Carbon fiber reinforced polyether-ether-ketone composite material design method and system based on machine learning guidance
By constructing a CF/PEEK composite material dataset through machine learning and optimizing parameters using the frog-leap algorithm, the problem of time-consuming and costly traditional design methods is solved, enabling fast and efficient composite material design that meets the performance requirements of diverse application scenarios.
Patent Information
- Application Number
- CN202410767776.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-14
- Publication Date
- 2025-12-16
AI Technical Summary
Traditional composite material design methods are time-consuming and resource-intensive, making it difficult to quickly meet the performance requirements of diverse application scenarios. In particular, in carbon fiber reinforced polyether ether ketone composites, complex parameter adjustments lead to experimental waste and unstable material properties.
A parameter-performance dataset for CF/PEEK composite materials was constructed using machine learning methods. A data model was established by selecting feature variables, and the configuration and process parameters were optimized using the frog leaping algorithm to achieve rapid design of materials that meet performance requirements.
With machine learning guidance, high-performance composite materials can be designed in a time-saving and labor-saving manner, saving costs, adapting to multi-objective optimization needs, and improving the efficiency of materials research.
Smart Images
Figure CN121148540A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of composite material design. BACKGROUND
[0002] (1) With the development of modern science and technology and the increasing demand of society, along with the rise of modern industry, the application scenarios of materials are becoming more and more diversified and are expanding to extreme environments. Traditional engineering materials have their advantages, but it is difficult to meet the needs of diversified application scenarios by relying on a single material. Single material is difficult to be compatible with multiple excellent properties, and due to its limitations, it will cause the service life of the material to be reduced, thereby causing unnecessary waste and even environmental pollution.
[0003] Composite materials are a kind of artificial materials composed of several materials with different properties in a certain proportion. As the name implies, composite materials are designed by different combinations and different composite methods of various materials. It can not only play the advantages of each component material, but also derive new comprehensive properties that cannot be achieved by each component material. This makes the existing material types increase exponentially, and the application field is greatly widened. Therefore, the compounding of materials is an inevitable trend to promote social progress and meet the needs of various fields, and it is necessary to study the composition and process of composite materials.
[0004] Composite materials include matrix and reinforcement. Carbon fiber (CF) is a common reinforcement, and the thermoplastic composite material reinforced by it has the characteristics of high strength and light weight. It has been applied since the 1960s, especially in the field of aerospace. With the continuous maturity of science and technology, carbon fiber reinforced resin matrix composite materials play an increasingly dominant role in aircraft load-bearing structural parts. Polyether ether ketone resin (PEEK) is a thermoplastic resin with excellent processability, which is easier to process than some thermosetting resins and is often used as the matrix of carbon fiber reinforced composite materials. The melting point of PEEK is 343℃, which is higher than that of general thermoplastic resins and can withstand harsh environments. PEEK also has corrosion resistance and can be used in aircraft ball bearings and other parts. Because PEEK is lighter than metal materials, it can effectively reduce the weight of the aircraft. PEEK has very stable chemical properties and better mechanical properties than most resin materials, so compared with other fiber reinforcements, the mechanical properties of CF / PEEK composite materials are the most excellent, and CF / PEEK composite materials have great potential in application.
[0005] (2) Because the composite material system is complex, the molding process is complex, and each parameter such as the type of material, the content of different components, the molding method, and the temperature and pressure in the molding process will affect the macro performance of the composite material, the research on the composite material involves multiple aspects. For example, with the increase of CF content, the mechanical properties of CF / PEEK composite material will be significantly improved, but too much CF content will cause that the CF fiber cannot be fully wrapped by PEEK resin, and the resin cannot fix the CF fiber, causing some defects in the composite material. For the molding process of the composite material, when the molding temperature is low, the resin cannot be fully melted, and when the molding temperature is too high, PEEK decomposition will occur, causing pores in the laminate. When the molding pressure is large, the resin will be squeezed out, and when the molding pressure is too small, the inside is not dense. At present, the research on these parameters still stays in the stage of changing a single variable and repeating the test, and there are already a large amount of test data, and the traditional test method not only consumes a long time, but also repeats a lot of unnecessary tests, causing material waste. In the existing research, there are many achievements on the influence of different composite material designs on mechanical properties, and in the big data era, it should be learned to use various existing resources and various open databases to improve the material research efficiency, and the material calculation design method emerges as the times require. The machine learning method can set a series of input and output parameters, extract information from the existing test data, process information, and output information according to different algorithms. Without test, the effect of completing the composite material design can be achieved, which saves time and labor, and the material parameters can be designed for the target performance, and repeated tests can be avoided. SUMMARY
[0006] The purpose of the present application is to provide a carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance, which can design material parameters meeting performance requirements according to engineering needs, so as to better enter industrial application.
