Parameter optimization method and device for composite material molding process

Abstract

The present invention discloses a parameter optimization method and device for a composite material forming process, and the present invention relates to the technical field of composite materials. The method comprises the following steps: obtaining a finished product with known mechanical characteristic parameters and its corresponding process parameters, and mapping to generate a training sample data set. A neural prediction network is established based on the training sample data set, and the process parameters are used as inputs and the mechanical characteristic parameters are used as labels for training to obtain a mechanical characteristic prediction model. A plurality of optimized process parameter combinations are generated, and their corresponding mechanical characteristic parameter values ​​are predicted to establish a fitness function. The optimal process parameter combination with the highest fitness value is screened out through a genetic algorithm. Finally, based on process flow parameters, such as the average length of the fiber and the curing time, the curing dynamic factor and the reaction factor are calculated, and the optimal process parameter combination is corrected to complete parameter optimization. The comprehensive performance of the product is improved and material waste in the production process is reduced.

Description

Technical Field

[0001] The invention relates to the technical field of composite materials, and in particular to a parameter optimization method and device for a composite material molding process. Background Art

[0002] With the continuous development of materials science and engineering technology, composite materials have been widely used in aerospace, automobile and construction fields due to their excellent mechanical properties and good designability. However, the molding process of composite materials is complex, involving the mutual influence and coordination of multiple parameters. These process parameters include fiber volume ratio, resin viscosity, average mold temperature, etc., which directly affect the mechanical properties of the finished product, such as tensile strength, flexural strength and interlaminar shear strength. Therefore, how to optimize the molding process parameters of composite materials to achieve high-performance finished products is a key technical problem that needs to be solved in the current material processing field.

[0003] Traditional optimization methods for composite molding processes often rely on experience and trial and error, resulting in long production cycles, high costs, and low efficiency. Under different molding conditions, the mechanical properties of composite materials show great uncertainty and complexity, making it difficult to reasonably predict them through simple models. In this case, the lack of effective mathematical models and calculation tools makes it difficult to achieve precise control of mechanical properties during the production process. Therefore, there is an urgent need for a data-driven optimization method to improve the robustness and repeatability of composite molding processes, thereby reducing production costs and improving product quality.

[0004] In addition, existing optimization methods often ignore the dynamic impact analysis of key factors such as curing dynamics and fiber distribution, resulting in deviations in optimization results in practical applications. For example, the resin may produce local defects due to uneven fiber distribution during flow, or incomplete curing reaction may lead to decreased mechanical properties. Therefore, how to comprehensively consider the complex relationship between process parameters and mechanical properties through scientific methods has become a key technical problem that needs to be solved in the field of composite material molding process optimization.

[0005] In the prior art, the publication number CN116663332B discloses a parameter optimization method, device, equipment and medium for a composite material forming process, wherein the method includes: using Hypermesh software to construct a target impregnation model of the composite material; importing the target impregnation model into the PAM-RTM software, and constructing a target porosity model in the PAM-RTM software; inputting multiple sets of input parameter sets of the composite material forming process into the PAM-RTM software to obtain the porosity distribution and impregnation rate distribution of the target impregnation model; wherein the porosity distribution includes the total porosity corresponding to each set of input parameter sets, and the impregnation rate distribution includes the impregnation rate corresponding to each set of input parameter sets; based on the porosity distribution and the impregnation rate distribution, the parameters of the composite material forming process are optimized. However, this scheme relies on specific software for modeling and simulation. When the parameters change, the modeling process needs to be repeated, and the scheme lacks analysis of the mechanical characteristics of the finished composite material. Therefore, this scheme may reduce the accuracy and effectiveness of parameter optimization.

[0006] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0007] The purpose of the present invention is to provide a method and device for optimizing parameters of a composite material molding process, so as to solve the problems raised in the above background technology.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] A parameter optimization method for a composite material molding process, the specific steps comprising:

[0010] Obtaining several composite product products with known mechanical characteristic parameters, collecting corresponding process parameters in the manufacturing process of the composite product products, mapping the obtained process parameters with the corresponding mechanical characteristic parameters of the composite product products one by one, and generating a training sample data set, wherein the mechanical characteristic parameters include tensile strength, bending strength and interlaminar shear strength, and the process parameters include fiber volume ratio, resin viscosity and average mold temperature;

[0011] Based on the data in the training sample data set, a neural prediction network is established, the process parameters in the training sample data set are used as the input of the model, and the corresponding mechanical characteristic parameters are used as labels to train the neural prediction network to obtain a mechanical characteristic prediction model;

[0012] Obtaining the value range of each process parameter, randomly generating several groups of optimized process parameter combinations within the value range, inputting different optimized process parameter combinations into the trained mechanical characteristic prediction model, obtaining the mechanical characteristic parameter prediction value corresponding to the combination, and establishing a fitness function according to the mechanical characteristic parameter prediction value, wherein the optimized process parameter combination includes three gene positions corresponding to the fiber volume ratio, the resin viscosity and the average mold temperature respectively;

[0013] A set of optimized process parameter combinations is taken as an individual, and the process parameters in the combination are taken as genes. Based on the fitness function, an optimal process parameter combination is obtained by a genetic algorithm, wherein the optimal process parameter combination is an optimized process parameter combination with the highest fitness function value;

[0014] The process parameters are obtained, and the curing kinetic factor and the fiber distribution factor are calculated based on the obtained process parameters. The parameters in the optimal process parameter combination are corrected according to the obtained curing kinetic factor and the curing reaction factor to obtain the precise values ​​of the process parameters, and the parameter optimization of the composite material molding process is completed. The process parameters include the average length of the mold fiber, the angle of the fiber relative to the resin flow direction, the injection flow rate of the resin, the curing time and the ambient humidity.

