Optimization design method of carbon fiber composite material

Abstract

The invention provides an optimization design method of a carbon fiber composite material. The optimization design method comprises the following steps: acquiring a crashworthiness requirement, an energy absorption requirement and a lightweight requirement of the carbon fiber composite material; constructing a target performance model of the carbon fiber composite material, and predicting performance parameters of the carbon fiber composite material under a high-speed impact working condition and a low-speed impact working condition according to the material parameters; carrying out iterative optimization on material parameters in the target performance model by adopting an NSGA-II multi-target genetic algorithm to generate a set of optimal solutions; and carrying out quantitative screening on a set of optimal solutions by adopting a grey correlation analysis method based on double-weight fusion, outputting a comprehensive optimal carbon fiber composite material structure scheme, and realizing optimal crashworthiness requirement, energy absorption requirement and lightweight requirement of the material. According to the method, the NSGA-II multi-target genetic algorithm and the grey relational degree analysis method based on double-weight fusion are adopted to realize material proportioning under the high-speed impact working condition and the low-speed impact working condition, so that the material performance is optimal.

Description

Technical Field

[0001] This invention belongs to the field of composite material design technology, and specifically relates to an optimization design method for carbon fiber composite materials. Background Technology

[0002] Carbon fiber composites, due to their excellent specific strength, specific stiffness, and outstanding energy absorption characteristics, have been widely used in structural design in aerospace, automotive, and rail transportation industries. Particularly in the field of collision safety, carbon fiber composite components can absorb impact energy through multiple damage modes, effectively protecting occupant safety.

[0003] However, the mechanical behavior of carbon fiber composites exhibits high anisotropy and process sensitivity, and their impact response is influenced by a combination of factors, including material composition, fiber orientation, and layup structure. In practical engineering applications, structural components often need to withstand different conditions simultaneously, such as low-velocity and high-velocity impacts. These two types of conditions have fundamentally different requirements for material properties: low-velocity impacts emphasize damage tolerance and structural integrity maintenance, while high-velocity impacts focus more on energy absorption efficiency and peak load control.

[0004] Existing optimization design methods for carbon fiber composites suffer from the following shortcomings when addressing the aforementioned problems: Traditional optimization methods often focus on single operating conditions, making it difficult to consider the comprehensive performance requirements under multiple operating conditions. When the design objective expands from a single operating condition to two operating conditions, the relationship between material parameters and performance response becomes more complex, and traditional single-objective optimization or simple weighted multi-objective optimization methods struggle to effectively handle performance conflicts between operating conditions. Most studies optimize material parameters and structural parameters separately, neglecting the strong coupling effect between them, making it difficult to achieve the global optimum in the optimization results. Summary of the Invention

[0005] This invention proposes an optimized design method for carbon fiber composite materials, which solves the problem that existing technologies cannot simultaneously meet the comprehensive performance requirements under various operating conditions, including high speed and low speed.

[0006] To address the aforementioned technical problems, this invention provides an optimized design method for carbon fiber composite materials, comprising the following steps:

[0007] Step S1: Obtain the impact resistance requirements, energy absorption requirements, and lightweight requirements for carbon fiber composite materials;

[0008] Step S2: Construct a target performance model for carbon fiber composite materials and predict the performance parameters of carbon fiber composite materials under high-speed impact and low-speed impact conditions based on material parameters.

[0009] Step S3: Use the NSGA-II multi-objective genetic algorithm to iteratively optimize the material parameters in the target performance model to generate a set of optimal solutions;

[0010] Step S4: The set of optimal solutions is quantitatively screened using a grey relational analysis method based on dual-weight fusion, and the comprehensive optimal carbon fiber composite material structure scheme is output to achieve the optimal requirements for material impact resistance, energy absorption and lightweighting.

[0011] Preferably, the carbon fiber composite material in step S1 is a material system with continuous carbon fiber as reinforcement, polyarylene ether nitrile as resin matrix, and phthalocyanine iron as carbon fiber surface microstructure building block.

[0012] Preferably, the impact resistance requirement in step S1 is to minimize the peak impact force; the energy absorption requirement is to maximize the energy absorption; and the lightweight requirement is to minimize the component mass.

[0013] Preferably, the target performance model in step S2 is a radial basis function surrogate model; the input of the radial basis function surrogate model is material parameters, which include material variables and structural variables. The material variables include at least the fiber volume fraction and the protrusion height of the phthalocyanine iron microstructure, and the structural variables include at least the layup angle, layup sequence, layup ratio, wall thickness, and tube length.

[0014] The output of the radial basis function surrogate model is a performance parameter, which includes at least the peak impact force under low-speed impact conditions, the maximum impact displacement under low-speed impact conditions, the energy absorbed under low-speed impact conditions, the peak impact force under high-speed impact conditions, the maximum impact displacement under high-speed impact conditions, the energy absorbed under high-speed impact conditions, and the component mass.

[0015] Preferably, the construction of the target performance model of the carbon fiber composite material in step S2 includes at least the following steps:

[0016] Step S21: Construct the complete stiffness matrix of the laminate based on the classical laminate theory, and establish a mapping model between design variables and initial mechanical properties;

[0017] Step S22: Establish an in-plane damage model and an interlaminar damage model based on the theory of continuous medium damage mechanics. The damage variables of the in-plane damage model include at least fiber-direction tensile-compression damage variables, transverse tensile-compression damage variables, and in-plane shear damage variables. The interlaminar damage model introduces interlaminar damage variables based on the cohesive region model.

[0018] Step S23: Establish an impact simulation model through finite element simulation, apply high-speed impact load and low-speed impact load to carbon fiber composite material, calculate and output the peak impact force, impact energy absorption, impact displacement and component mass under the dual conditions as the training sample benchmark values ​​of the radial basis function surrogate model.

[0019] Step S24: Based on the benchmark values ​​of the training samples, the shape parameters of the radial basis functions are determined using leave-one-out cross-validation, and the radial basis function surrogate model is constructed.

[0020] Preferably, the objective function of the NSGA-II multi-objective genetic algorithm in step S3 includes at least a weighted minimum impact force peak objective function, a weighted maximum energy absorption objective function, and a minimum mass objective function;

[0021] The objective function for the weighted minimum impact force peak value is:

[0022] ;

[0023] In the formula, The objective function is the weighted minimum impact force peak value. and These represent the peak impact forces under low-speed and high-speed impact conditions, respectively. This represents the theoretical minimum value of the peak impact force under dual operating conditions. This is the maximum peak impact force setting value; For low-speed impact conditions, the weighting coefficient is used. This represents the weighting coefficient for high-speed impact conditions.

[0024] The weighted maximum energy absorption objective function is:

[0025] ;

[0026] In the formula, and Energy absorption under low-speed impact conditions and high-speed impact conditions, respectively; This represents the theoretical maximum energy absorption value under dual operating conditions. The minimum allowable energy absorbed after weighting for both operating conditions;

[0027] The minimum mass objective function is:

[0028] ;

[0029] In the formula, For component quality; This represents the theoretical minimum mass of the component. Set the maximum component mass value.