[0007] To achieve the above purpose, the present application provides the following scheme:
[0008] The present application provides a carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance, wherein the method comprises the following steps:
[0009] S1, constructing a CF / PEEK composite material parameter-performance data set;
[0010] S2, screening the characteristic variables in the CF / PEEK composite material parameter-performance data set, and respectively establishing different CF / PEEK composite material parameter-performance data models according to the screened characteristic variables;
[0011] S3, selecting the optimal CF / PEEK composite material parameter-performance data model by comparing the determination coefficients;
[0012] S4, the optimal CF / PEEK composite parameter-performance data model is used to analyze the influence trend curve of each characteristic variable on the mechanical properties;
[0013] S5, the frog leap algorithm is used to optimize the configuration parameters and process parameters required for the mechanical properties.
[0014] Further, there is a preferred embodiment, and the above step S1 is specifically:
[0015] S11, a configuration parameter-process parameter-mechanical property data set is established according to the past test data;
[0016] S12, the material parameters in the data set are data cleaned and data arranged.
[0017] Further, there is a preferred embodiment, and the configuration parameters in the above configuration parameter-process parameter-mechanical property data set include carbon fiber volume fraction, carbon fiber mass fraction, carbon fiber diameter, number of filaments per strand of two-dimensional carbon fiber cloth and polyether ether ketone resin mesh number;
[0018] The above process parameters include melting temperature, molding pressure and holding time;
[0019] The above mechanical property parameters include tensile strength, Young's modulus, elongation, bending strength, bending modulus, interlaminar shear strength and interlaminar shear modulus.
[0020] Further, there is a preferred embodiment, and the above data arrangement formula is:
[0021]
[0022] Wherein, W CF is the CF mass fraction, V CF is the CF volume fraction, ρ CF is the CF density, ρ PEEK is the PEEK density.
[0023] Further, there is a preferred embodiment, and the above step S2 is specifically:
[0024] S21, the CF / PEEK composite parameter performance data set is analyzed by using partial least squares regression and mean square error calculation method, and 6 input parameters are determined;
[0025] S22, the importance of the parameters in the CF / PEEK composite parameter performance data set is calculated, and 6 input parameters and 3 output parameters are determined;
[0026] S23, according to the 6 input parameters, the corresponding model is established and trained by using Gaussian process regression, decision tree, support vector machine and integrated algorithm respectively.
[0027] Further, there is a preferred embodiment that the six input parameters are CF mass fraction, CF type, CF filament number, melting temperature, molding pressure and holding time.
[0028] The three output parameters are tensile strength, bending strength and interlaminar shear strength.
[0029] Further, there is a preferred embodiment that the coefficient of determination formula in step S3 is:
[0030]
[0031] wherein R 2 is the goodness of fit, SSR is the regression sum of squares, SSE is the residual sum of squares, SST is the total deviation split sum, y i is the true observed value, is the average of the true observation, is the fitted value.
[0032] The carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance provided by the application can be realized by computer software, and accordingly, the application also provides a carbon fiber reinforced polyether ether ketone composite material design system based on machine learning guidance, which comprises:
[0033] A storage device for constructing a CF / PEEK composite material parameter-performance data set;
[0034] A storage device for screening feature variables in the CF / PEEK composite material parameter-performance data set and establishing different CF / PEEK composite material parameter-performance data models according to the screened feature variables;
[0035] A storage device for selecting the optimal CF / PEEK composite material parameter-performance data model by comparing the coefficients of determination;
[0036] A storage device for analyzing the influence trend curve of each feature variable on the mechanical properties by using the optimal CF / PEEK composite material parameter-performance data model;
[0037] A storage device for optimizing the configuration parameters and process parameters required for the mechanical properties by using the frog leap algorithm.
