Model-based crash energy absorption component design method, system, medium, and apparatus

Abstract

The application discloses a model-based collision energy-absorbing component design method, system, medium and equipment, comprising randomly generating energy-absorbing component combinations following specific combination rules, setting key structure characteristic parameters according to the internal and external boundary constraints of spatial arrangement to form a first characteristic parameter combination; utilizing parameterization modeling and mesh generation technology to process the parameter combinations to create a finite element analysis model of the energy-absorbing component; performing performance testing on the model through a simulation algorithm to obtain a simulation result; training a neural network based on the simulation result to obtain a design model; encoding the energy-absorbing component and the characteristic parameters thereof into individuals in a genetic algorithm and performing iterative optimization, using the design model to evaluate individual fitness at each iteration, and outputting an optimal energy-absorbing component design when a preset condition is met. The application realizes efficient conversion from performance requirements to energy-absorbing component design by combining optimization algorithms and artificial intelligence models, and improves the design efficiency of vehicle body collision energy-absorbing components.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of automotive engineering, and relates to a model-based crash energy-absorbing component design method, system, medium and equipment. BACKGROUND

[0002] In the field of automotive engineering, the design of crash energy-absorbing components of a vehicle body is crucial to passenger safety. Traditionally, the design process relies on the experience of engineers and simulation methods such as finite element analysis, combined with experimental verification for optimization. Although this method is effective, it often requires a large amount of time and computing resources, especially in the early conceptual design stage, making it difficult to quickly evaluate the energy-absorbing efficiency of different structures.

[0003] The existing technology has some deficiencies. On the one hand, the application of past vehicle data is not high, leading to repetitive labor and prolonged design cycle. On the other hand, for the optimization of complex structures, it relies too much on the experience of engineers, which may result in less than ideal design solutions. If an accurate surrogate model is constructed, a large amount of simulation data is required, which not only consumes time but also costs a lot, especially in the case of tight project schedule, it is difficult to guarantee the design quality and efficiency. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a model-based crash energy-absorbing component design method, system, medium and equipment, which realizes efficient conversion from performance requirements to energy-absorbing component design by combining optimization algorithms and artificial intelligence models, and improves the design efficiency of crash energy-absorbing components of a vehicle body.

[0005] To achieve the above-mentioned purpose, in a first aspect, the application provides a model-based crash energy-absorbing component design method, comprising:

[0006] According to the connection relationship of each energy-absorbing component of each reference vehicle, randomly generate each energy-absorbing component combination conforming to the preset combination rule;

[0007] According to the preset internal and external boundary constraint parameters of the energy-absorbing component space arrangement, set a plurality of key structure feature parameters for each energy-absorbing component in each energy-absorbing component combination to obtain a first feature parameter combination corresponding to each energy-absorbing component combination;

[0008] According to the preset parameterized model generation algorithm and the grid generation algorithm, process each first feature parameter combination to obtain a finite element analysis model of each energy-absorbing component, and according to a preset simulation algorithm, simulate a preset non-energy-absorbing component finite element analysis model and each energy-absorbing component finite element analysis model to obtain each performance simulation result;

[0009] According to the data of each performance simulation result, first feature parameter combination and energy-absorbing component combination, train a preset original neural network model to obtain a design model;

[0010] The energy-absorbing components and the key structural feature parameters are combined as a plurality of population individuals of the genetic algorithm, and the population is iteratively optimized according to a preset genetic operation algorithm until a preset iteration number is reached or the population individuals meet a preset expected performance index, and the current energy-absorbing component combination and the corresponding key structural feature parameters are output as a design result; and in each iteration optimization, the design model is used to process the population individuals to obtain predicted performance indexes of the population individuals, and a fitness value is calculated according to deviations of the predicted performance indexes from the expected performance index.

[0011] Compared with the prior art, the embodiments of the application have the following beneficial effects: random combination ensures the diversity of the design scheme and avoids the influence of the subjective experience of engineers; the key structural feature parameters are set to define specific parameter ranges, so that subsequent simulation and analysis have clear targets and boundary conditions, and unnecessary waste of computing resources is reduced; the actual performance of different design schemes can be predicted in advance through simulation, and the cost and time consumption of physical experiments are reduced; the neural network model is trained in combination with the key structural feature parameters and the simulation results, a large amount of simulation data is processed by using AI technology, the mapping relationship between the input (structural feature) and the output (performance index) is quickly established, and the speed and accuracy of data analysis are improved; the genetic algorithm and the model cooperate with each other to automatically find the optimal solution in combination with the principle of evolution, without manual intervention, thereby further improving the design efficiency and ensuring the quality of the design result.

[0012] In some embodiments of the first aspect of the application, the random generation of the energy-absorbing component combinations that meet the preset combination rules according to the connection relationships of the energy-absorbing components of the reference vehicle types comprises:

[0013] Information of the energy-absorbing components of each reference vehicle type is obtained.

[0014] The information of the energy-absorbing components is converted into numerical parameters according to a preset encoding rule.

[0015] The energy-absorbing component combinations are randomly generated according to the connection relationships of the energy-absorbing components and the preset combination rules.

[0016] Compared with the prior art, the above embodiments have the following beneficial effects: the energy-absorbing component information of the reference vehicle types is converted into numerical parameters, and the component combinations are randomly generated based on the connection relationships and the combination rules, so that the energy-absorbing component designs of different vehicle types can be quickly transplanted and compared, and the repetitive work is reduced; in addition, the use of a unified encoding rule simplifies the subsequent processing process, improves the speed and accuracy of parameterized modeling, and thus speeds up the entire design cycle and reduces the development cost.

[0017] In some embodiments of the first aspect of the application, the inner and outer boundary constraint parameters of the preset spatial arrangement of energy-absorbing components are used to set a plurality of key structural characteristic parameters for each energy-absorbing component in each combination of energy-absorbing components, to obtain a first characteristic parameter combination corresponding to each combination of energy-absorbing components, including:

[0018] The inner and outer boundary constraint parameters of the preset spatial arrangement of energy-absorbing components are used to set a plurality of structural characteristic parameters for each energy-absorbing component;

[0019] According to the preset SGMW sensitivity analysis specification, the importance values of each of the structural characteristic parameters on the preset safety performance indicators are analyzed and calculated;

[0020] The importance values are sorted from large to small, and the structural characteristic parameters corresponding to a plurality of importance values ranked in the front are selected as key structural characteristic parameters.