[0030] Preferably, the step S3, which involves iteratively optimizing the material parameters in the target performance model using the NSGA-II multi-objective genetic algorithm, includes the following steps:

[0031] Step S31: Based on the sensitivity of material parameters to high-speed and low-speed impact conditions, divide the parameter value range into high-speed adaptation range, low-speed adaptation range, and trade-off range.

[0032] Step S32: Use the Latin hypercube sampling method to generate an initial design scheme within the range of material parameter values, with sampling covering the high-speed adaptation range, low-speed adaptation range, and trade-off range.

[0033] Step S33: Calculate the dominance relationship of each design scheme under the weighted minimum impact force peak objective function, the weighted maximum energy absorption objective function, and the minimum mass objective function, and classify the population according to the dominance level;

[0034] Step S34: Calculate the distance between each non-dominated solution and its neighboring solutions in the target space;

[0035] Step S35: Use roulette wheel selection and elite retention strategies for selection, use an improved partial matching crossover strategy to adapt the crossover operation of the ply sequence encoding, and use a position swap mutation strategy to adapt the mutation operation of the ply sequence encoding.

[0036] Step S36: Select a new generation of population according to the principles of priority based on dominance level and priority based on high crowding at the same level;

[0037] Step S37: Repeat steps S33 to S36 until the number of iterations reaches the maximum number of iterations or the convergence condition is met, and output the set of optimal solutions.

[0038] Preferably, the NSGA-II multi-objective genetic algorithm in step S3 introduces a dual-condition conflict coordination mechanism, which includes the following steps:

[0039] Step A: Define conflict quantification metrics This characterizes the degree of contradiction in performance parameters under dual operating conditions:

[0040] ;

[0041] In the formula, These are the performance parameters under high-speed impact conditions. These are the corresponding parameters under low-speed impact conditions; The average values ​​of parameters under both operating conditions;

[0042] Step B: Constructing the comprehensive difference coefficient Quantifying the performance differences between the two operating conditions:

[0043] ;

[0044] In the formula, , These are the target weights for crashworthiness and energy absorption, respectively. A dual-condition collision quantification index for crashworthiness targets; A quantitative indicator for dual-condition conflict of energy-absorbing targets;

[0045] Step C: Introduce conflict adaptation factor Schemes with small performance differences between the two operating conditions are assigned ranking weights:

[0046] ;

[0047] In the formula, The maximum value of the comprehensive difference coefficient among all solutions in the solution set;

[0048] Step D: Quantify conflict indicators The corresponding performance parameters increase the weight of congestion calculation.

[0049] Preferably, the step S4, which employs a grey relational analysis method based on dual-weight fusion to quantitatively filter the set of optimal solutions, includes the following steps:

[0050] Step S41: Extract core evaluation indicators, which include at least the peak impact force under low-speed impact conditions, the maximum impact displacement under low-speed impact conditions, the energy absorbed under low-speed impact conditions, the peak impact force under high-speed impact conditions, the maximum impact displacement under high-speed impact conditions, the energy absorbed under high-speed impact conditions, and the component mass. Normalize the core evaluation indicators according to the target type.

[0051] Step S42: Set the target weights for crashworthiness requirements Target weights for energy absorption requirements and the target weights for lightweight requirements Combined with operating condition weights and The target weight is decomposed into the weights of each core evaluation indicator;

[0052] Step S43: Take the maximum value of each normalized core evaluation index among all schemes as the ideal value, and construct the ideal optimal sequence;

[0053] Step S44: Quantify the degree of closeness of each scheme to the ideal optimal sequence for each indicator to obtain the correlation coefficient. Combine the indicator weights to perform a weighted summation of the correlation coefficients to obtain the comprehensive correlation degree. The expression for the correlation coefficient is:

[0054] ;

[0055] In the formula, Let be the correlation coefficient between the i-th scheme and the j-th index in the ideal optimal sequence; Let j be the index value of the j-th item in the ideal optimal sequence; Let be the normalized value of the j-th indicator of the i-th scheme; The resolution coefficient;

[0056] The expression for the comprehensive correlation degree is:

[0057] ;

[0058] In the formula, Let be the overall correlation degree of the i-th group of schemes; Let be the weight of the j-th indicator;

[0059] Step S45: Sort all schemes in the set of optimal solutions in descending order according to the comprehensive correlation degree, select the scheme with the highest comprehensive correlation degree as the comprehensive optimal scheme and output it.

[0060] Preferably, the step S42 of decomposing the target weight into the weights of each core evaluation indicator includes the following steps: setting internal weight allocation coefficients. , used to split the crashworthiness target weights, where As the internal weight of the peak impact force, 1- The internal weight of the impact displacement;

[0061] Weight of peak impact force under low-velocity impact conditions for:

[0062] ;

[0063] Weight of maximum impact displacement under low-speed impact conditions for:

[0064] ;

[0065] Weight of energy absorption under low-speed impact conditions for:

[0066] ;

[0067] Weight of peak impact force under high-speed impact conditions for:

[0068] ;

[0069] Weight of maximum impact displacement under high-speed impact conditions for:

[0070] ;

[0071] Weight of energy absorption under high-speed impact conditions for:

[0072] ;

[0073] Weight of component quality for:

[0074] .

[0075] The beneficial effects of the present invention include at least the following:

[0076] 1. It simultaneously considers the dual requirements of high-speed impact and low-speed impact conditions. By introducing condition weights and a dual-condition conflict coordination mechanism, it effectively solves the conflict between performance objectives under multiple conditions and realizes the comprehensive optimal design of carbon fiber composite materials under different impact conditions.

[0077] 2. By incorporating material variables such as fiber volume fraction and phthalocyanine iron microstructure protrusion height with structural variables such as layup angle, layup sequence, layup ratio, wall thickness, and tube length into a unified optimization framework, the synergistic optimization of material-structure parameters is achieved, giving full play to the designability advantages of carbon fiber composite materials.

[0078] 3. The proposed grey relational analysis method based on dual-weight fusion organically combines the target weight and the working condition weight to form a scientific and systematic decision-making mechanism, which can quickly determine the comprehensive optimal solution that meets the actual needs of the project from a large number of Pareto optimal solutions.

[0079] 4. The radial basis function surrogate model is used to replace the time-consuming finite element simulation for performance prediction, which greatly improves the optimization efficiency; the model parameters are determined by leave-one-out cross-validation, which ensures the prediction accuracy and generalization ability of the surrogate model. Attached Figure Description

[0080] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention. Detailed Implementation

[0081] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0082] like Figure 1As shown, this embodiment of the invention provides an optimized design method for carbon fiber composite materials, including the following steps:

[0083] Step S1: Obtain the impact resistance requirements, energy absorption requirements, and lightweight requirements for carbon fiber composite materials.