[0038] The application also provides a computer readable storage medium, which stores a computer program, and the computer program is run by a processor to execute the carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance provided by any one of the above.
[0039] The application further provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance.
[0040] The application has the following beneficial effects:
[0041] 1. The carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance provided by the application builds the configuration-process-performance relationship of the carbon fiber polyether ether ketone composite material, wherein the configuration-process involves the CF volume fraction, the CF mass fraction, the CF diameter, the number of monofilaments of the two-dimensional CF cloth per strand, the PEEK mesh number, the melting temperature, the molding pressure and the holding time; and the output is the mechanical property parameter, involving the tensile strength, the Young's modulus, the elongation rate, the bending strength, the bending modulus, the interlaminar shear strength and the interlaminar shear modulus. The influence law and mechanism of each parameter on the performance of the CF / PEEK composite material are studied by using the optimal parameter-performance data model, thereby providing theoretical guidance for the preparation of high-performance composite materials. Compared with the traditional test method, the method is time-saving and labor-saving, and the cost is saved.
[0042] 2. The carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance provided by the application is convenient to transplant and configure, and can quickly design the composition and process parameters meeting the performance requirements. The multi-objective optimization can realize two performance requirements as the target, or three, four or five performance requirements, and can be randomly combined, thereby fully embodying the possibility and advantage of multi-objective optimization. The multi-objective optimization is not only high target, but also has more advantages in the preparation of specified strength and modulus composite materials.
[0043] The application is suitable for the design and optimization of composite materials. BRIEF DESCRIPTION OF DRAWINGS
[0044] Figure 1 It is a flowchart of the carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance according to the first embodiment;
[0045] Figure 2 It is a decision tree model according to the third embodiment;
[0046] Figure 3 It is a schematic diagram of the decision boundary and the margin according to the third embodiment;
[0047] Figure 4 It is a schematic diagram of the nonlinear separable feature space to the linear separable feature space according to the third embodiment;
[0048] Figure 5is the ensemble framework of ensemble learning according to embodiment three;
[0049] Figure 6 is the basic flowchart of Boosting method according to embodiment three.
[0050] wherein, Positive is positive class sample, Negative is negative class sample, Hyper-plane is decision boundary, Margin is margin, Input date is input data, Final Predictions is final prediction result. DETAILED DESCRIPTION
[0051] In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular architectures, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.
[0052] It is to be understood that the terminology "including", "comprising", "consisting" and "consisting essentially of" used in the specification and the appended claims, indicates the presence of the stated features, integers, steps, operations, elements, and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.
[0053] It is also to be understood that the terminology used in the present specification and the appended claims is for the purpose of describing the particular embodiments only and is not intended to be limiting. As used in this specification and the appended claims, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
[0054] The specific embodiments of the present application will be further described with reference to the drawings and embodiments below. The following examples will help those skilled in the art to further understand the present application, but do not limit the present application in any form. It should be noted that for those skilled in the art, without departing from the concept of the present application, a number of changes and improvements can be made, which are within the scope of protection of the present application.
[0055] Embodiment one, the embodiment provides a carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance, the design method comprises the following steps:
[0056] S1, constructing CF / PEEK composite material parameter-performance dataset;
[0057] S2, screen the characteristic variables in the CF / PEEK composite parameter-performance data set, and establish different CF / PEEK composite parameter-performance data models according to the screened characteristic variables;
[0058] S3, select the optimal CF / PEEK composite parameter-performance data model by comparing the determination coefficients;
[0059] S4, analyze the influence trend curve of each characteristic variable on the mechanical properties by using the optimal CF / PEEK composite parameter-performance data model;
[0060] S5, optimize the configuration parameters and process parameters required for the mechanical properties by using the frog leap algorithm.
