New energy automobile body-in-white and battery pack structural performance collaborative optimization method and system

Abstract

The invention provides a new energy automobile body-in-white and battery pack structure performance collaborative optimization method and system, relates to the technical field of new energy automobile structure design and optimization, and aims to solve the problems of performance imbalance, low efficiency, high cost, difficulty in model prediction convergence and the like caused by separated optimization in the prior art. Comprising the steps that a body-in-white and battery pack integrated finite element model is constructed, and structural performance analysis is conducted on the body-in-white and a battery pack; the method comprises the following steps: constructing a particle swarm-neural network hybrid model, optimizing a BP neural network by using a particle swarm algorithm, generating a Pareto leading-edge solution set based on structural performance analysis by taking maximization of the bending rigidity of a body in white and the first-order modal frequency of a battery pack as an optimization target through an improved multi-target genetic algorithm, and adjusting the thicknesses of 12 key plates based on the Pareto leading-edge solution set. And determining an optimal plate thickness parameter combination. The problems existing in the prior art are solved, light weight is achieved, and meanwhile the structural performance of the body-in-white and the battery pack is balanced.

Description

Technical Field

[0001] This invention belongs to the field of new energy vehicle structural design and optimization technology, and particularly relates to a method and system for synergistic optimization of the structural performance of the body-in-white and battery pack of new energy vehicles. Background Technology

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] With the accelerated global energy structure transformation, the new energy vehicle industry has developed rapidly, and lightweight vehicle body design and battery pack structural reliability have become key factors in improving vehicle performance. Compared with traditional fuel vehicles, new energy vehicles, due to the addition of core components such as battery packs, need to withstand greater loads and more complex operating conditions, which places higher demands on body stiffness, modal characteristics, and lightweighting.

[0004] However, most technologies for optimizing vehicle body and battery pack structures employ separate optimization methods for the two, failing to fully consider the coupling effect between them. This results in limitations in analytical dimensions and performance imbalances between the vehicle body and battery pack. Furthermore, traditional structural optimization relies on experimental testing and trial-and-error methods, which are time-consuming, inefficient, and costly. Additionally, traditional BP neural networks suffer from convergence difficulties and local optima problems in multi-output prediction, making it difficult to find the global optimum. Moreover, lightweight design typically focuses on reducing material thickness or structural components, but weight reduction can lead to decreased stiffness or deteriorated vibration characteristics. Summary of the Invention

[0005] To overcome the shortcomings of the prior art, this invention provides a method and system for synergistic optimization of the structural performance of the body-in-white and battery pack of new energy vehicles. By constructing finite element models of the body-in-white and battery pack, accurate and effective analysis of the structural performance of the body-in-white and battery pack can be achieved. Furthermore, by integrating the PSO-BP neural network and the improved NSGA-II algorithm, the structural performance of the body-in-white and battery pack can be balanced while achieving lightweighting, thereby improving the overall safety and reliability of new energy vehicles.

[0006] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: The first aspect of this invention provides a method for synergistic optimization of the structural performance of the body-in-white and battery pack of a new energy vehicle, comprising: Construct an integrated finite element model of the body-in-white and battery pack; Based on the integrated finite element model of the body-in-white and battery pack, structural performance analysis is performed on the body-in-white and battery pack. Specifically: modal analysis is performed on the body-in-white to obtain modal data for solving the set conditions and determine the first torsional mode and the first bending frequency; stiffness analysis is performed on the body-in-white to calculate and obtain the bending stiffness and torsional stiffness as optimization constraints; plate-based sensitivity analysis is performed on the body-in-white to determine the optimization variables; constrained modal analysis is performed on the battery pack to determine the first modal frequency and shell stiffness; static analysis is performed on the battery pack to determine and optimize the structural strength and deformation characteristics of the battery pack under extreme conditions. A particle swarm optimization-neural network hybrid model was constructed, and the initial weights and thresholds of the BP neural network were optimized using the particle swarm optimization algorithm. Based on structural performance analysis, the optimization objectives were to maximize the bending stiffness of the body-in-white and the first-order modal frequency of the battery pack. A Pareto front solution set was generated using an improved multi-objective genetic algorithm. Based on the Pareto front solution set, the thicknesses of 12 key plates were adjusted to determine the optimal combination of plate thickness parameters.

[0007] As one implementation method, an integrated finite element model of the body-in-white and battery pack is constructed. The specific process is as follows: A finite element model of the body-in-white was created using the OptiStruct solver in HyperMesh software, employing a shell element model. The finite element model of the battery pack adopts the power battery model and is simulated using solid cells and BEAM cells.

[0008] As one implementation method, stiffness analysis is performed on the body-in-white to calculate and obtain the bending stiffness and torsional stiffness as optimization constraints. The specific process is as follows: Set up separate analysis conditions for bending stiffness and torsional stiffness; Based on the bending stiffness and torsional stiffness analysis conditions, stiffness parameters are calculated and obtained as optimization constraints.

[0009] As one implementation method, based on the bending stiffness and torsional stiffness analysis conditions, stiffness parameters are calculated and obtained as optimization constraints. The specific process is as follows: The Z-axis displacement cloud map is obtained by finite element solver, and the displacement values ​​of key measuring points are extracted. Calculate the bending stiffness and torsional stiffness based on the displacement values ​​of key measuring points; The calculated bending stiffness and torsional stiffness are compared with the design target values ​​as constraints in the optimization process.

[0010] A second aspect of the present invention provides a system for synergistic optimization of the structural performance of a new energy vehicle body-in-white and a battery pack, comprising: The finite element model building module is used to build an integrated finite element model of the body-in-white and the battery pack; The structural performance analysis module is used to perform structural performance analysis on the body-in-white and battery pack based on an integrated finite element model. Specifically, it performs modal analysis on the body-in-white to obtain modal data to solve set conditions and determine the first torsional mode and first bending frequency; it performs stiffness analysis on the body-in-white to calculate and obtain bending stiffness and torsional stiffness as optimization constraints; it performs plate-based sensitivity analysis on the body-in-white to determine optimization variables; it performs constrained modal analysis on the battery pack to determine the first modal frequency and shell stiffness; and it performs static analysis on the battery pack to determine and optimize the structural strength and deformation characteristics of the battery pack under extreme conditions. The multi-objective collaborative optimization module is used to construct a particle swarm optimization-neural network hybrid model. It uses the particle swarm optimization algorithm to optimize the initial weights and thresholds of the BP neural network. Based on structural performance analysis, it takes maximizing the bending stiffness of the body-in-white and the first-order modal frequency of the battery pack as optimization objectives. It generates a Pareto front solution set through an improved multi-objective genetic algorithm. Based on the Pareto front solution set, it adjusts the thickness of 12 preset key plates to determine the optimal combination of plate thickness parameters.