[0021] Compared with the prior art, the above embodiments have the following beneficial effects: the structural characteristic parameters are set according to the inner and outer boundary constraint parameters, and the key structural characteristic parameters are selected according to the SGMW sensitivity analysis specification, to ensure the importance of the selected parameters on the safety performance indicators; the sorting and selection mechanism ensures that the most important parameters are given priority, thereby improving the safety and reliability of the design scheme; at the same time, it helps to reduce unnecessary complexity and focus the design on the factors that have the greatest impact on the overall performance.

[0022] In some embodiments of the first aspect of the application, the first characteristic parameter combinations are processed according to the preset parameterized model generation algorithm and mesh generation algorithm to obtain a finite element analysis model of each energy-absorbing component, including:

[0023] According to a preset sampling algorithm, the first characteristic parameter combinations are extracted to obtain a plurality of sampled first characteristic parameter combinations;

[0024] The sampled first characteristic parameter combinations are processed according to the preset parameterized model generation algorithm and mesh generation algorithm to obtain a finite element analysis model of each energy-absorbing component.

[0025] Compared with the prior art, the above embodiments have the following beneficial effects: by sampling the first characteristic parameter combinations and using the parameterized model generation algorithm and mesh generation algorithm to construct the finite element analysis model, the computational resource bottleneck problem that may be encountered during full-data analysis is solved, the sampling method not only preserves the core characteristics of the original data but also reduces the computational burden, making large-scale simulation possible and promoting more extensive design exploration and optimization.

[0026] In some embodiments of the first aspect of the application, the training of the preset original neural network model according to the data of the performance simulation results, the first feature parameter combinations and the energy-absorbing component combinations to obtain the design model comprises:

[0027] According to the performance simulation results, the corresponding first feature parameter combinations and the corresponding energy-absorbing component combinations, an original mapping data set is constructed.

[0028] According to a preset preprocessing algorithm, the original mapping data set is processed to obtain a mapping data set; wherein the preprocessing algorithm comprises any one or a combination of more than one of the following: missing value processing, abnormal value processing, repeated value processing, renaming processing, format unification processing, normalization processing, feature encoding processing.

[0029] Compared with the prior art, the above embodiments have the following beneficial effects: on the basis of constructing the original mapping data set, the application of the preprocessing algorithm (such as missing value processing, abnormal value processing, etc.) can significantly improve the quality of the data, eliminate noise interference and enhance the effectiveness of model learning.

[0030] In some embodiments of the first aspect of the application, the training of the preset original neural network model according to the data of the performance simulation results, the first feature parameter combinations and the energy-absorbing component combinations to obtain the design model further comprises:

[0031] According to a preset K-fold cross-validation algorithm, a dropout algorithm and the mapping data set, the original neural network model is trained to obtain the design model.

[0032] Wherein the training adopts a preset AdamW algorithm to optimize the learning rate, and the algorithm is as follows:

[0033] m t =β1m t +(1-β1)g t ; Wherein m 1,init represents the first moment estimation of the current step t, g 2,init represents the gradient of the current step, β1 and β2 respectively represent the momentum decay coefficients of the first order and the second order, and v 1,init represents the second moment estimation of the current step t; wherein β1 and β2 are dynamically adjusted as follows: β1 = β 2,init · (1-t / T); β2 = β target · (1-t / T); wherein β target and β t respectively represent the initial values of the momentum decay coefficients of β1 and β2, and T is the total number of training steps; wherein the gradient clipping algorithm is as follows: Wherein τ represents the clipping threshold of the gradient.

[0034] Compared with the prior art, the above-mentioned embodiment has the following beneficial effects: the neural network model is trained by using K-fold cross-validation, dropout and AdamW optimization algorithm, so that the overfitting phenomenon is avoided, and the model has good generalization ability and stability; the first-order and second-order momentum decay coefficients are dynamically adjusted, and the gradient clipping algorithm is applied, so that the convergence speed and stability in the model training process are further improved.

[0035] In some embodiments of the first aspect of the application, the fitness value is calculated according to the deviation of each prediction performance indicator from the expected performance indicator, comprising:

[0036] The calculation formula for calculating the fitness value is as follows:

[0037] F(i) = -MSE(y(i), y target ); wherein F(i) represents the fitness value of the individual i of the population, y(i) represents the prediction performance indicator of the designed model, y target represents the expected performance indicator, and MSE represents the minimum mean square error.

[0038] Compared with the prior art, the above-mentioned embodiment has the following beneficial effects: the genetic algorithm and the model work together, automatically find the optimal solution according to the principle of evolution, without human intervention, further improve the design efficiency, and at the same time ensure the quality of the design result.

[0039] Secondly, the application also provides a model-based collision energy-absorbing component design system, comprising: a component combination module, a feature parameter setting module, a simulation module, a model training module and a processing module.

[0040] The component combination module is configured to randomly generate each energy-absorbing component combination according to the connection relationship of each energy-absorbing component of each reference vehicle and in accordance with a predetermined combination rule.

[0041] The feature parameter setting module is configured to set a plurality of key structural feature parameters for each energy-absorbing component in each energy-absorbing component combination according to the preset internal and external boundary constraint parameters of the energy-absorbing component space arrangement, and obtain a first feature parameter combination corresponding to each energy-absorbing component combination.

[0042] The simulation module is configured to process each first feature parameter combination according to a preset parameterized model generation algorithm and a mesh generation algorithm, obtain a finite element analysis model of each energy-absorbing component, and perform simulation on a preset non-energy-absorbing component finite element analysis model and each energy-absorbing component finite element analysis model according to a preset simulation algorithm, and obtain each performance simulation result.

[0043] The model training module is configured to train a preset original neural network model according to data of each performance simulation result, first characteristic parameter combination and energy-absorbing component combination, and obtain a design model.

[0044] The processing module is configured to represent each energy-absorbing component combination and key structure characteristic parameter as a plurality of population individuals of a genetic algorithm, and iteratively optimize the population according to a preset genetic operation algorithm until a preset iteration number is reached or the population individuals meet a preset expected performance index, and output the current energy-absorbing component combination and corresponding key structure characteristic parameters as a design result; and at each iteration optimization, process the population individuals according to the design model to obtain predicted performance indexes of the population individuals, and calculate fitness values according to deviations between the predicted performance indexes and the expected performance index.