[0061] In actual application, as shown in Figure 1 The carbon fiber reinforced polyether ether ketone composite design method based on machine learning guidance is specifically:
[0062] Step one, build a CF / PEEK composite parameter-performance data set; specifically, refer to CF fabric reinforced PEEK composite material literature and previous test results, set the input parameters as material parameters, and the material parameters include structural parameters and process parameters; among them, the structural parameters include CF volume fraction, CF mass fraction, CF diameter, number of filaments per strand of two-dimensional CF cloth and PEEK mesh number. Process parameters include melting temperature, molding pressure and holding time. The output is the mechanical property parameter, which includes tensile strength, Young's modulus, elongation, bending strength, bending modulus, interlaminar shear strength and interlaminar shear modulus; then perform feature screening, since the aggregated test data will have a small amount of experimental error, such as the same input but different output or missing part of the input parameter and low frequency data, it is necessary to clean up the aggregated data, that is, the purpose is to identify and correct the errors, incomplete, inaccurate or inconsistent parts in the data set, the specific steps include missing value processing, outlier processing, repeated value processing and text cleaning; In order to adapt to subsequent analysis and modeling, data arrangement is also needed, such as converting volume fraction to mass fraction, and numerically processing carbon fiber types by elastic modulus to make the data set more standardized.
[0063] Step two, screening the characteristic variables in the CF / PEEK composite parameter-performance data set, and establishing different CF / PEEK composite parameter-performance data models according to the screened characteristic variables; specifically: screening the characteristic variables in the CF / PEEK composite parameter-performance data set, first, using partial least squares (PLS) and mean square error calculation method (MSE) to analyze the parameters in the CF / PEEK composite parameter-performance data set constructed in step one, calculate the change of mean square prediction error with the number of input parameters, determine the optimal number of input parameters, and comprehensively consider the influence of input quantity on the mean square prediction error of tensile strength, bending strength and interlaminar shear strength, and too many input quantities will increase the complexity of the model. In order to balance efficiency and accuracy, after analysis and processing, it is determined that the input quantity is 6, so as to carry out the subsequent training of "material parameter-tensile strength" model, "material parameter-bending strength" model and "material parameter-interlaminar shear strength" model. After analyzing the optimal input quantity, further analysis determines that the six input material parameters are CF mass fraction, CF type, CF monofilament number, melting temperature, forming pressure and holding time, and the three output mechanical properties are tensile strength, bending strength and interlaminar shear strength, which can make the accuracy of the model optimal, mainly by calculating the importance of the above parameters to the predicted performance. According to the six input material parameters: CF mass fraction, CF type, CF monofilament number, melting temperature, forming pressure and holding time, and the three output mechanical properties: tensile strength, bending strength and interlaminar shear strength, the corresponding models are established and trained by using Gaussian process regression, decision tree, support vector machine and integrated algorithm. In order to test the stability of the model, each algorithm corresponds to 100 times of data model training.
[0064] Step three, selecting the optimal CF / PEEK composite parameter-performance data model by comparing the determination coefficient; wherein the determination coefficient is:
[0065]
[0066] Wherein, R 2 is the goodness of fit, SSR is the regression sum of squares, SSE is the residual sum of squares, SST is the total deviation split sum, y i is the true observed value, is the average of the true observation, is the fitted value.
[0067] Wherein, the regression sum of squares SSR is used to measure the degree of explanation of the independent variables to the dependent variable in the regression model, that is, the variation of the dependent variable explained by the regression model. The larger the SSR, the stronger the explanatory ability of the regression model to the dependent variable. The residual sum of squares SSE is used to represent the variation of the dependent variable that cannot be explained by the regression model, that is, the sum of squares of the residuals. The residual is the difference between the observed value of the dependent variable and the predicted value of the regression equation, and the SSE measures the variation of these residuals. The smaller the SSE, the better the fitting degree of the regression model to the data. The total deviation sum of squares SST represents the total variation of the dependent variable, that is, the sum of squares of the difference between the observed value of the dependent variable and its mean; SST = SSR + SSE.