[0015] Furthermore, the training sample set is generated by mapping the process parameters to the mechanical characteristic parameters of the corresponding composite material finished product one by one to form a corresponding grid, and recording the formed grid as the training sample data set;

[0016] Based on the data in the training sample set, a neural network prediction model is established, in which the long short-term memory network model LSTM model is selected as the base model, and the activation function and optimization algorithm are selected. The Tanh function is selected as the activation function, and Adam is selected as the optimization algorithm of the LSTM model; the formula of the Tanh function is:

[0017] ;

[0018] In the formula, Represents the Tanh function, independent variable Represents the weighted sum of the neuron's input, that is, the result of weighted summation of the input received by the neuron from the previous layer;

[0019] At the same time, the hyperparameters of the LSTM model are set, and the hyperparameters of the LSTM model include: the number of network layers, the number of iterations, the learning rate, the batch size, the number of training times, the batch processing number, and the number of hidden layer neurons;

[0020] The number of network layers is set to 4 layers, the number of iterations is set to 300, the learning rate is set to 0.001, the batch size is set to 32, the number of training times is set to 100, the batch size is set to 256, and the number of hidden layer neurons is 32;

[0021] The trained mechanical characteristic prediction model takes process parameter data as input and outputs mechanical characteristic parameter prediction values, which include tensile strength prediction values, bending strength prediction values ​​and interlaminar shear strength prediction values.

[0022] Furthermore, the logic underlying the random generation of several groups of optimized process parameter combinations within a value range is as follows: randomly generate several different data within the value range of each optimized process parameter as optional values ​​for the optimized parameter, and randomly select one value from all optional values ​​for the optimized parameter as the data for the corresponding gene position in the optimized process parameter combination.

[0023] Furthermore, a set of optimized process parameter combinations is taken as an individual, and the process parameters in the combination are taken as genes. The specific logic is as follows: all the data of the process parameters in the combination are encoded as corresponding genes, and the target parameters of the same type are alleles to each other, that is, there are three genes corresponding to a set of optimized process parameter combinations, including fiber volume ratio, resin viscosity and mold average temperature. All optimized process parameter combinations are encoded to obtain several individuals; an initial population is constructed based on all the obtained individuals, and the initial population is calibrated as ,and , represents the index of different individuals in the initial population, and u=1,2, , D, each individual There are 3 genes in each group, which correspond to a parameter value of fiber volume ratio, resin viscosity and average mold temperature, respectively, where D is the total number of individuals in the initial population.

[0024] Furthermore, the specific logic for obtaining the optimal target parameter combination through the genetic algorithm is: cyclically performing selection, crossover and mutation operations on the initial population to generate an iterative population containing multiple new individuals, and judging whether the maximum number of iterations has been reached. When the maximum number of iterations is greater than the maximum number of iterations, the individual with the highest fitness function value in the current population is selected as the optimal optimization parameter combination; otherwise, the generated iterative population is used as the initial population for iterative operations until the iteration termination condition is met, and the optimization process parameter combination with the largest fitness function value that appears during the iteration process is selected as the optimal target parameter combination, wherein the iteration termination condition is the set maximum number of iterations.

[0025] Furthermore, a fitness function is established according to the predicted values ​​of the mechanical characteristic parameters output by the mechanical characteristic prediction model, wherein the fitness function is calculated based on the formula:

[0026] ;

[0027] In the formula, represents the fitness function value of the individual, , and They are respectively the predicted values ​​of tensile strength, bending strength and interlaminar shear strength. , and They are the tensile strength target value, bending strength target value and interlaminar shear strength target value. , and are the weight coefficients of tensile strength, flexural strength and interlaminar shear strength, respectively, where and , and Both are greater than 0.

[0028] Furthermore, the curing dynamic factor and the curing reaction factor are calculated based on the obtained process flow parameters, wherein the curing dynamic factor is calculated based on the formula:

[0029] ;

[0030] In the formula, is the curing dynamic factor, is the average length of the fibers, is the angle between the fiber and the resin flow direction, is the injection flow rate of the resin, is the ambient humidity;

[0031] The formula for calculating the curing reaction factor is:

[0032] ;

[0033] In the formula, is the curing reaction factor, is the curing time, is the activation energy of the resin, is the gas constant, is the optimal average mold temperature within the optimal process parameter combination;

[0034] According to the obtained curing dynamic factor and curing reaction factor, the parameters in the optimal process parameter combination are corrected to obtain the precise value of the process parameter. The specific correction formula is as follows:

[0035] ;

[0036] ;

[0037] ;

[0038] In the formula, is the exact value of the fiber volume ratio, is the exact value of resin viscosity, is the precise value of the average mold temperature, and are the optimal fiber volume ratio and the optimal resin viscosity within the optimal process parameter combination, and are the weight coefficients of the curing dynamic factor and the curing reaction factor, respectively, where and and Both are greater than 0.