[0045] Compared with the prior art, the above embodiments have the following beneficial effects: random combination generation ensures the diversity of design schemes and avoids the influence of subjective experience of engineers; key structure characteristic parameters are set to define specific parameter ranges, so that subsequent simulation and analysis have clear targets and boundary conditions, and unnecessary waste of computing resources is reduced; simulation can predict the actual performance of different design schemes in advance, reducing the cost and time consumption of physical experiments; a neural network model is trained in combination with key structure characteristic parameters and simulation results, AI technology is used to process a large amount of simulation data, the mapping relationship between input (structure characteristic) and output (performance index) is quickly established, and the speed and accuracy of data analysis are improved; genetic algorithm and model cooperate to automatically find the optimal solution in combination with the principle of evolution, without manual intervention, further improving the design efficiency and ensuring the quality of the design result.

[0046] In a third aspect, the present application further provides a model-based collision energy-absorbing component design device, which comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the computer program implements the steps of the model-based collision energy-absorbing component design method when loaded into the processor.

[0047] In a fourth aspect, the present application further provides a computer readable storage medium, which stores a computer program, and the computer program implements the steps of the model-based collision energy-absorbing component design method when executed by a processor. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 FIG. 1 is a flowchart of a model-based collision energy-absorbing component design method provided in some embodiments of the present application.

[0049] Figure 2A structural schematic diagram of a model-based crash energy absorption component design system provided in some embodiments of the present application.

[0050] Figure 3 A structural diagram of a model-based crash energy absorption component design device provided in some embodiments of the present application. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0052] Embodiment one:

[0053] Please refer to Figure 1 A model-based crash energy absorption component design method provided in the embodiments of the present application includes steps S1 to S5:

[0054] Step S1: According to the connection relationship of each energy absorption component of each reference vehicle type, randomly generate each energy absorption component combination conforming to the preset combination rule.

[0055] Preferably, in some embodiments of the present application, step S1 can be implemented through the following preferred implementation, including steps S11-S13, specifically as follows:

[0056] S11: Obtain the information of each energy absorption component of each reference vehicle type;

[0057] S12: According to the preset coding rule, convert the information of each energy absorption component into a numerical parameter;

[0058] S13: According to the connection relationship of each energy absorption component and the preset combination rule, randomly generate each energy absorption component combination.

[0059] For example, in specific implementation, the crash beam number in the energy absorption component is defined as A, then the corresponding crash beam numbers of different vehicle types are A1, A2, …, An, the energy absorption box number in the energy absorption component is defined as B, then the corresponding energy absorption box numbers of different vehicle types are B1, B2, …, Bn, the front longitudinal beam number is C, and the corresponding front longitudinal beam numbers of different vehicle types are C1, C2, …, Cn, the subframe number is D, and the corresponding subframes of different vehicle types are D1, D2, …, Dn. Considering the combination between different types of components, according to the lapping structure and linking mode of the energy absorption component, the possible vehicle body crash energy absorption component combination scheme is determined, such as A1 B2C3D5, and a new energy absorption component combination scheme is obtained by combination.

[0060] In the preferred embodiment, the energy absorption component information of each reference vehicle model is converted into numerical parameters, and the component combinations are randomly generated based on the connection relationship and combination rules, so that the energy absorption component design between different vehicle models can be quickly transplanted and compared, reducing repetitive work. In addition, the use of a unified coding rule simplifies the subsequent processing flow, improves the speed and accuracy of parameterized modeling, thereby speeding up the entire design cycle and reducing development costs.

[0061] Step S2: According to the internal and external boundary constraint parameters of the preset energy absorption component space arrangement, set a plurality of key structure characteristic parameters for each energy absorption component in each energy absorption component combination, and obtain a first characteristic parameter combination corresponding to each energy absorption component combination.

[0062] Preferably, in some embodiments of the present application, step S2 can be implemented through the following preferred implementation, including steps S21 to S23, as follows:

[0063] S21: According to the internal and external boundary constraint parameters of the preset energy absorption component space arrangement, set a plurality of structure characteristic parameters for each energy absorption component;

[0064] For example, in specific implementation, the thickness of the anti-collision beam part T1, the thickness of the energy absorption box part T3, the thickness of the front longitudinal beam inner and outer plates T3 / T4, the thickness of the reinforcing plate part T5, the thickness of the auxiliary frame T6, etc. are taken as thickness characteristic parameters, and the cross-sectional dimensions of the energy absorption components: the cross-sectional height and width of the anti-collision beam H1 / H2, the cross-sectional width and height of the energy absorption box H3 / H4, the cross-sectional height and width of the front longitudinal beam H5 / H6, the cross-sectional width and height of the auxiliary frame H7 / H8, etc. are taken as cross-sectional characteristic parameters, the length and width of the key rib of the energy absorption box and the front longitudinal beam L1 / L2 are taken as structure characteristic variables, and the material elastic modulus E1 / E2 / … / En, material stress and strain curve LC1 / LC2… of the energy absorption component are taken as material control parameters.

[0065] S22: According to the preset SGMW sensitivity analysis specification, analyze and calculate the importance value of each structure characteristic parameter to the preset safety performance index;

[0066] S23: Sort the importance values from large to small, and select the structure characteristic parameters corresponding to the importance values in the front as key structure characteristic parameters.

[0067] In a specific implementation, the impact sensitivity analysis can output scatter plot matrix, impact bar chart, radar chart, dimension reduction projection chart and other visual analysis charts, to study the influence of different structural characteristic parameters on safety performance indicators (such as energy absorption, collision acceleration, mass, etc.), and the geometric constraints and interference risks of the structural characteristic parameters after combination, so as to determine the structural characteristic parameters and their ranges of the energy absorption component, filter out the structural characteristic parameters with less influence, and obtain a key structural characteristic parameter table, such as the structural characteristic parameters affecting the safety performance indicators in the top 15 or top 20.

[0068] In the preferred embodiment, steps S21-S23 set the structural characteristic parameters according to the inner and outer boundary constraint parameters, and select the key structural characteristic parameters according to the SGMW sensitivity analysis specification, to ensure the importance of the selected parameters for the safety performance indicators; the sorting selection mechanism ensures that the most important parameters are given priority, thereby improving the safety and reliability of the design scheme; at the same time, it helps to reduce unnecessary complexity and focus the design on the factors that have the greatest impact on the overall performance.