[0084] The carbon fiber composite material in this embodiment of the invention is a material system with continuous carbon fibers as reinforcement, polyarylene ether nitrile (PEN) resin as resin matrix, and phthalocyanine iron as the microstructure building block on the carbon fiber surface. Polyarylene ether nitrile (PEN) resin possesses excellent heat resistance, chemical corrosion resistance, and mechanical properties, making it suitable as a matrix material for high-performance composite materials. The phthalocyanine iron microstructure can effectively enhance the interfacial bonding strength between the carbon fibers and the resin matrix, improving the overall mechanical properties and damage tolerance of the composite material.

[0085] Performance requirements include impact resistance, energy absorption, and lightweighting: impact resistance requires minimizing peak impact force; energy absorption requires maximizing energy absorption; and lightweighting requires minimizing component mass.

[0086] The crashworthiness requirements include two core indicators: peak impact force and impact displacement. The weights of these two indicators are unified based on the coupling calculation of crashworthiness target weight and working condition weight, ensuring that the optimization priority of internal indicators of crashworthiness target is consistent and avoiding performance imbalance caused by weight splitting.

[0087] The parameters of the material system are initialized, including the range of values ​​for material variables and structural variables, as well as the intrinsic property parameters of the material.

[0088] Material variables include fiber volume fraction The protrusion height of the phthalocyanine iron microstructure Structural variables include ply angles. Layer sequence and layer ratio Wall thickness Manager .

[0089] Wherein, the layering order is a discrete ordered variable, defined as follows:

[0090] Angle type encoding: Maps the ply angle type to an integer encoding, according to the following rules. , , , ;

[0091] Ply sequence coding: The ply sequence is represented by an integer sequence, the sequence length is equal to the total number of ply n, the sequence order directly corresponds to the actual ply sequence, n is determined by the ply ratio and wall thickness, and is an even number to meet the requirements of symmetrical ply process.

[0092] The intrinsic properties of the materials include the elastic modulus, shear modulus, and Poisson's ratio of continuous carbon fibers, polyaryletheronitrile, and iron phthalocyanine.

[0093] Step S2: Construct a target performance model for carbon fiber composite materials and predict the performance parameters of carbon fiber composite materials under high-speed impact and low-speed impact conditions based on material parameters.

[0094] The target performance model in this embodiment of the invention is a radial basis function (RBF) surrogate model. The RBF surrogate model is a mathematical model based on interpolation theory, capable of constructing a continuous mapping relationship between input variables and output response based on known sample points. Compared to polynomial response surface models, the RBF model has better fitting ability for complex responses with high nonlinearity; compared to the Kriging model, the RBF model is more computationally efficient and more suitable for solving large-scale optimization problems.

[0095] The input to the radial basis function surrogate model is material parameters, which include material variables and structural variables.

[0096] The output of the radial basis function surrogate model consists of performance parameters, which include at least the peak impact force under low-speed impact conditions, the maximum impact displacement under low-speed impact conditions, the energy absorbed under low-speed impact conditions, the peak impact force under high-speed impact conditions, the maximum impact displacement under high-speed impact conditions, the energy absorbed under high-speed impact conditions, and the component mass. By predicting these seven performance parameters, the overall performance of the design scheme under dual impact conditions can be comprehensively evaluated.

[0097] Constructing a target performance model for carbon fiber composites includes at least the following steps:

[0098] Step S21: Construct a complete stiffness matrix for the laminated plate based on classical laminated plate theory, covering the mechanical properties under multi-directional stress states. The specific steps are as follows:

[0099] Step S211: Define a local coordinate system with the fiber direction as axis 1, the transverse direction as axis 2, and the thickness direction as axis 3.

[0100] The symmetric stiffness matrix of the k-th layer monolayer in the local coordinate system It is a 3×3 matrix:

[0101] ;

[0102] The formulas for calculating each stiffness component are as follows:

[0103] ;

[0104] ;

[0105] ;

[0106] ;

[0107] ;

[0108] In the formula, Let k be the elastic modulus of the k-th layer monolayer along the fiber direction; It is the transverse elastic modulus; Poisson's ratio in the plane; It is the inverse Poisson's ratio in the plane; This is the in-plane shear modulus.

[0109] Step S212: Define the global coordinate system: the direction of the impact load is the X-axis, the perpendicular direction is the Y-axis, and the ply angle is... The angle between the local X-axis and the global X-axis.

[0110] Local coordinate system Transform to global coordinate system, transformation matrix From the ply angle Sure:

[0111] ;

[0112] Stiffness matrix of a single-layer plate in the global coordinate system for:

[0113] .

[0114] Step S213: Integration of global stiffness matrix of laminated plate

[0115] The expression for the total thickness h of the laminate is:

[0116] .

[0117] Based on the thickness of each layer Calculate the extended stiffness matrix of the laminate. , , These correspond to in-plane tension / shear, coupling effect, and bending effect, respectively:

[0118] ;

[0119] ;

[0120] ;

[0121] in, Let be the coordinates of the upper surface of the k-th layer in the global thickness direction; Let be the coordinates of the lower surface in the global thickness direction.

[0122] Step S214: Output equivalent stiffness parameters

[0123] When the laminate is symmetrically plyed, the coupling matrix The final output is the complete equivalent stiffness matrix of the laminate in the global coordinate system. It covers multi-directional stiffness characteristics such as in-plane tension, transverse tension, and in-plane shear, providing a basis for stress-strain analysis under impact loads.

[0124] Step S22: Establish in-plane damage model and interlayer damage model by combining the theory of continuous medium damage mechanics.

[0125] Step S221: Define the in-plane damage model. The damage variables of the in-plane damage model include at least fiber-direction tensile-compression damage variables, transverse tensile-compression damage variables, and in-plane shear damage variables.

[0126] Define damage variables: tensile / compressive damage variables in the fiber direction Lateral tensile / compression damage variables In-plane shear damage variables .

[0127] Introducing damage variables , , Global stiffness matrix after damage for:

[0128] .

[0129] Damage variables are determined by the corresponding stress thresholds and damage evolution equations. Based on continuum damage mechanics and the Hashin criterion, combined with intrinsic material properties and corresponding physical experiments, the stress initiation thresholds for each damage mode are determined. An exponential evolution model is used to describe the gradual process of damage from initiation to complete failure.

[0130] Fiber orientation damage evolution Including tensile damage and compressive damage:

[0131] Tensile damage :

[0132] ;

[0133] In the formula, The equivalent plastic strain under tension along the fiber direction is calculated in real time by finite element simulation. The tensile fracture strain along the fiber direction is an intrinsic parameter of the material, obtained from physical experiments. p are evolution coefficients, determined by fitting the stress-strain curves from finite element simulations;

[0134] Compression damage :

[0135] ;

[0136] In the formula, The equivalent plastic strain under fiber-direction compression is calculated in real time by finite element simulation. The fiber-direction compressive fracture strain is an intrinsic parameter of the material, obtained from physical experiments. The evolution coefficient is determined by fitting the stress-strain curve from the finite element simulation.