[0039] The present invention also provides a parameter optimization device for a composite material forming process, wherein the parameter optimization device for a composite material forming process is used to execute the above-mentioned parameter optimization method for a composite material forming process, comprising:

[0040] A training parameter acquisition module is used to obtain a number of composite material products with known mechanical characteristic parameters, collect the corresponding process parameters in the manufacturing process of the composite material products, map the obtained process parameters with the mechanical characteristic parameters of the corresponding composite material products one by one, and generate a training sample data set, wherein the mechanical characteristic parameters include tensile strength, bending strength and interlaminar shear strength, and the process parameters include fiber volume ratio, resin viscosity and average mold temperature;

[0041] The prediction model training module is used to establish a neural prediction network based on the data in the training sample data set, use the process parameters in the training sample data set as the input of the model, and use the corresponding mechanical characteristic parameters as labels to train the neural prediction network to obtain a mechanical characteristic prediction model;

[0042] A mechanical characteristic prediction module is used to obtain the value range of each process parameter, randomly generate several groups of optimized process parameter combinations within the value range, input different optimized process parameter combinations into the trained mechanical characteristic prediction model, obtain the mechanical characteristic parameter prediction value corresponding to the combination, and establish a fitness function according to the mechanical characteristic parameter prediction value. The optimized process parameter combination includes three gene positions, which correspond to the fiber volume ratio, resin viscosity and mold average temperature respectively.

[0043] An optimal parameter determination module is used to take a group of optimized process parameter combinations as an individual, and the process parameters in the combination as genes, and obtain an optimal process parameter combination through a genetic algorithm based on a fitness function, wherein the optimal process parameter combination is an optimized process parameter combination with the highest fitness function value;

[0044] The precise parameter correction module is used to obtain process flow parameters, calculate the curing kinetic factor and fiber distribution factor based on the obtained process flow parameters, and correct the parameters in the optimal process parameter combination according to the obtained curing kinetic factor and curing reaction factor to obtain the precise values ​​of the process parameters, thereby completing the parameter optimization of the composite material molding process. The process flow parameters include the average length of the mold fiber, the angle of the fiber relative to the resin flow direction, the injection flow rate of the resin, the curing time and the ambient humidity.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] Based on data-driven, this solution can effectively handle complex nonlinear relationships by establishing a neural network prediction model, thereby achieving accurate prediction of the molding process. This method no longer relies on tedious experiments and experience, but uses existing finished product data for deep learning, which greatly shortens the optimization cycle and cost and improves production efficiency. At the same time, the optimal process parameter combination is found through genetic algorithms, which effectively improves the globality of the optimization, avoids the risk of falling into a local optimal solution, and makes the optimization results more reliable. In addition, the combination of process flow parameters with curing dynamic factors and fiber distribution factors significantly enhances the adaptability of this method in practical applications. By accurately correcting the process parameters, the curing process of the composite material can be better controlled to ensure that the final product reaches the best state in terms of mechanical properties. In the end, the optimized composite material molding process not only improves the comprehensive performance of the product, but also reduces material waste in the production process, achieving the goal of sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic diagram of the overall method flow of the present invention;

[0048] Figure 2 It is a schematic diagram of the overall structure of the device of the present invention. DETAILED DESCRIPTION

[0049] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0050] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0051] Example:

[0052] See also Figure 1 , the present invention provides a technical solution:

[0053] A parameter optimization method for a composite material molding process, the specific steps comprising:

[0054] Step 1: Obtain several finished composite materials with known mechanical characteristic parameters, collect the corresponding process parameters in the manufacturing process of the finished composite materials, map the obtained process parameters with the corresponding mechanical characteristic parameters of the finished composite materials one by one, and generate a training sample data set. The mechanical characteristic parameters include tensile strength, bending strength and interlaminar shear strength, and the process parameters include fiber volume ratio, resin viscosity and average mold temperature.

[0055] The collected process parameters are preliminarily sorted out, and abnormal data are eliminated. The abnormal data specifically refers to process parameters that exceed a reasonable range. The training sample set is generated by mapping the process parameters to the mechanical characteristic parameters of the corresponding composite material products one by one to form a corresponding grid, and the formed grid is recorded as a training sample data set; the training sample data set is stored in a file to facilitate subsequent model calls. The storage format can select a common format (such as CSV, JSON or SQL database) to ensure easy reading and operation. The resin viscosity is the initial viscosity of the resin during the processing process.

[0056] The tensile strength is mainly carried by the fiber, so the higher the fiber volume ratio, the greater the tensile strength of the material is usually. When the fiber volume ratio is too low, the resin matrix load ratio increases, resulting in a decrease in tensile strength. When the fiber volume ratio is too high, insufficient gaps between fibers may lead to incomplete resin impregnation, causing defects (such as pores or delamination), thereby reducing tensile strength. Flexural strength also depends on the fiber content and distribution. A high fiber volume ratio usually helps to improve the flexural strength of the composite material, but too high a fiber volume ratio may also lead to poor interface bonding or fiber brittle failure. A higher fiber volume ratio may lead to a decrease in interface performance, thereby reducing interlaminar shear strength, but an appropriate fiber content helps to improve the overall rigidity of the composite structure. The fiber volume ratio specifically refers to the volume ratio between the added fiber filler and the matrix.

[0057] Resin viscosity affects the fluidity of the resin between fibers and the uniformity of impregnation. Too high a resin viscosity may lead to incomplete fiber impregnation, resulting in pores or delamination defects, thereby reducing tensile strength. If the resin viscosity is too low, it is easy to flow or accumulate unevenly, resulting in insufficient resin or fiber enrichment in local areas, affecting strength. Flexural strength depends on the integrity of the matrix and the bonding performance of the fiber-resin interface. When the resin viscosity is appropriate, the interface bonding performance is better, thereby improving the flexural strength.