[0011] The above one or more technical solutions have the following beneficial effects: In this embodiment, by constructing an integrated finite element model of the body-in-white and the battery pack and combining it with a multi-objective genetic algorithm, the performance imbalance problem caused by traditional separate optimization is solved. This results in a 5.57% increase in the bending stiffness of the body-in-white (490.9 N / mm) and a 28.38% increase in the first-order constraint mode frequency of the battery pack (6.51 Hz), significantly enhancing the vehicle's resistance to deformation and vibration.

[0012] In this embodiment, an improved PSO-BP hybrid model and the NSGA-II algorithm are used for collaborative optimization. By constructing six coupling constraint indices, the problems of traditional BP neural networks being prone to getting trapped in local optima and the standard algorithm having insufficient convergence are overcome. At the same time, costs are reduced, efficiency is improved, and lightweighting is achieved while balancing the structural performance of the body-in-white and the battery pack.

[0013] Advantages of additional aspects of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. Attached Figure Description

[0014] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0015] Figure 1 This is a flowchart of the method for synergistic optimization of the structure and performance of the body-in-white and battery pack of a new energy vehicle according to Embodiment 1 of the present invention; Figure 2This is a first-order torsional mode shape diagram of Embodiment 1 of the present invention; Figure 3 This is a Z-direction displacement contour plot of the bending stiffness analysis in Embodiment 1 of the present invention; Figure 4 This is a Z-direction displacement contour plot of the torsional stiffness analysis in Embodiment 1 of the present invention; Figure 5 This is a simplified schematic diagram of the torsion angle according to Embodiment 1 of the present invention; Figure 6 The results of modal sensitivity analysis are from Embodiment 1 of the present invention; Figure 7 The results of stiffness sensitivity analysis are shown in Embodiment 1 of the present invention; Figure 8(a) is a displacement cloud diagram of the vertical bumpy working condition in Embodiment 1 of the present invention; Figure 8(b) is a displacement stress cloud diagram of the vertical bumpy working condition in Embodiment 1 of the present invention; Figure 9(a) is a displacement cloud diagram of the bumpy + emergency braking condition in Embodiment 1 of the present invention; Figure 9(b) is a stress cloud diagram of the bumpy + emergency braking condition in Embodiment 1 of the present invention; Figure 10(a) is a displacement cloud diagram of the bumpy + sharp left turn working condition in Embodiment 1 of the present invention; Figure 10(b) is a stress cloud diagram of the bumpy + sharp left turn condition in Embodiment 1 of the present invention; Figure 11(a) is a displacement cloud diagram of the bumpy + sharp right turn working condition in Embodiment 1 of the present invention; Figure 11(b) is a stress cloud diagram of the bumpy + sharp right turn working condition of Embodiment 1 of the present invention; Figure 12(a) is a displacement cloud diagram of the bumpy and rapid start-up acceleration condition in Embodiment 1 of the present invention; Figure 12(b) is a stress cloud diagram of the bumpy and rapid start-up acceleration condition in Embodiment 1 of the present invention; Figure 13 The variable distribution location is shown in Embodiment 1 of the present invention; Figure 14 This is a PSO-BP fitness curve diagram of Embodiment 1 of the present invention; Figure 15(a) is a scatter plot of the regression stiffness of the test set in Embodiment 1 of the present invention; Figure 15(b) is a scatter plot of the regression modes of the test set in Embodiment 1 of the present invention; Figure 16(a) shows a comparison of stiffness errors in Embodiment 1 of the present invention; Figure 16(b) shows a comparison of modal errors in Embodiment 1 of the present invention; Figure 17 This is the Pareto front solution set of Embodiment 1 of the present invention; Figure 18 This is an optimized bending stiffness displacement cloud diagram of Embodiment 1 of the present invention; Figure 19(a) shows the first mode shape and frequency value of the optimized battery pack according to Embodiment 1 of the present invention; Figure 19(b) shows the second mode shape and frequency value of the optimized battery pack according to Embodiment 1 of the present invention; Figure 19(c) shows the third mode shape and frequency value of the optimized battery pack according to Embodiment 1 of the present invention; Figure 19(d) shows the fourth mode shape and frequency value of the optimized battery pack according to Embodiment 1 of the present invention; Figure 19(e) shows the fifth mode shape and frequency value of the optimized battery pack according to Embodiment 1 of the present invention; Figure 19(f) shows the sixth mode shape and frequency value of the optimized battery pack in Embodiment 1 of the present invention. Detailed Implementation

[0016] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0017] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations of the present invention.

[0018] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0019] Example 1 This embodiment discloses a method for synergistic optimization of the structural performance of the body-in-white and battery pack of a new energy vehicle.

[0020] To more clearly illustrate this embodiment, the process of synergistic optimization of the structural performance of the new energy vehicle body-in-white and battery pack can be described in detail as follows: Methods for synergistic optimization of the structural performance of the body-in-white and battery pack in new energy vehicles include: S1. Construct an integrated finite element model of the body-in-white and battery pack; S2. Based on the integrated finite element model of the body-in-white and battery pack, structural performance analysis is performed on the body-in-white and battery pack. Specifically: modal analysis is performed on the body-in-white to obtain modal data to solve the set conditions and determine the first torsional mode and the first bending frequency; stiffness analysis is performed on the body-in-white to calculate and obtain the bending stiffness and torsional stiffness as optimization constraints; plate-based sensitivity analysis is performed on the body-in-white to determine the optimization variables; constrained modal analysis is performed on the battery pack to determine the first modal frequency and shell stiffness of the battery pack; static analysis is performed on the battery pack to determine and optimize the structural strength and deformation characteristics of the battery pack under extreme conditions. S3. Construct a particle swarm optimization-neural network hybrid model and use the particle swarm optimization algorithm to optimize the initial weights and thresholds of the BP neural network.

