Composite pipe bending stiffness prediction method and system and terminal based on deep learning

Abstract

The present application relates to the technical field of composite structure design, and particularly relates to a composite pipe bending stiffness prediction method based on deep learning, a system thereof and a terminal, the composite pipe bending stiffness prediction method comprising: testing the bending performance of a layer-laid composite pipe under cantilever beam conditions by using a finite element model to build a basic data set; processing the basic data set to obtain a standard data set; constructing a neural network model, training the neural network model by using the standard data set, and obtaining a pipe bending stiffness prediction model; and outputting the bending stiffness by using the model. By simulating the composite pipe simulation to obtain the data thereof, the preprocessed data is divided into layer-laid data and pipe geometry data as independent variables, and the bending stiffness as a dependent variable, a sample data set is made, a pipe bending stiffness prediction model is constructed, the pipe bending stiffness can be calculated by using the basic data, and the low-cost and fast and accurate design of the composite pipe is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of structural design, in particular to a composite pipe bending stiffness prediction method based on deep learning and a system and terminal thereof. BACKGROUND

[0002] Due to the excellent performance of composite materials such as high specific strength, specific stiffness, corrosion resistance, etc., more and more traditional metal alloy components are replaced by composite materials. Correspondingly, when the manufacturing industry develops towards intelligentization, mechanical arms as the mainstream manufacturing production machine industry begin to develop towards lightweight to improve work efficiency and reduce energy consumption.

[0003] As a common connecting rod structural component of mechanical arms, when composite materials are used to replace pipe joints, the precise design becomes more complex due to the anisotropic characteristics of composite materials. At present, the finite element method is mainly used for simulation design, but it consumes a large amount of computing power.

[0004] Therefore, the prior art still needs to be improved and developed. SUMMARY

[0005] In view of the deficiencies of the prior art, the purpose of the present application is to provide a composite pipe bending stiffness prediction method based on deep learning and a system and terminal thereof, aiming to solve the problem that the existing pipe bending stiffness prediction method consumes a large amount of computing power.

[0006] The technical scheme of the present application is as follows:

[0007] A composite pipe bending stiffness prediction method based on deep learning, comprising the steps of:

[0008] constructing a finite element model of a layer-laid composite pipe;

[0009] testing the bending performance of the layer-laid composite pipe under cantilever beam conditions using the finite element model, and building a basic data set;

[0010] standardizing and segmenting the basic data set to obtain a standard data set;

[0011] constructing a neural network model, and training the neural network model using the standard data set to obtain a pipe bending stiffness prediction model;

[0012] inputting layer-laid composite pipe parameters into the pipe bending stiffness prediction model, and outputting the bending stiffness of the layer-laid composite pipe.

[0013] The composite pipe bending stiffness prediction method based on deep learning, wherein the step of constructing a finite element model of a layer-laid composite pipe comprises:

[0014] In finite element software, set the basic properties of a single-layer material, construct a multi-layer material model, and set the angles of each layer;

[0015] Set the geometric data of the pipe fitting, establish the composite material pipe fitting using the multilayer material model, divide the mesh elements, and complete the construction of the finite element model of the layered composite material pipe fitting.

[0016] The method for predicting the bending stiffness of composite pipe fittings based on deep learning includes Young's modulus, shear modulus, Poisson's ratio, and mass density as the basic properties.

[0017] The deep learning-based method for predicting the bending stiffness of composite pipe fittings includes the following steps: First, testing the bending performance of the laminated composite pipe fittings under cantilever beam conditions using the finite element model and building a basic dataset.

[0018] By combining the layer data and the pipe fitting geometry data, and scanning each parameter range at a predetermined step size, multiple sets of independent variable data about the layer data and the pipe fitting geometry data are obtained.

[0019] A fixed constraint is added to one end of the laminated composite pipe fitting, and a static load is applied to the other end. Finite element simulation is performed on each set of independent variable data, and the maximum deflection of the laminated composite pipe fitting corresponding to each set of independent variable data is recorded. A basic dataset containing the laminated composite pipe fitting layup data, pipe fitting geometry data, and bending stiffness is constructed.

[0020] The aforementioned deep learning-based method for predicting the bending stiffness of composite pipe fittings, wherein the expression for standardizing and preprocessing the basic dataset is as follows:

[0021] Where X represents the input data, X mean X represents the mean of the data. std This represents the variance of the data.

[0022] The method for predicting the bending stiffness of composite material pipe fittings based on deep learning, wherein the step of constructing a neural network model includes: constructing a model based on a long short-term memory model combined with a fully connected neural network model, and setting training parameters;

[0023] The training parameters include batch size, training rounds, validation frequency, and learning rate.