[0058] The interlaminar shear strength is closely related to the interfacial bonding and the quality of the interlaminar resin. The resin viscosity affects the distribution and porosity of the matrix before curing, which in turn affects the interlaminar shear performance of the material.

[0059] The average mold temperature affects the fluidity and impregnation properties of the resin. A moderately high average mold temperature helps reduce resin viscosity and improve the uniformity of fiber impregnation, thereby improving tensile strength. The average mold temperature regulates the curing rate and residual stress, which in turn affects the bending strength. Too fast or too slow a curing rate will lead to uneven stress distribution and reduce bending performance. The average mold temperature affects the uniformity of resin curing and the interfacial adhesion properties. Too low a temperature may lead to insufficient curing, while too high a temperature may cause interfacial failure or material decomposition, thereby reducing interlaminar shear strength.

[0060] Step 2: Based on the data in the training sample data set, a neural prediction network is established. The process parameters in the training sample data set are used as the input of the model, and the corresponding mechanical characteristic parameters are used as labels to train the neural prediction network to obtain a mechanical characteristic prediction model.

[0061] Based on the data in the training sample set, a neural network prediction model is established, in which the long short-term memory network model LSTM model is selected as the base model, and the activation function and optimization algorithm are selected. The Tanh function is selected as the activation function, and Adam is selected as the optimization algorithm of the LSTM model; the formula of the Tanh function is:

[0062] ;

[0063] In the formula, Represents the Tanh function, independent variable Represents the weighted sum of the neuron's input, that is, the result of weighted summation of the input received by the neuron from the previous layer;

[0064] At the same time, the hyperparameters of the LSTM model are set, and the hyperparameters of the LSTM model include: the number of network layers, the number of iterations, the learning rate, the batch size, the number of training times, the batch processing number, and the number of hidden layer neurons;

[0065] The number of network layers is set to 4 layers, the number of iterations is set to 300, the learning rate is set to 0.001, the batch size is set to 32, the number of training times is set to 100, the batch size is set to 256, and the number of hidden layer neurons is 32;

[0066] The trained mechanical characteristic prediction model takes process parameter data as input and outputs mechanical characteristic parameter prediction values, which include tensile strength prediction values, bending strength prediction values ​​and interlaminar shear strength prediction values.

[0067] Step 3: Obtain the value range of each process parameter, randomly generate several groups of optimized process parameter combinations within the value range, input different optimized process parameter combinations into the trained mechanical characteristic prediction model, obtain the mechanical characteristic parameter prediction value corresponding to the combination, and establish a fitness function according to the mechanical characteristic parameter prediction value. The optimized process parameter combination includes three gene positions, which correspond to the fiber volume ratio, resin viscosity and average mold temperature respectively.

[0068] The logic for randomly generating several groups of optimized process parameter combinations within a value range is as follows: randomly generate several different data within the value range of each optimized process parameter as optional values ​​of the optimized parameter, and randomly select one value from all optional values ​​of the optimized parameter as the data of the corresponding gene position in the optimized process parameter combination.

[0069] The method for obtaining the range of each process parameter includes: determining the typical range of process parameters by consulting theoretical research, standard specifications and literature in related fields, and referring to industry standards or experimental specifications (such as ASTM, ISO, GB standards, etc.) for composite material molding processes. For example: fiber volume ratio: for most composite materials, the industry recommended range is usually 0.4 to 0.7; resin viscosity: when using RTM (resin transfer molding) process, the recommended range is 50 to 1000 mPa·s; mold average temperature: thermosetting resin is usually 80 Up to 150 Or look for experimental studies related to the materials and processes being studied, and obtain the range of process parameters used by other researchers.

[0070] Step 4: Take a group of optimized process parameter combinations as an individual, and the process parameters in the combination as genes. Based on the fitness function, the optimal process parameter combination is obtained through a genetic algorithm, wherein the optimal process parameter combination is the optimized process parameter combination with the highest fitness function value.

[0071] A set of optimized process parameter combinations is regarded as an individual, and the process parameters in the combination are regarded as genes. The specific logic is as follows: all the process parameter data in the combination are encoded as corresponding genes, and the target parameters of the same type are alleles to each other, that is, there are three genes corresponding to a set of optimized process parameter combinations, including fiber volume ratio, resin viscosity and mold average temperature. All optimized process parameter combinations are encoded to obtain several individuals; an initial population is constructed based on all the obtained individuals, and the initial population is calibrated as ,and , represents the index of different individuals in the initial population, and u=1,2, , D, each individual There are 3 genes in each group, which correspond to a parameter value of fiber volume ratio, resin viscosity and average mold temperature, respectively, where D is the total number of individuals in the initial population.

[0072] The specific logic for obtaining the optimal target parameter combination through the genetic algorithm is as follows: the initial population is cyclically selected, crossed and mutated to generate an iterative population containing multiple new individuals, and it is determined whether the maximum number of iterations has been reached. When the maximum number of iterations is greater than the maximum number of iterations, the individual with the highest fitness function value in the current population is selected as the optimal optimization parameter combination. Otherwise, the generated iterative population is used as the initial population for iterative operations until the iteration termination condition is met. The optimization process parameter combination with the largest fitness function value that appears during the iteration process is selected as the optimal target parameter combination, wherein the iteration termination condition is the set maximum number of iterations, wherein the maximum number of iterations is generally set between 200 and 500.

