure of the benchmark vehicle and changing the structure and welding process of the reinforcement plate, the torsional stiffness is increased by 367 N·m / °.

[0086] (2.) By benchmarking the structure of the upper joint of the D-pillar, it is found that the thicknesses of the two vehicle models are also basically the same, but the structures are quite different. As shown in Figure 12 , a relatively large-sized reinforcement plate (pink part) is arranged inside the joint of the benchmark vehicle, which is respectively connected to the inside of the D-pillar joint and the outer panel reinforcement plate. At the same time, the fillet radius of the upper joint of the D-pillar is smaller than that of the benchmark vehicle. To simplify the process and reduce the weight, the structure of the benchmark vehicle is approximated by using two small reinforcement plates. At the same time, the fillet of the upper joint of the D-pillar is increased by 10 mm, and the torsional stiffness is increased by 1323 N·m / °.

[0087] (3.) The space of the door sill part is filled with solid meshes as the optimization space, and a topology optimization model is established. As shown in Figure 13 .

[0088] Objective: Maximize the torsional stiffness.

[0089] Constraint: The volume fraction is not greater than 20%.

[0090] Process constraint: Apply a draft constraint in the Y direction.

[0091] According to the optimization results, material accumulation appears at the rear end of the sill beam, which is interpreted as adding a reinforcement plate inside the sill, achieving a 919 N·m / ° increase in torsional stiffness.

[0092] In summary, in this embodiment, the torsional stiffness, lightweight coefficient, and strain energy decomposition values of different assemblies of 36 different vehicle models are obtained. And based on the above data, a process for establishing and applying a white body stiffness target management system is established. It is possible to determine similar vehicle models according to the characteristics of the vehicle model, such as the sunroof structure, floor structure, vehicle type (SUV, MPV, SEDAN), and body torsional stiffness, and then determine the proportion of strain energy of the white body stiffness to achieve stiffness decomposition. Moreover, it is possible to scientifically select the benchmark vehicle, output the benchmark vehicle data for benchmarking analysis with the development vehicle. And the application results show that the application of the target management system can quickly locate the structural optimization position and effectively combine means such as topology optimization and benchmarking to achieve rapid performance improvement of the developed vehicle model.

Claims

1. A method for establishing a body-in-white stiffness target management system based on machine learning, characterized in that It includes the following steps: S1. Analyze the energy distribution of the torsional stiffness of 36 body-in-whites, and identify key design structure areas; S2. Establish a stiffness target model based on the body stiffness database: Analyze the relationship between torsional stiffness, body-in-white mass, and lightweight coefficient, and establish linear regression models respectively to obtain the torsional stiffness target K, K = -2230.6L + 27551 (4) In the formula, K is the static torsional stiffness of the body-in-white, N·m / °; L is the lightweight coefficient of the body-in-white S3. Decompose the body stiffness target based on cluster analysis: Determine similar vehicle models according to the characteristics of the vehicle model, and then determine the proportion of strain energy of the body-in-white stiffness to achieve stiffness decomposition; S4. Screen benchmark vehicles based on KNN analysis: Through classification analysis of the body stiffness database, select benchmark vehicles and output benchmark vehicle data for benchmark analysis with the development vehicle; S5. Optimize the body-in-white structure: By comparing the differences in structural strain energy between the benchmark vehicle and the development vehicle, the comparison between the development vehicle and the benchmark vehicle changes from a macroscopic stiffness comparison to a comparison of strain energy, directly obtaining the location where the performance differences between the two occur and determining the optimization direction.

2. The method for establishing a white body stiffness target management system based on machine learning according to claim 1, characterized in that: In step S1, the 36 body-in-whites include 5 sedans, 13 MPVs, and 18 SUVs; they are Fengxing T5, Alita, Dongfeng Peugeot 3008, Baojun 310, Jiangling Yushang, Volkswagen Golf, Toyota Highlander, Jianghuai Refine, Jianghuai Sihao A5, BYD Song MAX, Baojun 360, Baojun 510, Honda Jade, New Baojun RS5, Fengxing Yacht, Non-Sunroof Fengxing Yacht, Tesla Model 3, XPeng P7, Haval Euler, Peugeot 308, Jaguar-Land Rover, Ford Edge, Guangzhou Automobile Trumpchi GA8, Jianghuai Sihao X8, Non-Sunroof Fengxing T5, Volkswagen Tiguan L, Volkswagen Touran L, Non-Sunroof Volkswagen Tiguan L, Fengxing S50ev, New Baojun RS7, Baojun 730, Volkswagen Touran, Volkswagen ID4, Volkswagen ID6, Volvo XC60.

3. The method for establishing a white body stiffness target management system based on machine learning according to claim 1, wherein: In step S1, the energy of the torsional stiffness of the body-in-white refers to the total strain energy W, The total strain energy W is equal to the work done by the body when generating a torsional angle θ under the action of torque T, as shown in formula (1). Furthermore, the relationship between the torsional stiffness K and the strain energy W can be converted, as shown in formula (2). The definition of the lightweight coefficient L of the body-in-white is: In the formula, m is the body-in-white structure mass, Kg; K is the static torsional stiffness of the body-in-white, N·m / °; A is the body-in-white footprint area, m2, A = wheelbase × (front track + rear track) / 2..

4. The method for establishing a white body stiffness target management system based on machine learning according to claim 1, characterized in that: In step S3, the characteristics of the vehicle model include sunroof type, lightweight coefficient, torsional stiffness, glass, roof, front cabin, side panel, front floor, and rear floor.

5. The method for establishing a white body stiffness target management system based on machine learning according to claim 1, characterized in that: In step S3, the stiffness decomposition includes the rear floor and side panel of the body-in-white; the strain energy of the rear floor is decomposed downward into joint strain energy and beam structure strain energy; the strain energy of the side panel is decomposed downward into joint strain energy and beam structure strain energy.

Images