[0069] Step S3: According to the preset parameterized model generation algorithm and the grid generation algorithm, processing each first characteristic parameter combination to obtain a finite element analysis model of each energy absorption component, and according to a preset simulation algorithm, simulating the preset non-energy absorption component finite element analysis model and each energy absorption component finite element analysis model to obtain each performance simulation result.

[0070] Preferably, in some embodiments of the present application, the finite element analysis model of each energy absorption component in step S3 can be obtained by the following preferred implementation, including steps S31-S32, as follows:

[0071] S31: According to a preset sampling algorithm, extracting the first characteristic parameter combination to obtain a plurality of sampled first characteristic parameter combinations;

[0072] S32: According to the preset parameterized model generation algorithm and the grid generation algorithm, processing each of the sampled first characteristic parameter combinations to obtain a finite element analysis model of each energy absorption component.

[0073] In specific implementation, various sampling methods such as optimized Latin hypercube, Hammersley sampling, etc. can be used for sampling. After sampling, the parametric model of the developed vehicle and the benchmark vehicle body crash energy absorption components (crash beam, energy absorption box, front longitudinal beam, subframe, etc.) is established by using computer-aided parametric design software (SFE Concept) in the form of base point, base line and cross section scanning, mapping, etc. That is, in the SFE Concept software, first, the base point (Influence Point) of the crash energy absorption component is created for positioning the component position and component length, then the base line (Base Line) is generated by connecting the base points for component sweeping, the feature creation of the energy absorption component is realized by creating the base section (Base Section), and the creation of the crash key energy absorption component is realized by inserting multiple base sections in the form of base section sweeping along the base line; the interface between components is established by the parametric model, the rapid combination of the body energy absorption components is realized by using rigid, patch and common node, the input of the structural feature parameters is realized by the variable definition in the parametric model, and the model combination and adjustment time is shortened. In the parametric model, the information of the part features in the base section, the base point position, the part thickness, etc. is recorded as corresponding variables by the variable recording function, and the parameter variables are changed by modifying the recorded parameters. In the SFE Concept software, the parametric model of the body crash energy absorption component is converted into a finite element model for crash simulation analysis by using the mesh generation function, the connection between the finite element models of the body crash energy absorption components and the non-energy absorption components is established in the form of patch, common node and rigid at the boundary of the energy absorption component parametric model, the finite element analysis model for integrated simulation process calculation is obtained, and the integrated simulation analysis process is built in isight using DOE and simcode modules. The SFE Concept software generated mesh is added in the simcode module, the finite element analysis is performed by adding the finite element simulation software, the information transmission is realized by data transmission between the simcode modules, the body crash integrated simulation analysis process is established, the integrated simulation is performed, and the hourglass energy Eh, the mass increase parameter Addmass, the energy absorption Ei, the body acceleration peak value MaxA, the body effective acceleration EA, the vehicle crushing amount Displacement, the body mass mass, etc. are obtained in the simulation results of the body crash performance, wherein the ratio of the hourglass energy to the total energy is required to be less than 5%, the mass increase parameter is required to be less than the standard value 5%, and the normal operation of the vehicle crash simulation is ensured.

[0074] In the preferred embodiment, steps S31-S32 solve the computing resource bottleneck problem that may be encountered during full data analysis by sampling the first feature parameter combination and using a parameterized model generation algorithm and a mesh generation algorithm to construct a finite element analysis model. The sampling method not only retains the core characteristics of the original data, but also reduces the computing burden, making large-scale simulation possible and promoting more extensive design exploration and optimization.

[0075] Step S4: Training the pre-set original neural network model according to the data of the performance simulation results, the first feature parameter combination, and the energy absorption component combination to obtain a design model.

[0076] Preferably, in some embodiments of the present application, step S4 can be obtained by the following preferred implementation, including steps S41-S43, specifically as follows:

[0077] S41: Constructing an original mapping data set according to the performance simulation results and the corresponding first feature parameter combination and the corresponding energy absorption component combination.

[0078] S42: Processing the original mapping data set according to a pre-set preprocessing algorithm to obtain a mapping data set; wherein the preprocessing algorithm includes any one or a combination of the following: missing value processing, outlier processing, duplicate value processing, renaming processing, format unification processing, normalization processing, feature encoding processing.

[0079] For example, in specific implementation, after constructing the original mapping data set, the data set is cleaned and preprocessed. For missing values, such as sample data without results due to calculation failure, they are deleted. For abnormal values, such as noise data with abnormal energy or acceleration during calculation, clustering algorithms or density-based anomaly detection algorithms are used to identify and delete sample data containing abnormal values, thereby improving data quality. For duplicate values, they are identified and deleted by comparing feature values between samples or using a hash algorithm. At the same time, the sample data table header is adjusted to adapt to the AI artificial intelligence algorithm logic. Different formats in the original data are converted into a unified format to ensure data consistency and comparability. Converting data to a distribution with a mean of 0 and a standard deviation of 1 helps to ensure that data has similar weights during training. For category feature data, label encoding, one-hot encoding, and other methods are applied so that the machine learning model can understand and process it. Natural ordering categories use label encoding, unordered categories use one-hot encoding, and frequently occurring features use frequency encoding.

[0080] In the preferred embodiment, steps S41-S42, based on the construction of the original mapping data set, can significantly improve the quality of the data, eliminate noise interference, and enhance the effectiveness of model learning by applying preprocessing algorithms such as missing value processing and outlier processing.

[0081] S43: Train the original neural network model according to the preset K-fold cross-validation algorithm, dropout algorithm and the mapping dataset to obtain the design model;

[0082] The training process employs a pre-defined AdamW algorithm to optimize the learning rate, as follows:

[0083] m t =β1m t-1 +(1-β1)g t ; Where m t G represents the first moment estimate at the current step t. t Let β1 and β2 represent the first and second order momentum decay coefficients, respectively, and v represent the gradient of the current step. t This represents the second moment estimate for the current step t; where β1 and β2 are dynamically adjusted as follows: β1 = β 1,init ·(1-t / T); β2=β 2,init ·(1-t / T); where β 1,init and β 2,init Let represent the initial values ​​of the momentum decay coefficients of β1 and β2, respectively, and T be the total number of training steps; the gradient clipping algorithm is as follows: Where τ represents the gradient clipping threshold.