[0073] The fitness function is established according to the predicted values ​​of the mechanical characteristic parameters output by the mechanical characteristic prediction model, wherein the specific formula for calculating the fitness function is as follows:

[0074] ;

[0075] In the formula, represents the fitness function value of the individual, , and They are respectively the predicted values ​​of tensile strength, bending strength and interlaminar shear strength. , and They are the tensile strength target value, bending strength target value and interlaminar shear strength target value. , and are the weight coefficients of tensile strength, flexural strength and interlaminar shear strength, respectively, where and , and Both are greater than 0.

[0076] Among them, it should be noted that It is the fitness function of the individual. The larger its value is, the greater the fitness of the individual is. The smaller the error of tensile strength, bending strength and interlaminar shear strength with the target finished product is, the higher the quality of the finished product is.

[0077] Among them, the individual fitness function and , and Inversely proportional to the error between the target value and the predicted value, the larger the error, the higher the fitness function of the individual. The smaller it is, the lower the quality of the finished product.

[0078] Among them, the tensile strength target value, bending strength target value and interlaminar shear strength target value can be set according to actual production needs.

[0079] In many composite applications, tensile strength is usually the most critical performance indicator because it directly affects the material's ability to bear loads under tensile loads. Flexural strength is also important, but in some applications, its importance may be secondary to tensile strength. For example, in structural parts, tensile stress is often the main consideration. Interlaminar shear strength is usually more important in certain specific applications, but in many cases, its importance is not as high as tensile and flexural strength, so it is necessary to set and , and Both are greater than 0.

[0080] The initial population is cycled through selection, crossover and mutation operations, where the logic of the selection operation is: select individuals in the population according to the fitness in the sequence, and select two individuals from the current sequence. The formula for calculating the probability of being selected is:

[0081] ;

[0082] In the formula, represents the first individual in the initial population The fitness of each individual, For the The probability of an individual being selected is ,and is a positive integer, and the roulette method is used to select individuals. The system randomly generates a A random number in the interval is distributed in the selection interval of the corresponding individual to determine the individual to be selected in this round according to the generated probability of being selected;

[0083] The logic of performing the crossover operation is: performing the crossover operation on two selected individuals, exchanging the two selected individuals from a certain gene position and the chromosome fragment after the gene position, and obtaining two crossed individuals, wherein the crossover operation is performed on the two selected individuals, and the logic based on which the crossover operation is performed is: using a single-point crossover method to perform the crossover operation.

[0084] The logic of mutation operation is: the mutation operation is performed on the two new individuals generated by the crossover operation, and the mutation probability is set. The selected gene is mutated into one of its alleles, and the gene at the gene position is replaced with the mutated gene, thereby obtaining a new individual.

[0085] Step 5: Obtain process flow parameters, calculate the curing kinetic factor and fiber distribution factor based on the obtained process flow parameters, correct the parameters in the optimal process parameter combination according to the obtained curing kinetic factor and curing reaction factor, obtain the precise values ​​of the process parameters, and complete the parameter optimization of the composite material molding process. The process flow parameters include the average length of the fiber, the angle of the fiber relative to the resin flow direction, the injection flow rate of the resin, the curing time and the ambient humidity.

[0086] The curing dynamic factor and curing reaction factor are calculated based on the obtained process parameters, wherein the curing dynamic factor is calculated based on the formula:

[0087] ;

[0088] In the formula, is the curing dynamic factor, is the average length of the fibers, is the angle between the fiber and the resin flow direction, is the injection flow rate of the resin, is the ambient humidity.

[0089] It should be noted that the curing dynamic factor Used to describe the uniformity of fiber distribution in the mold and the degree of resistance to resin flow, curing dynamic factor The larger the value, the worse the uniformity of fiber distribution in the mold and the greater the degree of hindrance to resin flow.

[0090] The average length of the fibers Affects the wetting and distribution between fiber and resin. Longer fibers will make it more difficult for resin to penetrate, increase flow resistance, and reduce the uniformity of fiber distribution. Therefore, the average length of the fiber Solidification dynamic factor The average length range or nominal length of the fiber is usually given in the technical data sheet.

[0091] The angle of the fiber relative to the resin flow direction Affects the directionality of resin flow. If the angle between the fiber and the flow direction is smaller (i.e. the fiber arrangement is closer to the resin flow direction), the resin will flow more smoothly, making it easier to cover the fiber and promote the uniformity of the curing process. On the contrary, the larger the angle, the greater the resistance and the worse the fluidity. Therefore, the curing dynamic factor It is proportional to the angle between the fiber and the resin flow direction. Represents an inverse relationship.

[0092] Use resin flow modeling software (such as Moldflow, RTM-Worx) to simulate the interaction between resin and fiber during injection. The software can generate fiber direction field and flow direction field, and obtain the angle between these direction fields by calculating the angle between them. distribution.

[0093] Resin injection flow rate , directly affects the fluidity of the resin and the filling process, the injection flow rate of the resin The larger the size, the stronger the fluidity and the more uniform the distribution, so the curing dynamic factor The injection flow rate of the resin Inversely proportional.

[0094] Ambient humidity affects the curing rate of the resin. Higher humidity increases the moisture content of the resin, which reduces the viscosity of the resin, increases fluidity, and makes the distribution more uniform. Therefore, the curing dynamic factor and ambient humidity Inversely proportional.

[0095] The formula for calculating the curing reaction factor is:

[0096] ;

[0097] In the formula, is the curing reaction factor, is the curing time, is the activation energy of the resin, is the gas constant, is the optimal average mold temperature within the optimal process parameter combination.