[0137] Total damage in the fiber direction:

[0138] .

[0139] Transverse damage evolution Including tensile damage and compressive damage:

[0140] Tensile damage :

[0141] ;

[0142] In the formula, The accumulated fracture energy under transverse tensile stress is calculated in real time by simulation. The transverse tensile fracture energy is an intrinsic parameter of the material, obtained from physical experiments. The evolution coefficient is determined by fitting the stress-strain curve of the finite element simulation.

[0143] Compression damage :

[0144] ;

[0145] In the formula, The accumulated fracture energy under transverse compression is calculated in real time by simulation. The transverse compressive fracture energy is an intrinsic parameter of the material, obtained from physical experiments. The evolution coefficient is determined by fitting the stress-strain curve of the finite element simulation.

[0146] Total transverse damage:

[0147] .

[0148] In-plane shear damage evolution :

[0149] ;

[0150] In the formula, The in-plane shear equivalent plastic strain is calculated in real time by finite element simulation; The in-plane shear fracture strain is an intrinsic material parameter obtained from physical experiments. The evolution coefficient is determined by fitting the stress-strain curve from the finite element simulation.

[0151] Damage variable update logic: In each time step of the finite element simulation, the stress and equivalent plastic strain in each direction are calculated first; if the stress reaches the initial threshold of the corresponding damage mode, damage evolution is initiated and the damage variables are updated. , , .

[0152] Step S222: Define the interlaminar damage model.

[0153] Based on the cohesive region model, considering tensile / shear damage in the thickness direction, interlaminar damage variables are introduced. It corrects interlayer stiffness and covers interlayer delamination damage.

[0154] Step S223: Define damage coupling.

[0155] In-plane damage ( , , ) and interlaminar damage ( Coupling is achieved through cross-dimensional stress transfer and damage accumulation.

[0156] Based on the three-dimensional stress equilibrium equation of laminated plates, the stress transfer relationship between the in-plane and interlayer planes is clarified, and the transferred stress is calculated using a quantitative formula:

[0157] In-plane damage propagates to the interlaminar plane: In-plane fiber tension / compression, transverse tension / compression, and in-plane shear damage can lead to degradation of in-plane stiffness, making it impossible for local in-plane stress to be uniformly borne. Some stress is transferred to the thickness direction through the interlaminar interface, forming additional interlaminar stress.

[0158] The formula for calculating interlayer stress is:

[0159] ;

[0160] ;

[0161] ;

[0162] In the formula, The actual in-plane stress is calculated in real time by finite element simulation. For in-plane damage variables; is the stress transfer coefficient, which is obtained from the interfacial shear test of the laminate.

[0163] Interlaminar tensile and shear damage can lead to interlaminar stiffness degradation, preventing interlaminar stress from being effectively transferred to adjacent plies. This can cause in-plane stress concentration in adjacent plies, resulting in additional in-plane stress.

[0164] The formula for calculating in-plane additional stress is:

[0165] ;

[0166] ;

[0167] ;

[0168] In the formula, The actual interlayer stress is calculated in real time by the cohesive zone model; For interlaminar damage variables; The stress transfer coefficient from interlaminar to in-plane is obtained from interlaminar tensile tests on laminates.

[0169] A coupling correction factor is introduced to correct the critical values ​​for intra-layer and inter-layer damage:

[0170] (1) In-plane damage critical value correction

[0171] Critical value for fiber-direction tensile strength:

[0172] ;

[0173] Fiber orientation compression critical value:

[0174] ;

[0175] Transverse tensile critical value:

[0176] ;

[0177] Lateral compression critical value:

[0178] ;

[0179] In-plane shear critical value:

[0180] ;

[0181] In the formula, The basic critical value of in-plane damage under uncoupled conditions is determined by the Hashin criterion. This is the coupling correction coefficient, calibrated experimentally.

[0182] (2) Correction of critical value for interlaminar damage

[0183] Critical value for interlaminar tension:

[0184] ;

[0185] Interlaminar shear critical value:

[0186] ;

[0187] ;

[0188] In the formula, The critical value for interlayer damage under uncoupled conditions is determined by cohesive zone model tests. This is an interlayer coupling correction factor, calibrated experimentally.

[0189] Define coupling damage variables to quantify the overall damage level:

[0190] ;

[0191] In the formula, Weights for in-plane / interlayer damage; ,when When the value is too large, it is determined to be a critical state of global damage, triggering accelerated degradation of overall stiffness.

[0192] In each time step of the finite element simulation, stiffness failure is triggered according to the following procedure to ensure the continuity and realism of damage evolution:

[0193] 1. Calculate the actual total stress: Total in-plane stress =In-plane initial stress +In-plane additional stress ;

[0194] Total interlayer stress =Interlayer initial stress + Interlayer additional stress ;

[0195] 2. Damage evolution and update: If ,but The growth accelerates according to the evolution equation in step S221;

[0196] like ,but Accelerated growth is achieved according to the evolution equation of the cohesive region model.

[0197] Stiffness component failure determination: If a certain damage variable is too high, the corresponding stiffness component is reduced to zero; if... If the value is too high, the global stiffness matrix will return to zero, and the material will completely fail.

[0198] Step S23: Establish an impact simulation model through finite element simulation, apply high-speed impact load and low-speed impact load to carbon fiber composite material, calculate and output the peak impact force, impact energy absorption, impact displacement and component mass under the dual load conditions as the training sample benchmark values ​​of the radial basis function surrogate model.

[0199] An impact simulation model was established using finite element software, with the following specific settings:

[0200] Step S231: Calculate the complete equivalent stiffness matrix of the laminated plate. The damage model parameters are input into the simulation model to ensure the accuracy of the mechanical response calculation under multi-directional stress.

[0201] Step S232: Apply dual-condition loads of high-speed impact and low-speed impact to simulate contact, collision and energy transfer during the impact process.

[0202] Step S233: Calculate and output the peak impact force under dual operating conditions. Impact energy absorption Impact displacement The component mass m is recorded, along with data such as damage distribution and stiffness degradation history, which serve as the training sample benchmark values ​​for the surrogate model.

[0203] Step S24: Based on the baseline values ​​of the training samples, determine the shape parameters of the radial basis function using leave-one-out cross-validation and construct a radial basis function surrogate model. This specifically includes the following steps:

[0204] Step S241: Construct the proxy model.

[0205] Define the model structure:

[0206] ;

[0207] in, is the distance between the new design point and the training sample point in the design space; c is the shape parameter.

[0208] Methods for determining the shape parameter c include:

[0209] (1) Determining the initial range: First, normalize all design variables and calculate the minimum distance between sample points. Maximum distance The initial range of values ​​for c is determined based on the maximum and minimum distances.