[0084] For example, in practical implementation, taking a deep learning model as the original model to be trained, the data features of the training set are used as the input of the model, denoted as X = {x1, x2, ..., xn}, where xi represents the feature vector of the i-th sample; the predicted target value (collision energy absorption performance) of the training set is used as the output of the model, denoted as Y = {y1, y2, ..., yn}, where yi represents the true energy absorption performance value of the i-th sample; the input X is calculated through each layer of the neural network to finally obtain the predicted output of the model. In the forward propagation of the neural network, the activation output a(l) = f of the l-th hidden layer (l) (W (l) z (l-1) +b (l) ), where z (l-1) It is the output of the previous layer, W (l) b is the weight matrix of the l-th layer. (l) It is the bias vector, f (l) It is the activation function; the loss function can use the mean squared error (MSE) to measure the model's predictions. The difference between the model and the true value Y; to prevent overfitting, a Dropout strategy is used, randomly dropping a portion of neurons during training to reduce the model's dependence on specific neurons z.(l) Dropout(a (l) , p), where p is the dropout probability; the K-fold cross-validation technique can be introduced to improve the generalization ability of the model, and the training set is divided into K parts (for example, K = 5), and each time one part is taken as the validation set, and the remaining K-1 parts are taken as the training set, and K times of model training and evaluation are performed, and finally the average value of K times of evaluation results is taken as the final performance index of the model to improve the generalization ability of the model; the AdamW algorithm is used to optimize the learning rate, and the core of the algorithm is where m t represents the first moment estimate of the current step t, g t represents the gradient of the current step, β1 and β2 represent the first and second momentum decay coefficients, respectively, v t represents the second moment estimate of the current step t; wherein β1 and β2 are dynamically adjusted as follows: β1 = β 1,init · (1-t / T); β2 = β 2,init · (1-t / T); wherein β 1,init and β 2,init represent the initial values of the momentum decay coefficients of β1 and β2, respectively, and T is the total number of training steps; wherein the gradient clipping algorithm is as follows: where τ represents the gradient clipping threshold; in addition, multiple models of different structures can be trained, or multiple models can be trained using different machine learning algorithms, and the prediction results of these models can be fused using weighted averaging or voting, etc. For example, for weighted averaging, the fused prediction value can be represented as: where m represents the number of models, represents the prediction value of the jth model, w j represents the weight of the jth model, and Σ j＝1 m w j = 1, and the weights can be allocated according to the performance of the models on the validation set; finally, the performance of the final model is evaluated using the test set, the real test data is fed back to the model to form a closed-loop optimization mechanism, and the model performance is continuously improved.

[0085] In the preferred embodiment, step S43 uses K-fold cross-validation, dropout, and AdamW optimization algorithm to train the neural network model, avoids the occurrence of overfitting phenomenon, and ensures that the model has good generalization ability and stability; dynamically adjusting the first and second momentum decay coefficients and applying the gradient clipping algorithm further improves the convergence speed and stability in the model training process.

[0086] Step S5: combine each energy-absorbing component and key structural feature parameter into several population individuals of the genetic algorithm, and iteratively optimize the population according to the preset genetic operation algorithm until a preset iteration number is reached or the population individuals meet a preset expected performance index, output the current energy-absorbing component combination and corresponding key structural feature parameters as the design result; and at each iteration optimization, process each population individual according to the design model to obtain a predicted performance index of each population individual, and calculate an fitness value according to a deviation of each predicted performance index from the expected performance index.

[0087] Preferably, in some embodiments of the present application, the fitness value is calculated according to a deviation of each predicted performance index from the expected performance index in step S5, wherein the calculation formula of the fitness value is as follows:

[0088] F(i) = -MSE(y(i), y target ); wherein F(i) represents the fitness value of population individual i, y(i) represents the predicted performance index of the design model, y target represents the expected performance index, and MSE represents the minimum mean square error.

[0089] For example, in specific implementation, first, the target to be optimized, i.e. the expected performance index (such as minimizing peak acceleration, maximizing energy absorption, etc.), is determined, and the combination of energy-absorbing components and key structural feature parameters are characterized as parameters for algorithm processing. Different component combinations can be encoded using binary encoding, integer encoding or Gray encoding, etc. For example, assuming that there are 4 components, each component has 3 optional schemes, a 12-bit binary code can be used to represent a component combination, and every 3 bits represent the selection of a component; the structural feature parameters of each component (such as wall thickness, cross-sectional size, etc.) are represented as continuous numerical variables; secondly, the genetic algorithm is set, and the combination encoding of energy-absorbing components and the structural feature parameters are characterized as population individuals of the genetic algorithm. The initial population size is set to 100, i.e. 100 sets of component combination and structural feature parameter encoding are randomly generated, each individual represents a potential design scheme, the maximum iteration number is set to 500, the crossover rate is set to 0.8, and the mutation rate is set to 0.02. The above settings can ensure the diversity and search efficiency of the population, while balancing the calculation cost and optimization accuracy; then, the performance of each individual is predicted using the pre-trained design model, the encoding and feature parameters of the individual are used as the input of the model, the output of the model is the predicted performance index value, and the prediction result is used to calculate the fitness value of the genetic algorithm. The fitness is calculated according to the deviation of the prediction result from the expected performance index, and is defined as F(i) = -MSE(y(i), y target); wherein F(i) represents the fitness value of the individual i in the population, y(i) represents the predicted performance indicator of the design model, y target represents the expected performance indicator, by minimizing the mean square error (MSE), the genetic algorithm is guided to optimize the population; then, the genetic algorithm iteratively optimizes the population through operations such as selection, crossover and mutation, constantly improves the fitness of individuals in the population, and repeatedly executes the fitness value calculation and genetic operation steps until the maximum iteration number is reached or the individuals in the population meet the expected performance indicator. When the individuals in the population meet the expected performance indicator, the corresponding energy-absorbing component combination code and structural feature parameters are output, thereby completing the optimization design of the vehicle body energy-absorbing component. In order to verify the effectiveness of the optimization result, the simulation verification can be performed on the optimized design scheme, and the deviation between the simulation result and the expected performance indicator is compared.