[0021] S4. Based on structural performance analysis, with the optimization objectives of maximizing the bending stiffness of the body-in-white and the first-order modal frequency of the battery pack, a Pareto front solution set is generated using an improved multi-objective genetic algorithm. S5. Adjust the thickness of 12 key plates based on the Pareto front solution set to determine the optimal combination of plate thickness parameters.

[0022] like Figure 1 As shown, in step S1, an integrated finite element model of the body-in-white and the battery pack is constructed.

[0023] S1.1. Use the OptiStruct solver in HyperMesh software to create a finite element model of the body-in-white and use shell elements for simulation.

[0024] Body-in-white (BIW) is the main structure of a vehicle and mainly consists of seven parts: the body floor, body side panels, roof system, front compartment system, rear compartment system, body reinforcement structure, and instrument panel bracket and crossbeams.

[0025] In this embodiment, during the simulation analysis, the deformation of the body parts is linear elastic deformation. Therefore, the material cards of the body parts are all set to MAT1, and the attributes are set according to the specific parameters of the corresponding materials. The rigid connections of weld points and weld seams are simulated using RBE2 / RBE3 rigid elements, while other flexible connections are simulated using CBUSH / CBUSH1D elements or CELAS / CDAMP spring-damping elements.

[0026] Following the steps outlined above, the finite element model of the body-in-white has been successfully established. Subsequent steps include conducting modal and stiffness analyses to evaluate its structural performance in detail, and then coupling it with the power battery pack model to achieve structural optimization design.

[0027] S1.2 The finite element model of the battery pack adopts the power battery model and uses solid elements and BEAM elements for simulation.

[0028] The power battery pack of new energy vehicles mainly consists of battery cells, cooling system, battery management system (BMS), electrical connections and wiring, battery module mounting bracket, top housing of high voltage assembly, upper cover and lower housing.

[0029] A power battery model was used, with individual cells simulated using a combination of solid elements and BEAM elements. The material cards were set to *MAT_PLASTIC_BATTERY. The battery module mounting bracket, the top housing of the high-voltage assembly, the upper cover, and the lower housing were simulated using shell elements, with all material cards set to MAT1. RB3 elements were used to simulate the support and connection functions of the potting section, and the remaining parts were simulated using solid elements of equal mass.

[0030] After the above steps, the finite element model of the power battery pack for new energy vehicles has been successfully established. The subsequent work will involve constrained modal analysis and static analysis under operating conditions to evaluate its structural performance in detail, and then couple it with the established body-in-white model for structural optimization design.

[0031] In step S2, structural performance analysis is performed on the body-in-white and the battery pack based on the integrated finite element model of the body-in-white and the battery pack.

[0032] S2.1 Perform modal analysis on the body-in-white to obtain modal data to solve the set conditions and determine the first torsional mode and the first bending frequency.

[0033] First, modal analysis was performed on the body-in-white, solving for the first 20 modal data in the frequency range of 0-200Hz. Some modal values ​​and mode shapes are shown in Table 1. Figure 2 As shown.

[0034] Table 1. First eight modal frequencies and mode shapes of the body-in-white

[0035] Secondly, based on modal data analysis, the first-order torsional mode and first-order bending frequency of the body-in-white were preliminarily determined.

[0036] Modal data analysis revealed that first-order torsional and bending mode values ​​below 30Hz would cause resonance between the vehicle body and powertrain and external forces such as road surface and wheel unevenness. Prolonged and severe resonance could lead to fatigue damage to the vehicle body structure and reduced overall vehicle stiffness. Therefore, the values ​​of the first-order torsional and bending modes were closely monitored during the analysis and optimization process.

[0037] The first-order torsional and first-order bending modal frequencies vary depending on the vehicle model. In practical engineering applications, the first-order torsional frequency is generally concentrated around 35Hz, while the first-order bending frequency is concentrated between 40-50Hz. Through comparative analysis, it can be preliminarily determined that the first-order torsional and first-order bending frequencies of this model are within a reasonable range, fully verifying the effectiveness of this model in the simulation analysis process.

[0038] After the above steps, modal analysis can provide detailed information on the vibration characteristics of the vehicle body at different frequencies, providing a basis for modal design of connected components and systems, and preventing resonance between the vehicle body modes and the connected systems.

[0039] S2.2 Perform static analysis on the body-in-white stiffness. By setting bending stiffness and torsional stiffness analysis conditions, calculate the bending stiffness and torsional stiffness, and calculate and obtain the stiffness parameters as optimization constraints. Insufficient body rigidity can exacerbate body deformation, reduce the durability of critical components and load-bearing connections, thereby affecting the overall quality of the vehicle. It may also lead to a lower natural frequency of the body, making it more susceptible to resonance when subjected to external or internal excitation frequencies. Therefore, it is essential to analyze the bending and torsional stiffness of the body.

[0040] In this embodiment, the specific process is as follows: S2.2.1 Set up the bending stiffness and torsional stiffness analysis conditions respectively.

[0041] (1) Set up the bending stiffness analysis conditions.

[0042] Bending stiffness analysis simulates the vertical downward load generated when the vehicle body is fully loaded with passengers, fully loaded with cargo, or experiencing road bumps. Under these conditions, the vehicle body will undergo bending deformation.

[0043] In this embodiment, the boundary conditions for the bending stiffness analysis are set as follows: the left front shock absorber is constrained to translate in one direction (Z), the right front shock absorber is constrained to translate in both directions (Y and Z), the left rear shock absorber is constrained to translate in both directions (X and Z), and the right rear shock absorber is constrained to translate in all three directions (X, Y, and Z); a load of 1500N is simultaneously applied at the middle position of the left and right longitudinal beams of the door frame.

[0044] (2) Set the torsional stiffness analysis conditions.

[0045] Torsional stiffness analysis simulates the situation where a vehicle's left and right wheels experience uneven forces when one tire passes over a bumpy or uneven road surface, resulting in the vehicle body being subjected to asymmetrical loads. Under such conditions, the vehicle body is prone to torsional deformation, leading to a deterioration in the vehicle's structural performance.

[0046] In this embodiment, the working conditions for torsional stiffness analysis are set as follows: the left rear shock absorber is constrained to translate in the X, Y, and Z directions, and the right rear shock absorber is constrained to translate in the Y and Z directions; a torque of 2000 N / m is applied to the left and right sides of the front shock absorber.