[0098] It should be noted that the curing reaction factor It is used to describe the kinetics of the resin curing reaction, taking into account the combined effects of curing time, initial activation energy and external temperature. The larger the value, the higher the degree of curing and the better the curing effect.

[0099] Chemical reactions have a time accumulation effect. The longer the reaction time, the more complete the curing reaction and the higher the degree of reaction. Therefore, the curing time Curing reaction factor It is directly proportional to the curing time and is used to describe the curing effect. The curing time is the reaction time set by experience.

[0100] activation energy Located in the numerator of the exponential term, it represents the energy barrier required for the chemical reaction. The larger the value, the higher the energy barrier that the reaction needs to overcome, the more difficult the reaction is, and the worse the curing effect is. Therefore, the activation energy of the resin is Curing reaction factor Inversely proportional, through the exponential function Represents a nonlinear inverse relationship. The resin type is a common commercial resin (such as epoxy resin, phenolic resin, unsaturated polyester resin, etc.), and its activation energy can usually be obtained from the following ways: Many resin suppliers will provide curing kinetics-related data in TDS or MSDS, including activation energy; consult academic articles related to resins (through CNKI, ScienceDirect, SpringerLink and other databases) to obtain activation energy values ​​of similar resin systems.

[0101] Gas constant It is a physical constant that ensures the consistency of formula units and reflects the thermodynamic properties of the reaction process. It describes the relationship between energy and temperature. Gas constant Generally take .

[0102] The reaction rate increases with increasing temperature. This is because higher temperatures provide more molecular kinetic energy, allowing more molecules to overcome the activation energy barrier, thereby accelerating the reaction process. Therefore, the optimal average mold temperature Curing reaction factor Inversely proportional, through the exponential function Represents a nonlinear inverse relationship.

[0103] According to the obtained curing dynamic factor and curing reaction factor, the parameters in the optimal process parameter combination are corrected to obtain the precise value of the process parameter. The specific correction formula is as follows:

[0104] ;

[0105] ;

[0106] ;

[0107] In the formula, is the exact value of the fiber volume ratio, is the exact value of resin viscosity, is the precise value of the average mold temperature, and are the optimal fiber volume ratio and the optimal resin viscosity within the optimal process parameter combination, and are the weight coefficients of the curing dynamic factor and the curing reaction factor, respectively, where and and Both are greater than 0.

[0108] Among them, due to the curing dynamic factor The larger the value, the worse the uniformity of fiber distribution in the mold, and the greater the degree of obstruction to resin flow. Therefore, the fiber volume ratio needs to be appropriately reduced to improve resin fluidity and make the distribution more uniform. Therefore, the exact value of fiber volume ratio Solidification dynamic factor Inversely proportional, through the square root Indicates when the curing dynamic factor When it is large enough, it will have a significant impact on the fiber volume ratio, otherwise the impact will be small.

[0109] Due to the curing reaction factor The larger the value, the higher the degree of curing and the better the curing effect. Therefore, the resin viscosity can be appropriately reduced to save raw materials and achieve the target curing effect. The curing dynamic factor The larger the value, the worse the uniformity of fiber distribution in the mold, and the greater the degree of obstruction to resin flow. Therefore, the resin viscosity should be reduced to improve fluidity. Therefore, the exact value of resin viscosity is Solidification dynamic factor and curing reaction factor are inversely proportional, through the logarithmic function Indicates that as the curing reaction factor The increase in resin viscosity The influence of Narrow down the exact value of resin viscosity impact.

[0110] Curing reaction factor The larger the value, the higher the degree of curing and the better the curing effect. It also indicates that the curing reaction requires more heat and the average mold temperature may need to be increased. Therefore, the accurate value of the average mold temperature is Curing reaction factor Proportional, through the square Indicates its significant impact, solidification dynamic factor The larger the value, the worse the uniformity of fiber distribution in the mold, the greater the degree of obstruction to resin flow, and the smaller the value, the better the fiber distribution, and the average mold temperature can be slightly reduced to save energy. The exact value of the mold average temperature Proportional, expressed by a logarithmic function Indicates that as the curing dynamic factor The increase of the average temperature of the mold The impact gradually decreases.

[0111] Among them, the curing reaction factor is the core parameter of the resin curing process, which directly affects the chemical reaction degree of the resin, the crosslinking density and the mechanical properties of the final product (such as strength, durability, etc.). In the composite material molding process, chemical reaction control is the main contradiction. The curing dynamic factor is mainly related to the uniformity of fiber distribution and resin fluidity. Although its role is important, its priority is lower than the effect of the curing reaction on material properties during composite material molding. Therefore, the influence weight of the curing reaction factor is greater than the influence of fiber distribution, so it is set and and Both are greater than 0.

[0112] Among them, the logic for correcting the parameters within the optimal process parameter combination is: since the quality evaluation standard in the engineering processing process is generally a range standard, that is, all products within this range standard are qualified, therefore, by slightly reducing the mechanical characteristics of the product and further correcting the parameters within the optimal process parameter combination through the process flow parameters, the processing flow is smoother and the processing efficiency is further improved.