[0090] In the preferred embodiment, step S5 uses genetic algorithm and model in cooperation to automatically find the optimal solution based on the principle of evolution, without human intervention, further improving the design efficiency and ensuring the quality of the design result.

[0091] In summary, compared with the prior art, the above embodiments of the present application have the following beneficial effects: random combination generation ensures the diversity of the design scheme, avoiding the influence of the subjective experience of engineers; setting key structural feature parameters defines specific parameter ranges, providing clear targets and boundary conditions for subsequent simulation and analysis, reducing unnecessary waste of computing resources; simulation can predict the actual performance of different design schemes in advance, reducing the cost and time consumption of physical experiments; training the neural network model based on key structural feature parameters and simulation results, using AI technology to process a large amount of simulation data, quickly establishing the mapping relationship between input (structural features) and output (performance indicators), improving the speed and accuracy of data analysis; genetic algorithm and model work together, automatically finding the optimal solution based on the principle of evolution, without human intervention, further improving the design efficiency and ensuring the quality of the design result.

[0092] Embodiment two:

[0093] Please refer to Figure 2 The embodiment of the present application discloses a model-based collision energy-absorbing component design system, which comprises a component combination module M1, a feature parameter setting module M2, a simulation module M3, a model training module M4 and a processing module M5.

[0094] The component combination module M1 is used to randomly generate each energy-absorbing component combination that meets the preset combination rule according to the connection relationship of each energy-absorbing component of each reference vehicle model.

[0095] The component combination module M1 comprises an information acquisition unit, a data conversion unit and a combination unit.

[0096] The information acquisition unit is configured to acquire information of each energy absorption component of each reference vehicle model;

[0097] The data conversion unit is configured to convert the information of each energy absorption component into a numerical parameter according to a preset encoding rule;

[0098] The combination unit is configured to randomly generate each energy absorption component combination according to a connection relationship of each energy absorption component and a preset combination rule.

[0099] The component combination module M1 of the embodiment converts the energy absorption component information of each reference vehicle model into a numerical parameter and randomly generates a component combination based on the connection relationship and the combination rule, so that the energy absorption component design between different vehicle models can be quickly transplanted and compared, and the repetitive labor is reduced. In addition, the use of a unified encoding rule simplifies the subsequent processing flow, improves the speed and accuracy of parameterized modeling, and thus speeds up the entire design cycle and reduces the development cost.

[0100] The feature parameter setting module M2 is configured to set a plurality of key structure feature parameters for each energy absorption component in each energy absorption component combination according to preset inner and outer boundary constraint parameters of the energy absorption component spatial arrangement, to obtain a first feature parameter combination corresponding to each energy absorption component combination.

[0101] The feature parameter setting module M2 includes a parameter setting unit, an importance value calculation unit, and a parameter screening unit.

[0102] The parameter setting unit is configured to set a plurality of structure feature parameters for each energy absorption component according to preset inner and outer boundary constraint parameters of the energy absorption component spatial arrangement;

[0103] The importance value calculation unit is configured to analyze and calculate the importance value of each structure feature parameter on a preset safety performance index according to a preset SGMW sensitivity analysis specification;

[0104] The parameter screening unit is configured to sort the importance values from large to small, and select the structure feature parameters corresponding to a plurality of importance values in the front as key structure feature parameters.

[0105] The feature parameter setting module M2 of the embodiment sets structure feature parameters according to the inner and outer boundary constraint parameters, and optimizes key structure feature parameters according to the SGMW sensitivity analysis specification, to ensure the importance of the selected parameters on the safety performance index. The sorting and selection mechanism ensures that the most important parameters are given priority, thereby improving the safety and reliability of the design scheme. At the same time, it helps to reduce unnecessary complexity and makes the design more focused on the factors that have the greatest impact on the overall performance.

[0106] The simulation module M3 is configured to generate each energy-absorbing component finite element analysis model according to a preset parametric model generation algorithm and a mesh generation algorithm, and simulate the preset non-energy-absorbing component finite element analysis model and each energy-absorbing component finite element analysis model according to a preset simulation algorithm to obtain each performance simulation result.

[0107] The simulation module M3 comprises a sampling unit and a finite element model generation unit.

[0108] The sampling unit is configured to extract the first characteristic parameter combination according to a preset sampling algorithm to obtain a plurality of sampled first characteristic parameter combinations.

[0109] The finite element model generation unit is configured to generate each energy-absorbing component finite element analysis model according to a preset parametric model generation algorithm and a mesh generation algorithm.

[0110] The simulation module M3 of the embodiment can solve the computing resource bottleneck problem that may occur in full data analysis by sampling the first characteristic parameter combination and using the parametric model generation algorithm and the mesh generation algorithm to construct the finite element analysis model. The sampling method not only retains the core characteristics of the original data but also reduces the computing burden, making large-scale simulation possible and promoting more extensive design exploration and optimization.

[0111] The model training module M4 is configured to train a preset original neural network model according to the data of each performance simulation result, first characteristic parameter combination and energy-absorbing component combination to obtain a design model.

[0112] The model training module M4 comprises a data set construction unit and a preprocessing unit.

[0113] The data set construction unit is configured to construct an original mapping data set according to each performance simulation result and the corresponding first characteristic parameter combination and the corresponding energy-absorbing component combination.

[0114] The preprocessing unit is configured to process the original mapping data set according to a preset preprocessing algorithm to obtain a mapping data set. The preprocessing algorithm comprises any one or a combination of more than one of the following: missing value processing, abnormal value processing, duplicate value processing, renaming processing, format unification processing, normalization processing, and feature encoding processing.

[0115] The model training module M4 of the embodiment can significantly improve the quality of data, eliminate noise interference, and enhance the effectiveness of model learning by applying preprocessing algorithms (such as missing value processing and abnormal value processing) on the basis of constructing the original mapping data set.

[0116] The model training module M4 further comprises a training unit;

[0117] The training unit is configured to train the original neural network model according to a preset K-fold cross-validation algorithm, a dropout algorithm and the mapping data set to obtain the design model.

[0118] The training uses a preset AdamW algorithm to optimize a learning rate, and the algorithm is as follows:

[0119] m t =β1m t-1 +(1-β1)g t ; wherein m t represents a first moment estimation of a current step t, g t represents a gradient of the current step, β1 and β2 respectively represent first and second momentum decay coefficients, and v t represents a second moment estimation of the current step t; wherein β1 and β2 are dynamically adjusted as follows: β1=β 1,init ·(1-t / T);β2=β 2,init ·(1-t / T); wherein β 1,init and β 2,init respectively represent initial values of the momentum decay coefficients of β1 and β2, and T is a total training step number; wherein a gradient clipping algorithm is as follows: wherein τ represents a gradient clipping threshold.