[0047] S2.2.2 Based on the bending stiffness and torsional stiffness analysis conditions, calculate and obtain stiffness parameters as optimization constraints.

[0048] (1) Obtain the Z-direction displacement cloud map by finite element solver and extract the displacement values ​​of key measuring points.

[0049] After setting up the bending stiffness analysis conditions, import the file into the solver for calculation. Use post-processing software to set the calculation result type to Z-direction displacement, such as... Figure 3 As can be seen from the displacement cloud diagram, the overall displacement of the vehicle body is smooth without any major abrupt changes. The displacement is mainly concentrated near the loading points of the longitudinal beams of the left and right door frames, with the largest displacement along the negative Z-axis. Therefore, the Z-direction displacement of the two loading points can be used as the parameter basis for subsequent stiffness calculation of the bending stiffness value.

[0050] Import the torsional stiffness analysis case file into the solver for calculation, and use the post-processing software to view the Z-direction displacement contour plot of the vehicle body. The displacement contour plot clearly shows that the vehicle displacement gradually changes smoothly from the front bumper beam to the rear main body, with roughly symmetrical left and right displacements without significant abrupt changes. The Z-direction displacement is mainly concentrated at the loading points on both sides.

[0051] (2) Calculate the bending stiffness and torsional stiffness based on the displacement values ​​of the key measuring points.

[0052] 1) Calculate the bending stiffness based on the displacement values ​​of key measuring points.

[0053] The Z-direction displacements of the two loading points are taken as parameters for subsequent bending stiffness calculations. The bending stiffness calculation formula is as follows: (1) Where W is the bending stiffness value, in N / mm; F ​​is the sum of the applied load forces, in N; and б is the average of the absolute values ​​of the Z-direction displacements of the two loading points, in mm.

[0054] 2) Calculate the torsional stiffness based on the displacement values ​​of key measuring points.

[0055] First, the torque is calculated based on the torsional stiffness analysis of the working condition.

[0056] The formula for calculating torque is: (2) Where T is torque, in N / m; F is the load applied by the shock absorber, in N; and L is the Y-direction distance between the left and right shock absorbers, in m.

[0057] According to equation (2), the load applied to the left and right shock absorbers in opposite directions can be calculated to be 2072.8 N, specifically as follows: Figure 4 As shown.

[0058] Secondly, the torsion angle is calculated based on the displacement values ​​of key measuring points.

[0059] The torsion angle can be calculated by simplifying the vehicle body calculation, using the Z-axis displacement of the loading point and the Y-axis distance between the front shock absorbers. A simplified diagram is shown below. Figure 5 As shown.

[0060] The formula for calculating the torsion angle is: (3) in, θ Z1 and Z2 are the absolute values ​​of the Z-direction displacement of the loading points, in mm; L is the Y-direction distance between the two loading points, in mm.

[0061] Finally, the torsional stiffness is calculated based on the torque and torsional angle.

[0062] The formula for calculating torsional stiffness is: (4) Where K is the torsional stiffness, in N. m / °; T is the applied torque, in N / m; θ is the torsional angle, in degrees (°).

[0063] (3) Compare the calculated bending stiffness and torsional stiffness with the design target values ​​as constraints for the optimization process.

[0064] Substituting the corresponding loading point displacement and loading point spacing into formulas (1) and (4), the calculation results are shown in Table 2.

[0065] Table 2 Results of Bending Stiffness and Torsional Stiffness Analysis

[0066] The results show that the torsional stiffness of the body-in-white meets the design requirements of the development stage, but the bending stiffness obviously has not reached the target value, and there is a large room for optimization to further improve the performance.

[0067] S2.3. Perform panel sensitivity analysis on the body-in-white to determine the optimization variables.

[0068] The purpose of panel sensitivity analysis is to screen out the key panels that have the most significant impact on the output target from a large number of panel variables, accurately select optimization variables, and help find the best solution.

[0069] The structural parameter variable used is the thickness of the plate. The experimental variable is based on the initial plate thickness with a fluctuation of 20% above and below. The response is set as the first bending mode frequency, mass, first torsional mode frequency, bending stiffness and torsional stiffness for each group of variables.

[0070] The optimization variables include: the thickness of the roof, the left C-pillar reinforcement, the left rear wheel cover, the left longitudinal beam reinforcement, the left B-pillar reinforcement, the left B-pillar lower guard plate, the rear motor cover, the floor, the trunk rear panel, the floor crossbeam, the left door sill inner panel, and the battery pack upper shell.

[0071] Due to the large number of vehicle body panels and the massive computational workload, 38 key body panels were selected as the research objects to improve computational efficiency. For body panels that are symmetrical on the left and right sides, they were set as symmetry variables and moved to the same component layer for unified management, as shown in Table 3, which lists the name, code, and specific parameter values ​​of each body panel.

[0072] Table 3. Variable Design for Sensitivity Analysis

[0073] During sensitivity calculations, both positive and negative values ​​may appear simultaneously. A positive value indicates a positive correlation between the design variable and the response; that is, as the variable value increases, the response value also increases. A negative value indicates a negative correlation; that is, as the variable value increases, the response value decreases. Relying solely on sensitivity analysis results to select optimization variables will not intuitively reflect the influence between each response and the mass response, resulting in insignificant weight reduction of the optimized vehicle body. Therefore, relative sensitivity is used to select subsequent optimization variables. The formula for calculating relative optimization changes is: (5) in, S` i The relative sensitivity of each response; S i For each response sensitivity; S m For quality sensitivity.

[0074] To improve overall vehicle structural performance and achieve vehicle lightweighting, the two types of sheet metal components with the highest relative sensitivity analysis results for each response were selected for subsequent structural performance optimization, while the two types with the lowest sensitivity were selected for overall vehicle lightweighting optimization. Ultimately, 16 key sheet metal components, including the roof, C-pillar reinforcement, and rear wheel arches, were identified as optimization variables. Figure 6 and Figure 7 The results of the sensitivity analysis are shown. A total of 16 types of plates were selected as candidate variables for subsequent optimization. For the first-order torsional modal response, four types of plates were selected as candidate variables: T8.1 (8.2), T5.1 (5.2), T1, and T11.1 (11.2). For the first-order bending mode, four types of plates were selected as candidate variables: T22, T23.1 (23.2), T1, and T18. For the torsional stiffness response, four types of plates were selected as candidate variables: T8.1 (8.2), T20, T1, and T22. For the bending stiffness response, four types of plates were selected as candidate variables: T10.1 (10.2), T15.1 (15.2), T1, and T19.