[0210] (2) Cross-validation to select the optimal c: Leave-one-out cross-validation is used, combined with dual-quantitative evaluation indicators to select the optimal c. Evaluation indicators are calculated for the seven output indicators of the surrogate model, and the weighted average of the seven indicators is used as the selection criterion. Candidate values ​​within the initial range of c are traversed, and a surrogate model is constructed for each candidate c and the root mean square error is calculated. With the coefficient of determination Select those that simultaneously satisfy Minimum and The largest c is taken as the optimal value; if multiple cs satisfy the condition, the closest one is preferred. Candidate values ​​for 1.

[0211] The optimal c selected is verified from both physical meaning and optimization suitability aspects.

[0212] The optimal requirement is to satisfy This ensures that the interpolation characteristics of the radial basis functions conform to the mapping law between material parameters and properties; if This can easily lead to model overfitting; if This can easily lead to underfitting of the model, and is therefore deemed unreasonable.

[0213] The surrogate model built based on the optimal c has a small relative error in predicting performance parameters for any two sets of similar design variables, ensuring the local smoothness of the model in the design space and adapting to the iterative optimization requirements of the genetic algorithm.

[0214] If all the above criteria are met, c is adopted directly; if not, the search range of c is expanded, and cross-validation and rationality verification are performed again until a qualified c is selected.

[0215] Establish mapping relationships:

[0216] The proxy model output is The input is ;

[0217] ;

[0218] in, This is a 7-dimensional interpolation coefficient vector, corresponding to the 7 output dimensions.

[0219] The normalization method for the ply order encoding vector is as follows:

[0220] For each element in the integer encoding sequence ( Normalize (where n is the total number of layers):

[0221] ;

[0222] Normalized summary items:

[0223] ;

[0224] Used to convert discrete codes into continuous variables and to adapt the distance calculation to radial basis functions.

[0225] It is a 7-dimensional low-order polynomial vector, where each polynomial... Using a linear polynomial form, that is This is used to compensate for the bias of the radial basis function in the global fitting;

[0226] The output dimension is defined as:

[0227] Peak impact force under low-speed impact conditions;

[0228] : Maximum impact displacement under low-speed impact conditions;

[0229] Energy absorption under low-speed impact conditions;

[0230] Peak impact force under high-speed impact conditions;

[0231] : Maximum impact displacement under high-speed impact conditions;

[0232] Energy absorption under high-speed impact conditions;

[0233] Component quality;

[0234] , and The model was trained using sample data of seven indices obtained by the QR decomposition method and based on finite element simulation.

[0235] Step S242: Use leave-one-out cross-validation to verify the accuracy of the seven output indicators of the surrogate model, ensuring that the prediction error of each indicator meets the optimization design requirements.

[0236] Step S3: Use the NSGA-II multi-objective genetic algorithm to iteratively optimize the material parameters in the target performance model and generate a set of optimal solutions.

[0237] An objective function is constructed that includes minimum impact force, maximum energy absorption, and minimum mass. Weighting coefficients for different operating conditions are added to the objective function to improve the material's adaptability to the target. The weight of the low-velocity impact condition is set as follows: The weight of high-speed impact conditions is And satisfy ;

[0238] The expression for the weighted minimum impact peak objective function is:

[0239] ;

[0240] In the formula, The objective function is the weighted minimum impact force peak value. and These represent the peak impact forces under low-speed and high-speed impact conditions, respectively. This represents the theoretical minimum value of the peak impact force under dual operating conditions. This is the maximum peak impact force setting value; , These are the weights for low-speed impact conditions and high-speed impact conditions, respectively.

[0241] The expression for the weighted maximum energy absorption objective function is:

[0242] ;

[0243] In the formula, and Energy absorption under low-speed impact conditions and high-speed impact conditions, respectively; This represents the theoretical maximum energy absorption value under dual operating conditions. This represents the minimum allowable energy absorbed after weighting for both operating conditions.

[0244] The expression for the minimum mass objective function is:

[0245] ;

[0246] In the formula, For component quality; This represents the theoretical minimum mass of the component. Set the maximum component mass value.

[0247] At the same time, constraints are introduced:

[0248] Displacement constraints under low-speed impact conditions: ;

[0249] Displacement constraints under high-speed impact conditions: ;

[0250] During the iteration process of the NSGA-II algorithm, solutions that do not meet the displacement constraints are directly marked as invalid solutions and are not included in the non-dominated sorting.

[0251] Algorithm parameters include population size N and crossover probability. Probability of mutation and maximum number of iterations The range of values ​​for each parameter is determined by theoretical formulas, and is evaluated using a comprehensive evaluation function. Determine the optimal combination of parameters.

[0252] The theoretical lower bound of population size N for:

[0253] ;

[0254] in, Optimize empirical coefficients for multiple objectives; To design variable dimensions;

[0255] Theoretical upper limit for:

[0256] ;

[0257] in, This represents the maximum allowed time for a single set of iterations. The maximum number of iterations is preset. The computation time for the objective function of a single sample; This is the engineering margin coefficient.

[0258] Crossover probability correction range for:

[0259] ;

[0260] in, This provides a general crossover probability interval for multi-objective algorithms. These are discrete correction coefficients.

[0261] Population diversity The expression is:

[0262] ;

[0263] in, For the first The objective function value of the j-th sample; Let be the mean of the j-th objective function; The overall mean of the three objective functions.

[0264] During the iteration process, make To meet engineering requirements and verify the rationality of the crossover probability.

[0265] Mutation probability Theoretical lower limit for:

[0266] ;

[0267] Where s is the accuracy level of the average value of a single variable.

[0268] Probability of variation of discrete variables Set separately as:

[0269] ;

[0270] Where l represents the number of angle encoding types; This represents the total number of discrete design variables.

[0271] The theoretical upper limit is:

[0272] ;

[0273] in, It is the reciprocal of the proportion of invalid intervals in a single-variable process; This represents the algorithm's stability coefficient.

[0274] Maximum number of iterations The theoretical lower bound is determined by the convergence criterion. The decision made The convergence condition is met, thus determining the theoretical lower limit. , The expression is:

[0275] ;

[0276] in, Let G be the solution set coverage of generation G; The solution set coverage of generation G-10;

[0277] The theoretical upper limit is:

[0278] ;

[0279] in, This represents the maximum allowed time for a single set of iterations. Population size; The time taken to calculate a single sample.

[0280] Determining the optimality of parameter combinations:

[0281] ;

[0282] ;

[0283] ;

[0284] ;

[0285] in, For solution set coverage; To satisfy the number of effective non-dominated solutions; This represents the theoretical maximum number of nondominated solutions. For convergence speed; This represents the number of iterations required for the algorithm to converge to a stable solution set. This represents the maximum number of iterations. The percentage of invalid proposals; This represents the number of solutions that exceed process constraints after modification. This represents the total number of schemes after the mutation; , , These are the weighting coefficients of the parameters.