[0068] Step four, the optimal CF / PEEK composite parameter-performance data model is used to analyze the influence trend curve of each characteristic variable on the mechanical property; specifically, by inputting the material parameters into the optimal CF / PEEK composite parameter-performance data model, the mechanical property value is predicted, so that the influence trend curve of each material parameter on the mechanical property can be drawn.
[0069] Step five, the configuration parameters and process parameters required for the mechanical property are optimized by using the frog leap algorithm, that is, the material parameters including the configuration parameters and the process parameters can be back calculated by using the required mechanical property through the frog leap algorithm.
[0070] The carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance provided by the embodiment constructs the configuration-process-performance relationship of the carbon fiber polyether ether ketone composite material, the configuration-process involves the CF volume fraction, the CF mass fraction, the CF diameter, the number of monofilaments per strand of two-dimensional CF cloth, the PEEK mesh number, the melting temperature, the molding pressure, and the holding time; the output is the mechanical property parameter, which involves the tensile strength, the Young's modulus, the elongation rate, the bending strength, the bending modulus, the interlaminar shear strength, and the interlaminar shear modulus. By using the optimal parameter-performance data model, the influence law and mechanism of each parameter on the performance of the CF / PEEK composite material are studied, which provides theoretical guidance for the preparation of high-performance composite materials. Compared with the traditional test method, time and labor are saved, and cost is saved.
[0071] Embodiment two, the embodiment is an example of the data arrangement formula in the carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance described in embodiment one;
[0072] The data arrangement formula is specifically:
[0073]
[0074] Wherein, W CF is the CF mass fraction, VCF CF volume fraction, p CF CF density, p PEEK PEEK density.
[0075] In actual application, the data cleaning formula is used to normalize the data after data cleaning, such as converting the volume fraction into mass fraction, and the conversion formula is shown as formula (1). For example, the type of carbon fiber is numerically processed by the elastic modulus, so that the data set is more standardized to adapt to subsequent analysis and modeling.
[0076] Embodiment three, see Figure 1 and Figure 2 This embodiment is an example of the method of establishing a corresponding model and training in the carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance described in embodiment one.
[0077] Gaussian process regression is specifically: Gaussian process (GP) is a random process composed of a group of random variables subject to Gaussian distribution, wherein any variable subject to joint Gaussian distribution is uniquely determined by the mean function m(x) and the covariance function k(x, x'), as shown in the following formula. The covariance function is also called the kernel function.
[0078] f(x) ~ GP(m(x), k(x, x'))
[0079] Gaussian process regression uses Gaussian process prior. First, a group of observation values, i.e., CF / PEEK composite material parameter performance data set, is obtained, and then the Bayesian theorem is substituted into the Gaussian process for training, and the kernel function and the mean function are corrected. After multiple training, the posterior distribution is obtained, i.e., a group of mapping relationships, which can give a sample to obtain an output prediction value, and the expression is as follows, wherein ε i is an independent and identically distributed noise variable subject to N(0, σ2).
[0080] y (i) = f(x (i) ) + ε (i) , i = 1, …, n
[0081] The decision tree is specifically: the CF / PEEK composite material parameter performance data set is divided into a training set and a test set, and the decision tree model is trained using the training set. The decision tree model will construct a tree according to the characteristics of the data set and the target variable, and divide the data set into subsets recursively and select the best division method, wherein the decision tree model is shown as formula (3). Figure 2
[0082] Support vector machine is specifically: in a linear SVM model, each data point in the CF / PEEK composite parameter performance data set is plotted as a point in an n-dimensional space, a decision boundary is found by finding a separating hyperplane to separate positive and negative samples, and the hyperplane equation is w T x+b=0, in the two-dimensional case, this way represents a straight line, and in the three-dimensional case, the equation represents a plane, first determine the two feature vectors closest to the hyperplane, and the two feature vectors are support vectors. Then determine the distance between the hyperplane and the support vectors in the positive and negative two categories, and the distance is set as the margin. The greater the distance, the better the determined decision boundary, and a schematic diagram of the decision boundary and the margin is shown in Figure 3 Although it is a nonlinear model, it can still use a kernel function to map the data to a new feature space, and find a hyperplane in this higher-dimensional new space, as shown in Figure 4 .