[0113] See also Figure 2 The present invention also provides a parameter optimization device for a composite material forming process, wherein the parameter optimization device for a composite material forming process is used to execute the above-mentioned parameter optimization method for a composite material forming process, comprising:

[0114] A training parameter acquisition module is used to obtain a number of composite material products with known mechanical characteristic parameters, collect the corresponding process parameters in the manufacturing process of the composite material products, map the obtained process parameters with the mechanical characteristic parameters of the corresponding composite material products one by one, and generate a training sample data set, wherein the mechanical characteristic parameters include tensile strength, bending strength and interlaminar shear strength, and the process parameters include fiber volume ratio, resin viscosity and average mold temperature;

[0115] The prediction model training module is used to establish a neural prediction network based on the data in the training sample data set, use the process parameters in the training sample data set as the input of the model, and use the corresponding mechanical characteristic parameters as labels to train the neural prediction network to obtain a mechanical characteristic prediction model;

[0116] A mechanical characteristic prediction module is used to obtain the value range of each process parameter, randomly generate several groups of optimized process parameter combinations within the value range, input different optimized process parameter combinations into the trained mechanical characteristic prediction model, obtain the mechanical characteristic parameter prediction value corresponding to the combination, and establish a fitness function according to the mechanical characteristic parameter prediction value. The optimized process parameter combination includes three gene positions, which correspond to the fiber volume ratio, resin viscosity and mold average temperature respectively.

[0117] An optimal parameter determination module is used to take a group of optimized process parameter combinations as an individual, and the process parameters in the combination as genes, and obtain an optimal process parameter combination through a genetic algorithm based on a fitness function, wherein the optimal process parameter combination is an optimized process parameter combination with the highest fitness function value;

[0118] The precise parameter correction module is used to obtain process flow parameters, calculate the curing kinetic factor and fiber distribution factor based on the obtained process flow parameters, and correct the parameters in the optimal process parameter combination according to the obtained curing kinetic factor and curing reaction factor to obtain the precise values ​​of the process parameters, thereby completing the parameter optimization of the composite material molding process. The process flow parameters include the average length of the mold fiber, the angle of the fiber relative to the resin flow direction, the injection flow rate of the resin, the curing time and the ambient humidity.

[0119] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0120] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0121] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0122] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A parameter optimization method for a composite material molding process, characterized in that: The specific steps include: Obtaining several composite product products with known mechanical characteristic parameters, collecting corresponding process parameters in the manufacturing process of the composite product products, mapping the obtained process parameters with the corresponding mechanical characteristic parameters of the composite product products one by one, and generating a training sample data set, wherein the mechanical characteristic parameters include tensile strength, bending strength and interlaminar shear strength, and the process parameters include fiber volume ratio, resin viscosity and average mold temperature; Based on the data in the training sample data set, a neural prediction network is established, the process parameters in the training sample data set are used as the input of the model, and the corresponding mechanical characteristic parameters are used as labels to train the neural prediction network to obtain a mechanical characteristic prediction model; Obtaining the value range of each process parameter, randomly generating several groups of optimized process parameter combinations within the value range, inputting different optimized process parameter combinations into the trained mechanical characteristic prediction model, obtaining the mechanical characteristic parameter prediction value corresponding to the combination, and establishing a fitness function according to the mechanical characteristic parameter prediction value, wherein the optimized process parameter combination includes three gene positions corresponding to the fiber volume ratio, the resin viscosity and the average mold temperature respectively; A set of optimized process parameter combinations is taken as an individual, and the process parameters in the combination are taken as genes. Based on the fitness function, an optimal process parameter combination is obtained by a genetic algorithm, wherein the optimal process parameter combination is an optimized process parameter combination with the highest fitness function value; The process parameters are obtained, and the curing kinetic factor and the fiber distribution factor are calculated based on the obtained process parameters. The parameters in the optimal process parameter combination are corrected according to the obtained curing kinetic factor and the curing reaction factor to obtain the precise values ​​of the process parameters, thereby completing the parameter optimization of the composite material molding process. The process parameters include the average length of the fiber, the angle of the fiber relative to the resin flow direction, the injection flow rate of the resin, the curing time and the ambient humidity.

2. The parameter optimization method for composite material forming process according to claim 1, characterized in that: The method for generating the training sample data set is as follows: mapping the process parameters to the mechanical characteristic parameters of the corresponding composite material finished product one by one to form a corresponding grid, and recording the formed grid as the training sample data set; Based on the data in the training sample set, a neural network prediction model is established, in which the long short-term memory network model LSTM model is selected as the base model, and the activation function and optimization algorithm are selected. The Tanh function is selected as the activation function, and Adam is selected as the optimization algorithm of the LSTM model; the formula of the Tanh function is: ； In the formula, Represents the Tanh function, independent variable Represents the weighted sum of the neuron's input, that is, the result of weighted summation of the input received by the neuron from the previous layer; At the same time, the hyperparameters of the LSTM model are set, and the hyperparameters of the LSTM model include: the number of network layers, the number of iterations, the learning rate, the batch size, the number of training times, the batch processing number, and the number of hidden layer neurons; The number of network layers is set to 4 layers, the number of iterations is set to 300, the learning rate is set to 0.001, the batch size is set to 32, the number of training times is set to 100, the batch size is set to 256, and the number of hidden layer neurons is 32; The trained mechanical characteristic prediction model takes process parameter data as input and outputs mechanical characteristic parameter prediction values, which include tensile strength prediction values, bending strength prediction values ​​and interlaminar shear strength prediction values.

3. The parameter optimization method for composite material forming process according to claim 2, characterized in that: The logic for randomly generating several groups of optimized process parameter combinations within a value range is as follows: randomly generate several different data within the value range of each optimized process parameter as optional values ​​of the optimized parameter, and randomly select one value from all optional values ​​of the optimized parameter as the data of the corresponding gene position in the optimized process parameter combination.