[0286] The NSGA-II multi-objective genetic algorithm is used to iteratively optimize the material parameters in the target performance model, including the following steps:

[0287] Step S31: Based on the sensitivity of material parameters to high-speed and low-speed impact conditions, divide the parameter value range into high-speed adaptation range, low-speed adaptation range, and trade-off range.

[0288] Step S32: Using the Latin hypercube sampling method, generate [samples] within the range of values ​​for the design variables. The initial design scheme ensures a uniform distribution of samples within the design space. The sampling rules for the ply sequence coding include: randomly selecting n elements from the angle coding set {0,1,2,3} to form an initial coding sequence, and verifying the symmetric ply constraint to ensure the coding sequence satisfies:

[0289] ;

[0290] Where m = 1, 2, 3, ..., n / 2, if the condition is not met, resampling will be performed.

[0291] Sampling needs to cover the high-speed adaptation range, the low-speed adaptation range, and the trade-off range to avoid local optima.

[0292] Step S33: For each design scheme, calculate its... , , The dominance relationship is determined by classifying the population according to the dominance level. The first level represents non-dominated solutions, the second level represents solutions dominated only by the first level, and so on.

[0293] Step S34: For each non-dominated solution, calculate its distance to neighboring solutions in the target space, i.e., the crowding degree:

[0294] ;

[0295] Where i is the index of the current non-dominated solution in the set of non-dominated solutions after sorting in ascending order by the value of the j-th objective function; This represents the j-th objective function value of the i-th non-dominated solution after sorting. Let j be the objective value of the adjacent solutions of the i-th non-dominant solution after sorting; Let be the maximum and minimum values ​​of the j-th objective among all non-dominated solutions.

[0296] Step S35: Employing a roulette wheel + elite retention strategy, non-dominated solutions and solutions with high crowding are prioritized for inclusion in the parent population. The crossover operation uses an improved partial matching crossover strategy adapted to the layer sequence encoding; continuous variables still use single-point crossover, executed synchronously, including the following steps:

[0297] To address the characteristics of layer-by-layer encoding that allows duplicate elements and satisfies symmetry constraints, an improved crossover logic is designed, with the following steps:

[0298] Step S3501: Randomly select two intersection points for parent generation 1 with encoding sequence S1 and parent generation 2 with encoding sequence S2. , ,and ;

[0299] Step S3502: Extract the segment before the intersection of parent generation 1 Extract the segment before the intersection of parent generation 2. ;

[0300] Step S3503: Construct fragment mapping relationship: for and For each element in the set, a bidirectional mapping is established, so that even if elements are repeated, they are still mapped sequentially according to their position. , ;

[0301] Step S3504: Generate offspring skeleton: Retain the segments other than the crossover points of parent generation 1, leave the crossover point positions empty for the time being, and form the preliminary offspring sequence. ;

[0302] Step S3505: Fill the cross segments: Fill the cross segments of parent generation 2. Fill to The intersection point is determined, and the symmetrical position of the intersection point is filled in according to the symmetry constraint. to ,make sure ;

[0303] Step S3506: Conflict Correction: If, after filling the cross segment, the area outside the cross point contains elements that conflict with the mapping relationship, replace them with the corresponding elements of parent generation 2 according to the mapping relationship to ensure the validity of the encoding;

[0304] Step S3507: Verify symmetric plying constraints: If the child generation encoding does not satisfy:

[0305] ;

[0306] Then, elements in asymmetrical positions are corrected;

[0307] Step S3508: For continuous variables, use the single-point crossover method: randomly select a crossover point, swap the variable values ​​after the crossover point of parent generation 1 and parent generation 2, and generate the parameters of the continuous variable of the offspring generation;

[0308] The mutation operation uses a position-swap mutation strategy to adapt to the layer sequence encoding, while continuous variables still use a random mutation method. The operations are performed synchronously and include the following steps:

[0309] Step S3511: According to discrete mutation probability Randomly select two symmetrical positions m and n+1-m;

[0310] Step S3512: Swap the codes at these two positions to ensure that the symmetry constraint is still satisfied after mutation;

[0311] Step S3513: If the mutated coding sequence contains invalid combinations, such as three consecutive codes with the same angle, then mutate again until a valid code is generated.

[0312] Step S36: Merge the parent N-group schemes and the offspring N-group schemes into 2N-group schemes, and repeat the non-dominated sorting and crowding calculation; select N-group schemes as the new generation population according to the principle of priority of dominance level and priority of high crowding at the same level.

[0313] Step S37: Repeat steps S33 to S36 until the maximum number of iterations is reached. When the number and distribution of non-dominated solutions in the population do not change significantly over 10 consecutive generations, convergence is determined; the Pareto optimal solution set is output, which covers the trade-offs between different objectives.

[0314] In actual working conditions, the requirements for key parameters in high-speed impact conditions are opposite to those in low-speed impact conditions. High-speed impact conditions require... Smaller size to reduce peak impact force; low-velocity impact conditions require... While larger values ​​are used to increase energy absorption, differences exist in the adaptation requirements for parameters such as layer sequence and wall thickness, necessitating a coordination mechanism. Therefore, the NSGA-II multi-objective genetic algorithm in this embodiment of the invention also introduces a dual-condition conflict coordination mechanism, which includes the following steps:

[0315] Step A: Based on the 7-dimensional output data of the surrogate model, define the conflict quantification index ΔP to characterize the degree of contradiction in performance parameters under dual operating conditions:

[0316] ;

[0317] In the formula, These are the performance parameters under high-speed impact conditions. These are the corresponding parameters under low-speed impact conditions; This represents the average parameter values ​​under both operating conditions.

[0318] Step B: Define the conflict resolution logic for the NSGA-II algorithm:

[0319] (1) Population initialization bias: When performing Latin hypercube sampling, the range of conflict parameters is sampled in a denser manner to ensure coverage of the high-speed adaptation interval, low-speed adaptation interval, and trade-off interval, thus avoiding local optima.

[0320] (2) Non-dominated sorting optimization: In fast non-dominated sorting, a conflict adaptation factor K is introduced. Schemes with small performance differences between the two working conditions are given a sorting weight of K and are given priority to be included in the high dominance level, guiding the algorithm to evolve towards conflict coordination.

[0321] Overall difference coefficient:

[0322] ;

[0323] ;

[0324] ;

[0325] ;

[0326] ;

[0327] In the formula, , These are the target weights for crashworthiness and energy absorption, respectively. A dual-condition collision quantification index for crashworthiness targets; A quantitative indicator for dual-condition conflict of energy-absorbing targets; These represent the peak impact forces under low-speed and high-speed impact conditions, respectively. They represent the energy absorption under low-speed and high-speed impact conditions, respectively.

[0328] The expression for the conflict adaptation factor is:

[0329] ;

[0330] In the formula, This is the maximum value of the comprehensive difference coefficient among all schemes in the solution set.

[0331] In non-dominated ranking, the dominance level score of a solution is equal to the base level score multiplied by K, where the base level scores are 1, 2, 3, ..., with higher-level solutions having lower scores. When comparing the priorities of solutions, the dominance level scores are compared first, with lower scores indicating higher priorities. If the scores are the same, the crowding is compared next, with higher crowding indicating higher priorities.