[0083] Integrated algorithm is specifically: ensemble learning is used to train the training set in the CF / PEEK composite parameter performance data set by a plurality of individual machine learning models, and then integrated by group decision to obtain good prediction effect, so as to reduce the generalization classification error caused by only one model, and is suitable for classification and regression problems, as shown in Figure 5 . In the ensemble regression model, variance is used to evaluate the generalization error of a certain individual learner.
[0084] Ensemble learning mainly includes Bagging, Boosting, Stacking and several other methods, and the Boosting method is used in the embodiment. All individual learners of the Boosting method are of one kind. The basic process is to first train a weak learner-base model 1 from the samples, then use this weak learner to predict different training samples, increase the weight of the samples with large prediction error, so that the subsequent weak learner pays more attention to this sample, and then train the next weak learner. After the number of weak learners reaches a specified number, a strong learner is obtained by Adaboost algorithm through a series of linear combinations, as shown in Figure 6 .
[0085] Embodiment four, this embodiment is an example of step S5 in the carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance described in embodiment one;
[0086] In step S4, the influence trend curve of each characteristic variable on the mechanical property is obtained, i.e., the influence trend curve of each material parameter on the mechanical property is known, and finally the frog leap algorithm is used to back calculate the material parameters from the required mechanical property, wherein the frog leap algorithm is used to simulate the route planning of some different groups of frogs to find the place with the most food, and the core idea is that the frogs in the same group can share information. Each frog represents a solution to the problem, each group finds a local optimal solution, and after each group evolves to a certain extent, information is shared between groups to find a global optimal solution. The main mathematical model is as follows: if the solution of Pw is better than Pworst, replace the worst individual.
[0087]
[0088] wherein D represents the next step length, Dmax represents the maximum allowed step length, rand() represents a random number between 0 and 1, P best represents the best solution of the group, P worst worst represents the worst solution of the group; and Pw represents the next direction.
[0089] Embodiment five, the embodiment provides a carbon fiber reinforced polyether ether ketone composite material design system based on machine learning guidance, the system comprises:
[0090] a storage device for constructing a CF / PEEK composite material parameter-performance data set;
[0091] a storage device for screening characteristic variables in the CF / PEEK composite material parameter-performance data set, and establishing different CF / PEEK composite material parameter-performance data models according to the screened characteristic variables;
[0092] a storage device for selecting the optimal CF / PEEK composite material parameter-performance data model by comparison of determination coefficients;
[0093] a storage device for analyzing the influence trend curve of each characteristic variable on the mechanical property by using the optimal CF / PEEK composite material parameter-performance data model;
[0094] a storage device for optimizing the configuration parameters and process parameters of the required mechanical property by using the frog leap algorithm.
[0095] Embodiment six, the embodiment provides a computer readable storage medium, and the computer readable storage medium stores a computer program, and the computer program is run by a processor to execute the carbon fiber reinforced polyether ether ketone composite material design method based on machine learning guidance in any one of embodiments one to four.
[0096] Embodiment seven, the embodiment provides a computer device, the device comprises a memory and a processor, the memory has stored a computer program, when the processor runs the computer program stored in the memory, the processor executes the machine learning guided carbon fiber reinforced polyether ether ketone composite material design method of any one of the embodiments one to four.
[0097] In addition, the terms "first", "second", "third", etc. are used only for descriptive purposes and are not to be construed as indicating or implying relative importance or an indicated number of technical features. Therefore, the features defined with "first", "second", etc. can explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "a plurality of" is at least two, for example, two, three, etc., unless otherwise specifically limited.
[0098] In the description of the present application, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present application, the illustrative description of the above terms is not necessarily for the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. In addition, those skilled in the art can combine and combine different embodiments or examples described in the present application and the features of different embodiments or examples without contradiction.
[0099] The above only describes the embodiments of the present application and does not limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the scope of claims of the present application.