[0075] S2.4 Perform constrained modal analysis on the battery pack to obtain modal data that solves the set conditions, and determine the first-order modal frequency and shell stiffness of the battery pack.

[0076] First, constrained modal analysis is performed on the battery pack to obtain modal data with a frequency range of 0-200Hz and the first 20 orders.

[0077] During normal driving, a car is subjected to external excitations such as road surface and wheel unevenness. These external excitations are transmitted to the battery pack through the vehicle body structure. At the same time, the battery pack is also affected by the operating frequency of each drive system.

[0078] Constrained modal analysis was performed on the battery pack, simulating its installation on the vehicle body. The results of the constrained modal analysis were compared with the road excitation frequency and the operating frequencies of each powertrain system to avoid resonance phenomena that could affect the battery pack's lifespan.

[0079] (1) Full constraints are applied to the 6 degrees of freedom of the 34 mounting holes of the battery pack.

[0080] (2) Perform modal analysis in the frequency range of 0-200Hz and extract the number of the first 20 modes.

[0081] The 34 mounting holes on the front, back, left, right sides, and top of the battery pack were fully constrained across 6 degrees of freedom. The frequency range was 0-200Hz, and the first 20 orders were used for the solution. The external excitation frequency received by an economical electric vehicle traveling at high speed on a flat road is approximately 20Hz. Therefore, only the first six modal frequencies were extracted for reference analysis, and the results are shown in Table 4.

[0082] Table 4. First six constrained modes and mode shapes of the battery pack

[0083] Secondly, the results of the battery pack constrained modal analysis are compared with the road excitation frequency and the operating frequencies of each power system to determine the first-order modal frequency and shell stiffness of the battery pack.

[0084] The battery pack is primarily subjected to external excitation from road surface and wheel unevenness, as well as operating frequency excitation from various drive systems. The main consideration is the external excitation transmitted to the battery pack via the vehicle body structure from the road surface vibration frequency. Table 5 lists common road surface excitation frequencies in China. Comparative analysis shows that the first-order mode excitation frequency of the battery pack is lower than that of corrugated and smooth roads, making resonance highly likely during driving. Therefore, constrained modal analysis preliminarily concludes that the upper shell of the battery pack has relatively weak stiffness, and structural optimization can be used to improve its structural strength in the future.

[0085] Table 5 Common road surface excitation frequency values ​​in China

[0086] S2.5. Perform static analysis on the battery pack. By setting five typical working condition boundary conditions, determine and optimize the structural strength and deformation characteristics of the battery pack under extreme conditions.

[0087] (1) Set five typical working condition boundary conditions, including vertical bump, bump + emergency braking, and bump + sharp turn working conditions.

[0088] During actual vehicle operation, various complex road conditions and extreme operating conditions will be encountered, posing a severe test to the rigidity performance of the vehicle body and battery pack housing. Static analysis focuses on the structural strength, deformation, and safety performance of the battery pack under static mechanical conditions.

[0089] Taking into account complex road conditions and extreme working conditions, and based on a review and analysis of relevant literature, five typical working conditions were selected as the subjects of subsequent research. Table 6 shows the boundary condition settings for these five working conditions.

[0090] Table 6 Static Analysis Working Condition Setting Parameters

[0091] (2) Apply full constraints to the 34 mounting holes of the battery pack to simulate the actual installation state and obtain displacement cloud diagram and stress distribution. Boundary conditions were set in the HyperMesh preprocessing software to apply full constraints to the 34 original mounting holes on the front, back, left and right sides and top of the battery pack, restricting its 6 degrees of freedom to simulate the vehicle driving state, and the output results were defined as displacement deformation and stress distribution.

[0092] (3) Based on the displacement cloud map and stress distribution, determine the input conditions for battery pack structure optimization.

[0093] To ensure that the battery pack housing meets the strength and stiffness requirements under various extreme working conditions and effectively avoids structural damage, the design specifies that its maximum displacement deformation shall not exceed 3mm, and the maximum stress shall be lower than the material yield strength (210MPa).

[0094] As shown in Figures 8(a), 8(b), 9(a), 9(b), 10(a), 10(b), 11(a), 11(b), 12(a), and 12(b), the displacement under the five operating conditions is mainly concentrated in the center of gravity region. The maximum displacement deformation is 0.99 mm, occurring under the vertical bumpy condition, and meets the structural strength design requirements. The stress concentration area is mainly located at the top mounting holes. In the "bumpy + sharp left turn" and "bumpy + sharp right turn" conditions, due to the influence of inertia and gravity, stress concentration occurs at the top mounting holes on the right and left sides. In other conditions, the stress concentration point is located at the top mounting holes in the middle of the upper shell. The maximum stress value under the five operating conditions is 192 MPa, occurring in the "bumpy + sharp right turn" condition, but this value is still lower than the material's yield stress. In summary, the structural strength of this battery pack meets the design requirements, and its upper shell has considerable room for optimization.

[0095] In step S3, a particle swarm optimization-neural network hybrid model is constructed, and the initial weights and thresholds of the BP neural network are optimized using the particle swarm optimization algorithm.

[0096] Based on the analysis results of the structural performance of the body-in-white and the battery pack, the thickness of 12 key plates was finally determined as the optimization variable. The specific information and distribution of the variables are shown in Table 7. The optimization objectives were determined to be to maximize the bending stiffness of the body-in-white and the first-order modal frequency of the battery pack. In order to ensure that the optimization process is carried out without sacrificing the performance of other structures, the first-order bending and torsional frequency of the body-in-white, the torsional stiffness of the body-in-white, and the total mass of the body were determined as constraints.

[0097] Table 7 Basic Information on Optimization Variables

[0098] The optimization objective is to maximize the bending stiffness of the body-in-white and the first-order modal frequency of the battery pack. To ensure that the optimization process is carried out without sacrificing the performance of other structures, the first-order bending and torsional frequencies of the body-in-white, the torsional stiffness of the body-in-white, and the total mass of the body are defined as constraints.

[0099] S3.1 Construct a particle swarm-neural network hybrid model.

[0100] (1) Establish a BP neural network structure that includes an input layer, a hidden layer and an output layer.