[0332] Step C: Increase the weight of congestion calculation for the performance dimensions corresponding to the conflict parameters.

[0333] ;

[0334] in, For the first The dual-condition conflict quantification index corresponding to each objective function ensures that the scheme with better conflict coordination effect obtains higher crowding degree and is retained in the population.

[0335] Step D: Precisely screen conflicting solutions:

[0336] (1) Pareto solution set classification: The output Pareto optimal solution is divided into three categories according to the conflict handling effect: high-speed adaptation, low-speed adaptation, and conflict coordination.

[0337] (2) Weight-oriented screening: combining working condition weights , Differentiated selection was conducted for the three types of solutions:

[0338] like Prioritize retaining low-speed adaptation and conflict coordination solutions, and eliminate high-speed adaptation solutions;

[0339] like Prioritize retaining high-speed adaptation and conflict coordination solutions, and eliminate low-speed adaptation solutions;

[0340] like Only conflict coordination solutions are retained to ensure that the requirements of both working conditions are met.

[0341] (3) Final decision-making closed loop: Input the selected conflict coordination solutions into the grey relational analysis in step S4, and combine the target weight to achieve comprehensive optimization, so as to ensure that the conflict handling is accurately matched with the application scenario.

[0342] Step S4: Employ a grey relational analysis method based on dual-weight fusion to quantitatively filter the set of optimal solutions, outputting the comprehensive optimal carbon fiber composite material structure scheme to achieve the optimal requirements for material impact resistance, energy absorption, and lightweighting. This includes the following steps:

[0343] Step S41: Extract 7 core evaluation indicators, including Normalization is performed according to the target requirements.

[0344] Minimize metrics ( ):

[0345] ;

[0346] Maximize the index ( ):

[0347] ;

[0348] in, The original data for the j-th objective of the i-th scheme. , These are the maximum and minimum values ​​of the j-th target, respectively.

[0349] Step S42: Based on the application scenario, prioritize the three objectives and set their weights:

[0350] ;

[0351] satisfy ,in, Weights for crashworthiness targets; As the weight of the energy absorption target; To reduce the target weights.

[0352] Combined with operating condition weights , The target weights are decomposed into 7 indicators. Internal weight allocation coefficients are set. Used to split the crashworthiness target weights , The internal weight of the peak impact force is not explicitly required and is set by default. To ensure a balanced and compatible internal performance index for crashworthiness, 1- This represents the internal weight of the impact displacement.

[0353] Formulas for calculating the weights of the 7 indicators:

[0354] Weight: ;

[0355] Weight:

[0356] Weight: ;

[0357] Weight: ;

[0358] Weight: ;

[0359] Weight: ;

[0360] m-weight: .

[0361] Step S43: For each normalized index The maximum value of this index among all Pareto schemes is taken as the ideal value. ,Right now:

[0362] ;

[0363] Combining indicator weights To ensure that the ideal sequence meets the principle of prioritizing high-weight indicators; to construct the ideal optimal sequence. .

[0364] Step S44: Design the resolution coefficient based on the difference in composite material indicators Calculate the coefficient of variation for each individual indicator:

[0365] ;

[0366] in, Let j be the standard deviation of the normalized index. The mean of the j-th normalization index is calculated from the normalized data of all schemes in the Pareto solution set.

[0367] Calculate the comprehensive value of the index difference:

[0368] .

[0369] Quantify the degree of similarity between the i-th scheme and the j-th index of the ideal optimal sequence:

[0370] ;

[0371] in, The value range is [0,1]. The closer it is to 1, the closer the j-th index of the i-th scheme is to the ideal state.

[0372] Combining indicator weights We obtain the comprehensive correlation degree of the i-th scheme by weighted summation of the correlation coefficients:

[0373] ;

[0374] in, The value range is [0,1]. The closer it is to 1, the better the overall performance of the scheme.

[0375] Step S45: For all solutions in the Pareto solution set, sort them by overall correlation degree. Sort in descending order; select The largest possible solution is considered the optimal solution overall.

[0376] Output the complete parameter system of the optimal solution, including:

[0377] Material parameters: fiber volume fraction The protrusion height of the phthalocyanine iron microstructure ;

[0378] Structural parameters: ply angle Layer sequence and layer ratio Wall thickness Manager ;

[0379] Dual-mode ;

[0380] Among them, the output component mass m is a globally unique value that satisfies The requirement for lightweight design.

[0381] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described; only preferred embodiments of the present invention are illustrated. The descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. As long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0382] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this invention should be determined by the appended claims.

Claims

1. An optimization design method for carbon fiber composite materials, characterized in that, Includes the following steps: Step S1: Obtain the impact resistance requirements, energy absorption requirements, and lightweight requirements for carbon fiber composite materials; Step S2: Construct a target performance model for carbon fiber composite materials and predict the performance parameters of carbon fiber composite materials under high-speed impact and low-speed impact conditions based on material parameters. Step S3: Use the NSGA-II multi-objective genetic algorithm to iteratively optimize the material parameters in the target performance model to generate a set of optimal solutions; Step S4: The set of optimal solutions is quantitatively screened using a grey relational analysis method based on dual-weight fusion, and the comprehensive optimal carbon fiber composite material structure scheme is output to achieve the optimal requirements for material impact resistance, energy absorption and lightweighting.

2. The optimized design method for carbon fiber composite materials according to claim 1, characterized in that: The carbon fiber composite material mentioned in step S1 is a material system with continuous carbon fiber as reinforcement, polyarylene ether nitrile as resin matrix, and phthalocyanine iron as the microstructure building block of carbon fiber surface.

3. The optimized design method for carbon fiber composite materials according to claim 1, characterized in that: The impact resistance requirement in step S1 is to minimize the peak impact force; the energy absorption requirement is to maximize the energy absorption; and the lightweight requirement is to minimize the component mass.

4. The optimized design method for carbon fiber composite materials according to claim 1, characterized in that: The target performance model mentioned in step S2 is a radial basis function surrogate model; the input of the radial basis function surrogate model is material parameters, which include material variables and structural variables. The material variables include at least the fiber volume fraction and the protrusion height of the phthalocyanine iron microstructure, and the structural variables include at least the layup angle, layup sequence, layup ratio, wall thickness, and tube length. The output of the radial basis function surrogate model is a performance parameter, which includes at least the peak impact force under low-speed impact conditions, the maximum impact displacement under low-speed impact conditions, the energy absorbed under low-speed impact conditions, the peak impact force under high-speed impact conditions, the maximum impact displacement under high-speed impact conditions, the energy absorbed under high-speed impact conditions, and the component mass.