[0120] The model training module M4 of the embodiment uses K-fold cross-validation, dropout and AdamW optimization algorithm to train the neural network model, avoids overfitting phenomenon and ensures that the model has good generalization ability and stability; the first and second momentum decay coefficients are dynamically adjusted and the gradient clipping algorithm is applied, which further improves the convergence speed and stability in the model training process.

[0121] The processing module M5 is configured to combine and represent key structural feature parameters of each energy absorption component as a plurality of population individuals of a genetic algorithm, and iteratively optimize the population according to a preset genetic operation algorithm until a preset iteration number is reached or the population individuals meet a preset expected performance index, and output the current energy absorption component combination and corresponding key structural feature parameters as a design result; and at each iteration optimization, process the population individuals according to the design model to obtain predicted performance indexes of the population individuals, and calculate fitness values according to deviations of the predicted performance indexes from the expected performance index.

[0122] The processing module M5 comprises a fitness value calculation unit.

[0123] The fitness value calculation unit is configured to calculate a fitness value according to the following formula:

[0124] F(i) = -MSE(y(i), y target ); wherein F(i) represents the fitness value of the population individual i, y(i) represents the prediction performance index of the design model, y target represents the expected performance index, and MSE represents the minimized mean square error.

[0125] The processing module M5 of the embodiment automatically finds the optimal solution by the genetic algorithm and the model in cooperation, in combination with the principle of evolution, without manual intervention, further improves the design efficiency, and ensures the quality of the design result.

[0126] In summary, compared with the prior art, the above embodiments have the following beneficial effects: random generation of combinations ensures the diversity of design schemes and avoids the influence of subjective experience of engineers; setting key structural feature parameters defines specific parameter ranges, so that subsequent simulation and analysis have clear targets and boundary conditions, reducing unnecessary waste of computing resources; simulation can predict the actual performance of different design schemes in advance, reducing the cost and time consumption of physical experiments; training a neural network model in combination with key structural feature parameters and simulation results, using AI technology to process a large amount of simulation data, quickly establishing the mapping relationship between input (structural features) and output (performance indicators), improving the speed and accuracy of data analysis; genetic algorithm and model cooperate, combined with the principle of evolution to automatically find the optimal solution, without manual intervention, further improving the design efficiency, and ensuring the quality of the design result.

[0127] The division of each module described above is only logical functional division, and there can be another division way in actual implementation, for example, multiple modules can be combined or integrated into another system.

[0128] Embodiment three:

[0129] Figure 3 A structural diagram of the model-based collision energy-absorbing component design device is shown. As shown in the figure, the model-based collision energy-absorbing component design device can include a processor N1, a memory N2, a data interface N3, and a communication bus N4. Figure 3

[0130] The processor N1, the memory N2, and the data interface N3 complete mutual communication through the communication bus N4; the data interface N3 is used for data communication with other devices such as input devices or output devices; the processor N1 is used for executing a program N5, and can specifically execute related steps in the above model-based collision energy-absorbing component design method embodiment.

[0131] ​In particular, program N5 can include program code comprising computer-executable instructions.

[0132] Processor N1 can be a central processing unit, CPU, or an application specific integrated circuit, ASIC, or one or more integrated circuits configured to perform the functions of the embodiments. The one or more processors of the model-based crash energy management component design apparatus can be of the same type or different types. For example, one or more CPUs and one or more ASICs.

[0133] Memory N2 is configured to store program N5. Memory N2 can include a volatile RAM memory and can also include a non-volatile memory, such as at least one disk memory.

[0134] The algorithms and displays presented herein are not inherently related to any particular computer, virtual system, or other apparatus. Furthermore, embodiments are not described with reference to any particular programming language.

[0135] Embodiment four:

[0136] The embodiments also provide a computer readable storage medium storing at least one executable instruction, which, when executed on a model-based crash energy management component design apparatus / system, causes the model-based crash energy management component design apparatus / system to perform the model-based crash energy management component design method of any of the method embodiments.

[0137] In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the application can be practiced without these specific details. In other instances, well-known methods, structures and techniques have not been described in detail in order to not obscure importantly-meritorious aspects of the embodiments of the application. Like numbers refer to like elements throughout the description. Reference

[0138] As will be appreciated by one skilled in the art, aspects of the embodiments can be embodied as a system, method or computer program product. Accordingly, aspects of the embodiments can take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, micro-code, etc.) or an embodiment combining software and hardware aspects that can all generally be referred to herein as a "circuit," "module" or "system." Furthermore, aspects of the embodiments can take the form of a program of instructions (i.e., computer-readable storage medium) embodied in a computer readable medium (e.g., a non-transitory computer readable medium such as magnetic or optical disks or memory chips) or transmitted over a modulated carrier signal.

Claims

1. A model-based crash energy absorption component design method, characterized by, The method comprises the following steps: According to the connection relationship of each energy-absorbing component of each reference vehicle, randomly generate each energy-absorbing component combination that meets the preset combination rule; According to the preset internal and external boundary constraint parameters of the energy-absorbing component space arrangement, set a plurality of key structure characteristic parameters for each energy-absorbing component in each energy-absorbing component combination, and obtain a first characteristic parameter combination corresponding to each energy-absorbing component combination; According to the preset parameterization model generation algorithm and the grid generation algorithm, process each first characteristic parameter combination to obtain an energy-absorbing component finite element analysis model, and according to a preset simulation algorithm, simulate a preset non-energy-absorbing component finite element analysis model and each energy-absorbing component finite element analysis model to obtain each performance simulation result; According to the data of each performance simulation result, first characteristic parameter combination and energy-absorbing component combination, train a preset original neural network model to obtain a design model; Characterize each energy-absorbing component combination and key structure characteristic parameter as a plurality of population individuals of a genetic algorithm, and iteratively optimize the population according to a preset genetic operation algorithm until a preset iteration number is reached or the population individuals meet a preset expected performance index, output the current energy-absorbing component combination and corresponding key structure characteristic parameters as a design result; And at each iteration optimization, process each population individual according to the design model to obtain a predicted performance index of each population individual, and calculate an adaptability value according to the deviation between each predicted performance index and the expected performance index; According to the data of each performance simulation result, first characteristic parameter combination and energy-absorbing component combination, train a preset original neural network model to obtain a design model, comprising: According to the preset K-fold cross-validation algorithm, dropout algorithm and mapping data set, train the original neural network model to obtain the design model; wherein the mapping data set is obtained according to each performance simulation result and corresponding first characteristic parameter combination and corresponding energy-absorbing component combination.