[0101] Backpropagation (BP) neural networks are multi-layer feedforward neural networks that consist of an input layer, hidden layers, and an output layer. The main principle is to adjust the weights and thresholds through the backpropagation algorithm. Each layer contains multiple independent neurons, and the layers are connected by weights. They are widely used to handle nonlinear problems.

[0102] First, after the input layer receives the raw data, it initializes the data according to the selected adaptive activation function, performs forward propagation to calculate the output of each layer, and quantifies the difference between the predicted value and the true value. The formula is as follows: (6) in, M i Mean absolute error; y i and y^ i These are the actual and predicted values.

[0103] Then, the output value is fed back to the input layer, the error is recalculated, and the process continues with the error and weights from the previous layer. (7) in, M k Mean absolute error; ω k For connection weights; f`(z j ) This is the derivative of the activation function.

[0104] Finally, through each iteration, the weights are continuously redefined and the error is reduced to achieve the best prediction effect. However, in this embodiment, under the special case of multiple inputs and multiple outputs, due to the high computational complexity and the complex correlation between the inputs and outputs, traditional BP neural networks are prone to convergence difficulties and getting trapped in local optima.

[0105] (2) Set the initial parameters of the particle swarm optimization algorithm.

[0106] Particle Swarm Optimization (PSO) is a global optimization algorithm inspired by the social behavior of social animals. It first sets initial parameters, then calculates the fitness value of the current position based on the user-defined optimization conditions, and updates the individual and global optima. During the optimization iteration process, the velocity and position of the particles are updated using the following formula: (8) (9) in, V ij (t+1) , V ij (t) Let be the velocity of the individual particle at times t+1 and t; c 1. c 2 represents the individual learning factor and the social learning factor, used to adjust the weights of their optimal movement position; X ij (t+1) , Xij (t) Let be the position of the individual particle at times t+1 and t; ω Inertial weights are used to control the inertia of particles to maintain their current velocity, balancing global exploration and local exploitation. r 1. r 2 is a random constant, which is mainly used to increase the diversity of the group search; pbest ij The optimal position for the individual; gbest This is the optimal position for the group.

[0107] The iteration terminates when the maximum number of iterations is reached or the change in the globally optimal fitness is less than a threshold, as shown in the formula: (10) in, gbest(fitness t ) , gbest(fitness t-1 ) Let be the fitness values ​​of the global optimal solution for the entire particle swarm at the t-th and t-1-th iterations; The threshold value is set.

[0108] Traditional BP neural networks are prone to convergence difficulties and local optima due to the correlation between initial parameters and input / output. Given the global search capability of the PSO optimization algorithm, it is possible to use the PSO optimization algorithm to optimize the weights and thresholds of the BP neural network, so that the position of each particle corresponds to the weights and thresholds of each BP neural network, and continuously iterate to achieve the optimal effect.

[0109] like Figure 13 As shown, the initial parameters of the PSO optimization algorithm were set. After multiple experiments, ω was set to 0.9, and c1 and c2 were set to 2.

[0110] (3) Encode the weights and thresholds of the BP neural network into particle positions.

[0111] The weights and thresholds of the BP neural network are encoded into the position updates of each particle, so that the position of each particle corresponds to the weights and thresholds of each BP neural network.

[0112] (4) Using the error function of the BP neural network as the fitness function, the initial parameters of the neural network are iteratively optimized by the particle swarm optimization algorithm.

[0113] The error function of a backpropagation (BP) neural network is used as the basis for calculating the fitness of each particle. The process is iterated continuously to achieve the optimal result.

[0114] (5) Train and predict the particle swarm-neural network hybrid model.

[0115] The constructed particle swarm optimization-backpropagation (PSO-BP) hybrid model is trained, and the weights and thresholds of the BP neural network are continuously updated. The iteration terminates when the termination condition is met.

[0116] Using 80% of the original data as the training set and 20% as the test set, we used a traditional BP neural network and a pre-constructed PSO-BP hybrid model to predict four constraints and two optimization objectives.

[0117] To verify the applicability and convergence of the improved prediction model, such as Figure 14 As shown in the fitness curve, the fitness value drops sharply in the initial stage of iteration, indicating that the optimization efficiency is high in the early stage. After the 20th iteration, it gradually stabilizes and the fitness value reaches about 0.25. Therefore, it can be proved that the improved model has high matching performance and stable convergence.

[0118] To further verify the rationality of the regression model, after the model iteration was completed, regression scatter plots and error comparison plots for the two optimization objectives in the test set were generated. As shown in Figures 15(a) and 15(b), based on the observation results of the regression scatter plots and error plots, the improved prediction model has good prediction performance in the prediction process of both optimization objectives. The fitted curves are uniformly close to the Y=X curve and both show the advantages of high accuracy and low bias.

[0119] As shown in Figures 16(a) and 16(b), the maximum absolute error of the first-order mode prediction of the battery pack in the BP prediction model is 1.04%, and the error curve fluctuates wildly. However, in the optimized PSO-BP prediction model, the maximum error is only 0.76%, and the error curve is relatively stable. Since the bending stiffness value is much larger than the modal frequency value, its prediction results further verify the robustness of PSO-BP in complex nonlinear problems. The bending stiffness shows a large error in the BP prediction model, with a maximum absolute error of 6.68% and a large range of error curve fluctuations. In the optimized PSO-BP prediction model, the maximum absolute error is only 0.21%. The above is sufficient to prove that the improved PSO-BP hybrid model has higher accuracy and stability.

[0120] In step S4, based on structural performance analysis, with the optimization objectives of maximizing the bending stiffness of the body-in-white and the first-order modal frequency of the battery pack, a Pareto front solution set is generated using an improved multi-objective genetic algorithm.

[0121] The Multi-Objective Genetic Algorithm (NSGA-II) is an efficient and accurate optimization algorithm capable of adapting to complex constraints and multi-objective requirements. An improved PSO-BP prediction model combined with the NSGA-II optimization algorithm is used to optimize two problems, generating a two-dimensional Pareto front. The specific process is as follows: (1) Set the initial parameters for the multi-objective genetic algorithm.

[0122] Initial parameters were set for the multi-objective genetic algorithm (NSGA-II): population size of 150, crossover probability of 0.8, and mutation probability of 0.05.

[0123] (2) Establish an optimization objective function with the goal of maximizing the bending stiffness of the body-in-white and the first-order modal frequency of the battery pack. Set the first-order torsional frequency, first-order bending frequency, torsional stiffness and body-in-white mass as constraints.