5. The optimized design method for carbon fiber composite materials according to claim 4, characterized in that: The construction of the target performance model for carbon fiber composite materials in step S2 includes at least the following steps: Step S21: Construct the complete stiffness matrix of the laminate based on the classical laminate theory, and establish a mapping model between design variables and initial mechanical properties; Step S22: Establish an in-plane damage model and an interlaminar damage model based on the theory of continuous medium damage mechanics. The damage variables of the in-plane damage model include at least fiber-direction tensile-compression damage variables, transverse tensile-compression damage variables, and in-plane shear damage variables. The interlaminar damage model introduces interlaminar damage variables based on the cohesive region model. Step S23: Establish an impact simulation model through finite element simulation, apply high-speed impact load and low-speed impact load to carbon fiber composite material, calculate and output the peak impact force, impact energy absorption, impact displacement and component mass under the dual conditions as the training sample benchmark values ​​of the radial basis function surrogate model. Step S24: Based on the benchmark values ​​of the training samples, the shape parameters of the radial basis functions are determined using leave-one-out cross-validation, and the radial basis function surrogate model is constructed.

6. The optimized design method for carbon fiber composite materials according to claim 1, characterized in that: The objective functions of the NSGA-II multi-objective genetic algorithm in step S3 include at least the weighted minimum impact force peak objective function, the weighted maximum energy absorption objective function, and the minimum mass objective function; The objective function for the weighted minimum impact force peak value is: ； In the formula, The objective function is the weighted minimum impact force peak value. and These represent the peak impact forces under low-speed and high-speed impact conditions, respectively. This represents the theoretical minimum value of the peak impact force under dual operating conditions. This is the maximum peak impact force setting value; For low-speed impact conditions, the weighting coefficient is used. This represents the weighting coefficient for high-speed impact conditions. The weighted maximum energy absorption objective function is: ； In the formula, and Energy absorption under low-speed impact conditions and high-speed impact conditions, respectively; This represents the theoretical maximum energy absorption value under dual operating conditions. This represents the minimum allowable energy absorbed after weighting for both operating conditions. The minimum mass objective function is: ； In the formula, For component quality; This represents the theoretical minimum mass of the component. Set the maximum component mass value.

7. The optimized design method for carbon fiber composite materials according to claim 6, characterized in that: Step S3, which involves iteratively optimizing the material parameters in the target performance model using the NSGA-II multi-objective genetic algorithm, includes the following steps: Step S31: Based on the sensitivity of material parameters to high-speed and low-speed impact conditions, divide the parameter value range into high-speed adaptation range, low-speed adaptation range, and trade-off range. Step S32: Use the Latin hypercube sampling method to generate an initial design scheme within the range of material parameter values, with sampling covering the high-speed adaptation range, low-speed adaptation range, and trade-off range. Step S33: Calculate the dominance relationship of each design scheme under the weighted minimum impact force peak objective function, the weighted maximum energy absorption objective function, and the minimum mass objective function, and classify the population according to the dominance level; Step S34: Calculate the distance between each non-dominated solution and its neighboring solutions in the target space; Step S35: Use roulette wheel selection and elite retention strategies for selection, use an improved partial matching crossover strategy to adapt the crossover operation of the ply sequence encoding, and use a position swap mutation strategy to adapt the mutation operation of the ply sequence encoding. Step S36: Select a new generation of population according to the principles of priority based on dominance level and priority based on high crowding at the same level; Step S37: Repeat steps S33 to S36 until the number of iterations reaches the maximum number of iterations or the convergence condition is met, and output the set of optimal solutions.

8. The optimized design method for carbon fiber composite materials according to claim 7, characterized in that: The NSGA-II multi-objective genetic algorithm described in step S3 introduces a dual-condition conflict coordination mechanism, which includes the following steps: Step A: Define conflict quantification metrics This characterizes the degree of contradiction in performance parameters under dual operating conditions: ； In the formula, These are the performance parameters under high-speed impact conditions. These are the corresponding parameters under low-speed impact conditions; The average values ​​of parameters under both operating conditions; Step B: Constructing the comprehensive difference coefficient Quantifying the performance differences between the two operating conditions: ； In the formula, , These are the target weights for crashworthiness and energy absorption, respectively. A dual-condition collision quantification index for crashworthiness targets; A quantitative indicator for dual-condition conflict of energy-absorbing targets; Step C: Introduce conflict adaptation factor Schemes with small performance differences between the two operating conditions are assigned ranking weights: ； In the formula, The maximum value of the comprehensive difference coefficient among all solutions in the solution set; Step D: Quantify conflict indicators The corresponding performance parameters increase the weight of congestion calculation.

9. The optimized design method for carbon fiber composite materials according to claim 6, characterized in that: Step S4, which involves using a grey relational analysis method based on dual-weight fusion to quantitatively filter the set of optimal solutions, includes the following steps: Step S41: Extract core evaluation indicators, which include at least the peak impact force under low-speed impact conditions, the maximum impact displacement under low-speed impact conditions, the energy absorbed under low-speed impact conditions, the peak impact force under high-speed impact conditions, the maximum impact displacement under high-speed impact conditions, the energy absorbed under high-speed impact conditions, and the component mass. Normalize the core evaluation indicators according to the target type. Step S42: Set the target weights for crashworthiness requirements Target weights for energy absorption requirements and the target weights for lightweight requirements Combined with operating condition weights and The target weight is decomposed into the weights of each core evaluation indicator; Step S43: Take the maximum value of each normalized core evaluation index among all schemes as the ideal value, and construct the ideal optimal sequence; Step S44: Quantify the degree of closeness of each scheme to the ideal optimal sequence for each indicator to obtain the correlation coefficient. Combine the indicator weights to perform a weighted summation of the correlation coefficients to obtain the comprehensive correlation degree. The expression for the correlation coefficient is: ； In the formula, Let be the correlation coefficient between the i-th scheme and the j-th index in the ideal optimal sequence; Let j be the index value of the j-th item in the ideal optimal sequence; Let be the normalized value of the j-th indicator of the i-th scheme; The resolution coefficient; The expression for the comprehensive correlation degree is: ； In the formula, Let be the overall correlation degree of the i-th scheme; Let be the weight of the j-th indicator; Step S45: Sort all schemes in the set of optimal solutions in descending order according to the comprehensive correlation degree, select the scheme with the highest comprehensive correlation degree as the comprehensive optimal scheme and output it.

10. The optimized design method for carbon fiber composite materials according to claim 9, characterized in that: Step S42, which involves decomposing the target weight into the weights of each core evaluation indicator, includes the following steps: setting internal weight allocation coefficients. , used to decompose the crashworthiness target weights, where As the internal weight of the peak impact force, 1- The internal weight of the impact displacement; Weight of peak impact force under low-velocity impact conditions for: ； Weight of maximum impact displacement under low-speed impact conditions for: ； Weight of energy absorption under low-speed impact conditions for: ； Weight of peak impact force under high-speed impact conditions for: ； Weight of maximum impact displacement under high-speed impact conditions for: ； Weight of energy absorption under high-speed impact conditions for: ； Weight of component quality for: 。

Images