2. A model-based crash energy management component design method according to claim 1, wherein, According to the connection relationship of each energy-absorbing component of each reference vehicle, randomly generate each energy-absorbing component combination that meets the preset combination rule, comprising: Obtain information of each energy-absorbing component of each reference vehicle; According to a preset encoding rule, convert the information of each energy-absorbing component into a numerical parameter; According to the connection relationship of each energy-absorbing component and the preset combination rule, randomly generate each energy-absorbing component combination.

3. A model-based crash energy management component design method according to claim 2, wherein, According to the preset internal and external boundary constraint parameters of the energy-absorbing component space arrangement, set a plurality of key structure characteristic parameters for each energy-absorbing component in each energy-absorbing component combination, and obtain a first characteristic parameter combination corresponding to each energy-absorbing component combination, comprising: According to the preset internal and external boundary constraint parameters of the energy-absorbing component space arrangement, set a plurality of structure characteristic parameters for each energy-absorbing component; According to the preset SGMW sensitivity analysis specification, analyze and calculate the importance value of each structure characteristic parameter to the preset safety performance index; Sort each importance value from large to small, and select the structure characteristic parameters corresponding to a plurality of importance values ranked in the front as key structure characteristic parameters.

4. A model-based crash energy management component design method according to claim 3, wherein, The generating algorithm of the preset parameterized model and the mesh generating algorithm process each first characteristic parameter combination to obtain a finite element analysis model of each energy absorption component, including: According to a preset sampling algorithm, the first characteristic parameter combination is extracted to obtain a plurality of sampled first characteristic parameter combinations; According to a preset parameterized model generating algorithm and a mesh generating algorithm, each of the sampled first characteristic parameter combinations is processed to obtain a finite element analysis model of each energy absorption component.

5. A model-based crash energy management component design method according to claim 4, wherein, The data of each performance simulation result, first characteristic parameter combination and energy absorption component combination is used to train a preset original neural network model to obtain a design model, including: According to each performance simulation result and the corresponding first characteristic parameter combination and the corresponding energy absorption component combination, an original mapping data set is constructed; According to a preset preprocessing algorithm, the original mapping data set is processed to obtain a mapping data set; wherein the preprocessing algorithm includes any one or a combination of the following: missing value processing, outlier processing, duplicate value processing, renaming processing, format unification processing, normalization processing, feature encoding processing.

6. A model-based crash energy management component design method as in claim 5, wherein, The data of each performance simulation result, first characteristic parameter combination and energy absorption component combination is used to train a preset original neural network model to obtain a design model, including: The training uses a preset AdamW algorithm to optimize the learning rate, and the algorithm is as follows: The fitness value is calculated according to the deviation between each predicted performance index and the expected performance index, including: where m t denotes the first moment estimate of the current step t, g t denotes the gradient of the current step t, t-1 denotes the previous step of the current step t, β1 and β2 denote the momentum decay coefficients of the first and second order, respectively, v t denotes the second moment estimate of the current step t; where β1 and β2 are dynamically adjusted as follows: β1 = β 1,init · (1 - t / T); β2 = β 2,init · (1 - t / T); where β 1,init and β 2,init denote the initial values of the momentum decay coefficients of β1 and β2, respectively, T is the total number of training steps; where the clipping algorithm of the gradient is as follows: where τ denotes the clipping threshold of the gradient.

7. A model-based crash energy management component design method according to claim 6, wherein, The calculation formula of the fitness value is as follows: including: F(i) = -MSE(y(i), y target ); where F(i) denotes the fitness value of population individual i, y(i) denotes the predicted performance indicator of the design model, y target denotes the expected performance indicator, and MSE denotes the minimized mean square error.

8. A model-based crash energy management component design system, characterized by, The component combination module, the characteristic parameter setting module, the simulation module, the model training module and the processing module; The component combination module is used to randomly generate each energy absorption component combination according to the connection relationship of each energy absorption component of each reference vehicle, which meets the preset combination rule; The characteristic parameter setting module is used to set a plurality of key structural characteristic parameters for each energy absorption component in each energy absorption component combination according to the internal and external boundary constraint parameters of the preset energy absorption component space arrangement to obtain the corresponding first characteristic parameter combination of each energy absorption component combination; The simulation module is used to process each first characteristic parameter combination according to a preset parameterized model generating algorithm and a mesh generating algorithm to obtain a finite element analysis model of each energy absorption component, and to simulate a preset non-energy absorption component finite element analysis model and each energy absorption component finite element analysis model according to a preset simulation algorithm to obtain each performance simulation result; The model training module is used to train a preset original neural network model according to the data of each performance simulation result, first characteristic parameter combination and energy absorption component combination to obtain a design model; ​ The processing module is configured to combine each energy-absorbing component and a key structural feature parameter into a plurality of population individuals of a genetic algorithm, and iteratively optimize the population according to a preset genetic operation algorithm until a preset iteration number is reached or the population individuals satisfy a preset expected performance index, and output a current energy-absorbing component combination and corresponding key structural feature parameters as a design result; and at each iteration optimization, process the population individuals according to the design model to obtain a predicted performance index of each population individual, and calculate a fitness value according to a deviation between each predicted performance index and the expected performance index. The model training module comprises a training unit. The training unit is configured to train an original neural network model according to a preset K-fold cross-validation algorithm, a dropout algorithm and a mapping data set to obtain the design model, wherein the mapping data set is obtained according to each performance simulation result, a corresponding first feature parameter combination and a corresponding energy-absorbing component combination.

9. A model-based crash energy management component design apparatus comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the computer program comprises instructions for: The computer program, when loaded into the processor, implements the steps of the model-based collision energy-absorbing component design method according to any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program, when executed by the processor, implements the steps of the model-based collision energy-absorbing component design method according to any one of claims 1-7.

Images