[0124] The objective function is optimized, and the formula is: (11) in, y 1 (X) and y 2 (X) These are optimization objective 1 and optimization objective 2, respectively. X M2 This is the first-order torsional frequency value; X M3 This represents the first-order bending torsional frequency value. X m For the body-in-white weight; X n This represents the torsional stiffness value of the body-in-white.

[0125] (3) Based on the multi-objective genetic algorithm, the particle swarm-neural network hybrid model is iteratively optimized.

[0126] Based on the pre-configured multi-objective genetic algorithm (NSGA-II), 100 optimization iterations are performed to obtain... Figure 17 Pareto front solution set and Table 8 Comparison of performance parameters before and after optimization.

[0127] Table 8 Comparison of performance parameters before and after optimization

[0128] Based on the solution set of independent variables, the plate was modified sequentially and the simulation values ​​of each performance were recalculated. Figure 18It can be observed that the vehicle body displacement changes are symmetrically distributed, and the Z-direction displacement of both loading points decreases. As shown in Figures 19(a) to 19(f), the battery pack modal results show that the low-order modal frequencies are still concentrated in the upper shell, but all frequencies have been significantly improved. According to the comparison of the results in Table 8, the first-order frequency of the battery pack for Target 1 has been improved by 28.38% (6.51Hz), far exceeding the common road excitation frequency value, which greatly reduces the risk of resonance. The bending stiffness for Target 2 has been improved by 5.57% (490.9N / mm), exceeding the target value of 287.9N / mm, and the total vehicle weight has decreased by 0.22% (2kg). Therefore, this optimization has significantly improved the structural performance of this pure electric vehicle under the premise of considering vehicle lightweighting.

[0129] In step S5, the thickness of 12 key plates is adjusted based on the Pareto front solution set to determine the optimal combination of plate thickness parameters.

[0130] The optimal solution set of independent variables corresponding to the optimal solution of the optimization objective is the optimal combination of plate thickness parameters [4.38, 1.33, 0.96, 1.4, 1.86, 2.1, 0.97, 0.62, 1.06, 1.38, 1.37, 0.44].

[0131] The plates are modified sequentially based on the optimal combination of plate thickness parameters.

[0132] To address the need for coordinated optimization of lightweighting and structural reliability in new energy vehicles, a multi-objective optimization framework integrating the vehicle body and battery pack was constructed. A finite element model covering body-in-white modes, static stiffness, and battery pack constraint modes was established based on a specific vehicle model. Static simulation analyses under five operating conditions were performed using the integrated body-in-white and battery pack finite element models. A novel approach was adopted: integrating the PSO-BP neural network with an improved NSGA-II algorithm to collaboratively optimize the thickness of 12 key plate components. After optimization, the bending stiffness of the body-in-white increased by 5.57% to 9287.9 ​​N / mm, the first-order modal frequency of the battery pack increased by 28.38% to 29.45 Hz, and the weight of the body-in-white was reduced by 2 kg (0.22%). This embodiment overcomes the limitations of traditional experimental methods in terms of operational complexity and time consumption, achieving structural performance balance between the body-in-white and battery pack through Pareto front solution sets, providing an effective solution for the collaborative design of multiple systems in new energy vehicles.

[0133] Example 2 The purpose of this embodiment is to provide a system for synergistic optimization of the structural performance of the body-in-white and battery pack of a new energy vehicle, including: The finite element model building module is used to build an integrated finite element model of the body-in-white and the battery pack; The structural performance analysis module is used to perform structural performance analysis on the body-in-white and battery pack based on an integrated finite element model. Specifically, it performs modal analysis on the body-in-white to obtain modal data to solve set conditions and determine the first torsional mode and first bending frequency; it performs stiffness analysis on the body-in-white to calculate and obtain bending stiffness and torsional stiffness as optimization constraints; it performs plate-based sensitivity analysis on the body-in-white to determine optimization variables; it performs constrained modal analysis on the battery pack to determine the first modal frequency and shell stiffness; and it performs static analysis on the battery pack to determine and optimize the structural strength and deformation characteristics of the battery pack under extreme conditions. The multi-objective collaborative optimization module is used to construct a particle swarm-neural network hybrid model, and to optimize the initial weights and thresholds of the BP neural network using the particle swarm algorithm; Based on structural performance analysis, with the optimization objectives of maximizing the bending stiffness of the body-in-white and the first-order modal frequency of the battery pack, a Pareto front solution set is generated using an improved multi-objective genetic algorithm. Based on the Pareto front solution set, the thickness of 12 key plates was adjusted to determine the optimal combination of plate thickness parameters.

[0134] The method steps in Example 1 are implemented based on a system for collaborative optimization of the structural performance of the body-in-white and battery pack of new energy vehicles.

[0135] Example 3 The purpose of this embodiment is to provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the above-described method.

[0136] Example 4 The purpose of this embodiment is to provide a computer-readable storage medium.

[0137] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the above method.

[0138] Example 5 The purpose of this embodiment is to provide a computer program product containing instructions that, when run on a computer, cause the computer to perform the methods and functions involved in any of the above embodiments. The steps and methods involved in the apparatus of the above embodiments correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0139] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0140] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for synergistic optimization of the structural performance of the body-in-white and battery pack of a new energy vehicle, characterized in that, include: Construct an integrated finite element model of the body-in-white and battery pack; Based on the integrated finite element model of the body-in-white and battery pack, structural performance analysis is performed on the body-in-white and battery pack. Specifically, modal analysis is performed on the body-in-white to obtain modal data to solve the set conditions and determine the first torsional mode and the first bending frequency; stiffness analysis is performed on the body-in-white to calculate and obtain the bending stiffness and torsional stiffness as optimization constraints; and plate sensitivity analysis is performed on the body-in-white to determine the optimization variables. Constrained modal analysis was performed on the battery pack to determine the first-order modal frequencies and shell stiffness; static analysis was performed on the battery pack to determine and optimize the structural strength and deformation characteristics of the battery pack under extreme conditions. A particle swarm optimization-neural network hybrid model is constructed, and the initial weights and thresholds of the BP neural network are optimized using the particle swarm optimization algorithm. Based on structural performance analysis, the optimization objectives are to maximize the bending stiffness of the body-in-white and the first-order modal frequency of the battery pack. A Pareto front solution set is generated using an improved multi-objective genetic algorithm. Based on the Pareto front solution set, the preset thickness of key plates is adjusted to determine the optimal combination of plate thickness parameters.