Claims
1. A design method for carbon fiber reinforced polyetheretherketone composite materials based on machine learning guidance, characterized in that, The method is as follows: S1. Construct a parameter-performance dataset for CF / PEEK composite materials; S2. Filter the feature variables in the CF / PEEK composite material parameter-performance dataset, and establish different CF / PEEK composite material parameter-performance data models according to the filtered feature variables. S3. Select the optimal CF / PEEK composite material parameter-performance data model by comparing the coefficients of determination; S4. The influence trend curve of each characteristic variable on mechanical properties is analyzed using the optimal CF / PEEK composite material parameter-performance data model; S5. The frog-leap algorithm is used to optimize the configuration parameters and process parameters of the required mechanical properties.
2. The design method for carbon fiber reinforced polyetheretherketone composite materials based on machine learning guidance according to claim 1, characterized in that, Step S1 specifically involves: S11. Establish a configuration parameter-process parameter-mechanical property dataset based on past test data; S12. Clean and organize the material parameters in the dataset.
3. The design method for carbon fiber reinforced polyetheretherketone composite materials based on machine learning guidance according to claim 2, characterized in that, The configuration parameters include carbon fiber volume fraction, carbon fiber mass fraction, carbon fiber diameter, number of monofilaments per strand of two-dimensional carbon fiber cloth, and mesh count of polyetheretherketone resin. The process parameters include melting temperature, molding pressure, and holding time; The mechanical properties include tensile strength, Young's modulus, elongation, flexural strength, flexural modulus, interlaminar shear strength, and interlaminar shear modulus.
4. The design method for carbon fiber reinforced polyetheretherketone composite materials based on machine learning guidance according to claim 2, characterized in that, The data processing formula is as follows: Among them, W CF V represents the CF quality fraction. CF ρ is the volume fraction of CF. CF ρ is the density of CF. PEEK The density is PEEK.
5. The design method for carbon fiber reinforced polyetheretherketone composite materials based on machine learning guidance according to claim 1, characterized in that, Step S2 specifically involves: S21. The partial least squares regression and mean square error calculation methods are used to analyze the parameter-performance dataset of the CF / PEEK composite material to determine 6 input parameters; S22. Calculate the importance of the parameters in the CF / PEEK composite material parameter-performance dataset, and determine the 6 input parameters and 3 output parameters; S23. Based on the 6 input parameters, establish the corresponding models using Gaussian process regression, decision tree, support vector machine and ensemble algorithm respectively, and train them.
6. The design method for carbon fiber reinforced polyetheretherketone composite materials based on machine learning guidance according to claim 5, characterized in that, The six input parameters are: CF mass fraction, CF type, CF monofilament number, melting temperature, molding pressure, and holding time. The three output parameters are tensile strength, flexural strength, and interlaminar shear strength.
7. The design method for carbon fiber reinforced polyetheretherketone composite materials based on machine learning guidance according to claim 1, characterized in that, The formula for the determination coefficient in step S3 is: Among them, R 2 For goodness of fit, SSR is the regression sum of squares, SSE is the residual sum of squares, and SST is the total sum of deviations. i For the actual observed values, This is the average value of the actual observations. These are the fitted values.
8. A design system for carbon fiber reinforced polyetheretherketone composite materials based on machine learning guidance, characterized in that, The system includes: Storage device for constructing parameter-performance datasets of CF / PEEK composite materials; A storage device for filtering feature variables in the CF / PEEK composite material parameter-performance dataset and establishing different CF / PEEK composite material parameter-performance data models based on the filtered feature variables; Storage device for selecting the optimal CF / PEEK composite material parameter-performance data model by comparing the coefficients of determination; A storage device for analyzing the trend curve of the influence of each characteristic variable on mechanical properties using the optimal CF / PEEK composite material parametric-performance data model; A storage device for optimizing configuration parameters and process parameters for desired mechanical properties using a leapfrog algorithm.
9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the design method for carbon fiber reinforced polyetheretherketone composite materials based on machine learning guidance as described in any one of claims 1-7.
10. A computer device, characterized in that, The device includes a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the design method for carbon fiber reinforced polyether ether ketone composite materials based on machine learning guidance as described in any one of claims 1-7.