4. The parameter optimization method for composite material forming process according to claim 3, characterized in that: A set of optimized process parameter combinations is regarded as an individual, and the process parameters in the combination are regarded as genes. The specific logic is as follows: all the process parameter data in the combination are encoded as corresponding genes, and the target parameters of the same type are alleles to each other, that is, there are three genes corresponding to a set of optimized process parameter combinations, including fiber volume ratio, resin viscosity and mold average temperature. All optimized process parameter combinations are encoded to obtain several individuals; an initial population is constructed based on all the obtained individuals, and the initial population is calibrated as ,and , represents the index of different individuals in the initial population, and u=1,2, , D, each individual There are 3 genes in each group, which correspond to a parameter value of fiber volume ratio, resin viscosity and average mold temperature, respectively, where D is the total number of individuals in the initial population.

5. The parameter optimization method for composite material forming process according to claim 4, characterized in that: The specific logic for obtaining the optimal target parameter combination through the genetic algorithm is as follows: cyclically perform selection, crossover and mutation operations on the initial population to generate an iterative population containing multiple new individuals, determine whether the maximum number of iterations has been reached, and when it is greater than the maximum number of iterations, select the individual with the highest fitness function value in the current population as the optimal optimization parameter combination; otherwise, use the generated iterative population as the initial population for iterative operations until the iteration termination condition is met, and select the optimization process parameter combination with the largest fitness value that appears during the iteration process as the optimal target parameter combination, wherein the iteration termination condition is the set maximum number of iterations.

6. The parameter optimization method for composite material forming process according to claim 4, characterized in that: The fitness function is established according to the predicted values ​​of the mechanical characteristic parameters output by the mechanical characteristic prediction model, wherein the specific formula for calculating the fitness function is as follows: ； In the formula, represents the fitness function value of the individual, , and They are respectively the predicted values ​​of tensile strength, bending strength and interlaminar shear strength. , and They are the tensile strength target value, bending strength target value and interlaminar shear strength target value. , and are the weight coefficients of tensile strength, flexural strength and interlaminar shear strength, respectively, where and , and Both are greater than 0.

7. The parameter optimization method for composite material molding process according to claim 1, characterized in that: The curing dynamic factor and curing reaction factor are calculated based on the obtained process parameters, wherein the curing dynamic factor is calculated based on the formula: ； In the formula, is the curing dynamic factor, is the average length of the fibers, is the angle between the fiber and the resin flow direction, is the injection flow rate of the resin, is the ambient humidity; The formula for calculating the curing reaction factor is: ； In the formula, is the curing reaction factor, is the curing time, is the activation energy of the resin, is the gas constant, is the optimal average mold temperature within the optimal process parameter combination; According to the obtained curing dynamic factor and curing reaction factor, the parameters in the optimal process parameter combination are corrected to obtain the precise value of the process parameter. The specific correction formula is as follows: ； ； ； In the formula, is the exact value of the fiber volume ratio, is the exact value of resin viscosity, is the precise value of the average mold temperature, and are the optimal fiber volume ratio and the optimal resin viscosity within the optimal process parameter combination, and are the weight coefficients of the curing dynamic factor and the curing reaction factor, respectively, where and and Both are greater than 0.

8. A parameter optimization device for composite material molding process, characterized in that: The parameter optimization device for a composite material forming process is used to execute the parameter optimization method for a composite material forming process according to any one of claims 1 to 7, comprising: A training parameter acquisition module is used to obtain a number of composite material products with known mechanical characteristic parameters, collect the corresponding process parameters in the manufacturing process of the composite material products, map the obtained process parameters with the mechanical characteristic parameters of the corresponding composite material products one by one, and generate a training sample data set, wherein the mechanical characteristic parameters include tensile strength, bending strength and interlaminar shear strength, and the process parameters include fiber volume ratio, resin viscosity and average mold temperature; The prediction model training module is used to establish a neural prediction network based on the data in the training sample data set, use the process parameters in the training sample data set as the input of the model, and use the corresponding mechanical characteristic parameters as labels to train the neural prediction network to obtain a mechanical characteristic prediction model; A mechanical characteristic prediction module is used to obtain the value range of each process parameter, randomly generate several groups of optimized process parameter combinations within the value range, input different optimized process parameter combinations into the trained mechanical characteristic prediction model, obtain the mechanical characteristic parameter prediction value corresponding to the combination, and establish a fitness function according to the mechanical characteristic parameter prediction value. The optimized process parameter combination includes three gene positions, which correspond to the fiber volume ratio, resin viscosity and mold average temperature respectively. An optimal parameter determination module is used to take a group of optimized process parameter combinations as an individual, and the process parameters in the combination as genes, and obtain an optimal process parameter combination through a genetic algorithm based on a fitness function, wherein the optimal process parameter combination is an optimized process parameter combination with the highest fitness function value; The precise parameter correction module is used to obtain process flow parameters, calculate the curing kinetic factor and fiber distribution factor based on the obtained process flow parameters, and correct the parameters in the optimal process parameter combination according to the obtained curing kinetic factor and curing reaction factor to obtain the precise values ​​of the process parameters, thereby completing the parameter optimization of the composite material molding process. The process flow parameters include the average length of the mold fiber, the angle of the fiber relative to the resin flow direction, the injection flow rate of the resin, the curing time and the ambient humidity.

Images