2. The method for synergistic optimization of the structural performance of the body-in-white and battery pack of a new energy vehicle as described in claim 1, characterized in that, The specific process for constructing an integrated finite element model of the body-in-white and battery pack is as follows: A finite element model of the body-in-white was created using the OptiStruct solver in HyperMesh software, employing a shell element model. The finite element model of the battery pack adopts the power battery model and is simulated using solid cells and BEAM cells.

3. The method for synergistic optimization of the structural performance of the body-in-white and battery pack of a new energy vehicle as described in claim 1, characterized in that, Stiffness analysis was performed on the body-in-white, and the bending stiffness and torsional stiffness were calculated and obtained as optimization constraints. The specific process is as follows: Set up separate analysis conditions for bending stiffness and torsional stiffness; Based on the bending stiffness and torsional stiffness analysis conditions, stiffness parameters are calculated and obtained as optimization constraints.

4. The method for synergistic optimization of the structural performance of the body-in-white and battery pack of a new energy vehicle as described in claim 3, characterized in that, Based on the bending stiffness and torsional stiffness analysis conditions, the stiffness parameters are calculated and obtained as optimization constraints. The specific process is as follows: The Z-axis displacement cloud map is obtained by finite element solver, and the displacement values ​​of key measuring points are extracted. Calculate the bending stiffness and torsional stiffness based on the displacement values ​​of key measuring points; The calculated bending stiffness and torsional stiffness are compared with the design target values ​​as constraints in the optimization process.

5. The method for synergistic optimization of the structural performance of the body-in-white and battery pack of a new energy vehicle as described in claim 1, characterized in that, Constrained modal analysis is performed on the battery pack to solve for the modal data under the given conditions, and to determine the first-order modal frequencies and shell stiffness of the battery pack. The specific process is as follows: The 6 degrees of freedom of the 34 mounting holes of the battery pack are fully constrained; Modal analysis was performed in the frequency range of 0-200Hz to extract the number of the first 20 modes. The results of the battery pack constrained modal analysis were compared with the road excitation frequency and the operating frequencies of each power system to determine the first-order modal frequency and shell stiffness of the battery pack.

6. The method for synergistic optimization of the structural performance of the body-in-white and battery pack of a new energy vehicle as described in claim 1, characterized in that, Static analysis of the battery pack was performed to determine and optimize its structural strength and deformation characteristics under extreme conditions. The specific process is as follows: Five typical working condition boundary conditions are set, including vertical bumps, bumps + sudden braking, and bumps + sharp turns. Full constraints were applied to the 34 mounting holes of the battery pack to simulate the actual installation state and obtain displacement contour maps and stress distribution. Based on displacement contour maps and stress distribution, the input conditions for battery pack structure optimization are determined.

7. The method for synergistic optimization of the structural performance of the body-in-white and battery pack of a new energy vehicle as described in claim 1, characterized in that, A hybrid particle swarm optimization-neural network model is constructed, and the initial weights and thresholds of the backpropagation (BP) neural network are optimized using the particle swarm optimization algorithm. The specific process is as follows: Establish a BP neural network structure that includes an input layer, hidden layers, and an output layer; Set the initial parameters for the particle swarm optimization algorithm; The weights and thresholds of the BP neural network are encoded as particle positions; The initial parameters of the neural network are iteratively optimized using the particle swarm optimization algorithm, with the error function of the BP neural network serving as the fitness function. Training and prediction of a particle swarm-neural network hybrid model.

8. The method for synergistic optimization of the structural performance of the body-in-white and battery pack of a new energy vehicle as described in claim 1, characterized in that, Based on structural performance analysis, with the optimization objectives of maximizing the bending stiffness of the body-in-white and the first-order modal frequency of the battery pack, a Pareto front solution set is generated using an improved multi-objective genetic algorithm. The specific process is as follows: Initial parameter settings for the multi-objective genetic algorithm; An optimization objective function is established and constraints are set. The optimization objectives are to maximize the bending stiffness of the body-in-white and the first-order modal frequency of the battery pack, and the constraints are the first-order torsional frequency, the first-order bending frequency, the torsional stiffness, and the mass of the body-in-white. Based on a multi-objective genetic algorithm, the particle swarm-neural network hybrid model is iteratively optimized to generate a Pareto front solution set.

9. The method for synergistic optimization of the structural performance of the body-in-white and battery pack of a new energy vehicle as described in claim 8, characterized in that, The objective function and constraints are given by the following formula: ； in, y 1 (X) and y 2 (X) These are optimization objective 1 and optimization objective 2, respectively. X M2 This is the first-order torsional frequency value; X M3 This represents the first-order bending torsional frequency value. X m For the body-in-white weight; X n This represents the torsional stiffness value of the body-in-white.

10. A system for synergistic optimization of the structural performance of the body-in-white and battery pack of a new energy vehicle, characterized in that, include: The finite element model building module is used to build an integrated finite element model of the body-in-white and the battery pack; The structural performance analysis module is used to perform structural performance analysis on the body-in-white and battery pack based on the integrated finite element model of the body-in-white and battery pack. Specifically, it performs modal analysis on the body-in-white to obtain modal data to solve the set conditions and determine the first torsional mode and the first bending frequency; it performs stiffness analysis on the body-in-white to calculate and obtain the bending stiffness and torsional stiffness as optimization constraints; and it performs plate-based sensitivity analysis on the body-in-white to determine the optimization variables. Constrained modal analysis was performed on the battery pack to determine the first-order modal frequencies and shell stiffness; static analysis was performed on the battery pack to determine and optimize the structural strength and deformation characteristics of the battery pack under extreme conditions. The multi-objective collaborative optimization module is used to construct a particle swarm optimization-neural network hybrid model. It uses the particle swarm optimization algorithm to optimize the initial weights and thresholds of the BP neural network. Based on structural performance analysis, it takes maximizing the bending stiffness of the body-in-white and the first-order modal frequency of the battery pack as optimization objectives. It generates a Pareto front solution set through an improved multi-objective genetic algorithm. Based on the Pareto front solution set, it adjusts the preset thickness of key plates to determine the optimal combination of plate thickness parameters.

Images