[0024] The deep learning-based method for predicting the bending stiffness of composite material pipe fittings includes the following step: training the neural network model using the standard dataset to obtain the pipe fitting bending stiffness prediction model.

[0025] The input data of the neural network model is x = [x1, x2], where x1 corresponds to the layup data of the composite material and x2 corresponds to the geometric data of the pipe fitting.

[0026] If x1 passes through the LSTM layer, then the formula for inputting x1 into the processing layer is L(x1) = Lstm(x1, m), where Lstm(x1, m) represents the forward propagation function of the LSTM network and m represents the number of kernels in the LSTM layer.

[0027] If x2 passes through a fully connected layer, the formula for inputting x2 into the processing layer is F1(x2) = Dense(x2; w1, b1), where w1 and b1 are the weights and bias parameters of the fully connected layer, respectively.

[0028] x1 and x2 are concatenated according to the feature dimension, and the formula is C(x) = Concat(x1, x2), where Concat represents the concatenation processing function;

[0029] The spliced ​​data is input into two fully connected layers for processing to obtain the pipe fitting bending stiffness prediction model; the formula for inputting the spliced ​​data into two fully connected layers for processing is F2(x)=Dense(Dense(x;w2,b2);w3,b3), where w2 and w3 represent the weights of the two fully connected layers, and b2 and b3 represent the biases of the two fully connected layers, respectively.

[0030] The loss function for training materials in a neural network is Where y, These are represented as predicted and true values, respectively, and n represents the total number of training samples.

[0031] The deep learning-based method for predicting the bending stiffness of composite pipe fittings includes parameters such as length, radius, thickness, and layup angle.

[0032] A composite material pipe fitting bending stiffness prediction system for implementing the deep learning-based composite material pipe fitting bending stiffness prediction method includes:

[0033] The first model building module is used to build the finite element model of the laminated composite pipe fitting;

[0034] The data construction module is used to test the bending performance of the laminated composite pipe fittings and build a basic dataset.

[0035] A data processing module is used to process the basic dataset;

[0036] The second model building module is used to build and train neural network models;

[0037] The data output module is used to output the bending stiffness of the composite material pipe fittings.

[0038] A terminal includes a memory, a processor, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the deep learning-based composite material pipe bending stiffness prediction method.

[0039] Beneficial Effects: This invention provides a method, system, and terminal for predicting the bending stiffness of composite material pipe fittings based on deep learning. The method includes the following steps: constructing a finite element model of the laminated composite material pipe fitting; testing the bending performance of the laminated composite material pipe fitting under cantilever beam conditions using the finite element model to build a basic dataset; performing standardization preprocessing and segmentation on the basic dataset to obtain a standard dataset; constructing a neural network model and training the neural network model using the standard dataset to obtain a pipe fitting bending stiffness prediction model; inputting the laminated composite material pipe fitting parameters into the pipe fitting bending stiffness prediction model and outputting the bending stiffness of the laminated composite material pipe fitting. This invention acquires relevant data through simulation of composite material pipe fittings, preprocesses the data, and divides the preprocessed data into layer layup data, pipe fitting geometry data, and bending stiffness. Layer layup data and pipe fitting geometry data are used as independent variables, and bending stiffness is used as the dependent variable to create a sample dataset and construct a pipe fitting bending stiffness prediction model. In this way, the bending stiffness of composite material pipe fittings can be calculated from basic data, achieving low-cost and rapid accurate design of composite material pipe fittings. Compared to the traditional reverse design process of pipe fittings based on finite element analysis, the neural network model used in this invention can be reused after only one training, while finite element analysis requires remodeling for each design and has a certain skill level. The neural network model constructed in this invention, once trained, can be easily called to achieve multiple reverse designs of pipe fittings, greatly improving design efficiency and reducing computational consumption. Attached Figure Description

[0040] Figure 1 This is a flowchart illustrating the method for predicting the bending stiffness of composite material pipe fittings based on deep learning, as described in this invention.

[0041] Figure 2 This is a detailed flowchart illustrating the deep learning-based method for predicting the bending stiffness of composite pipe fittings according to the present invention.

[0042] Figure 3 This is a schematic diagram of the finite element model of the CFRP layer pipe laying component in Example 1;

[0043] Figure 4 This is a schematic diagram of the finite element model of the CFRP pipe fitting under load in Example 1;

[0044] Figure 5 This is a schematic diagram of the basic neural network model in Example 1;

[0045] Figure 6 This is a graph of the training-validation loss values ​​of the neural network in Example 1;

[0046] Figure 7 This is a graph showing the prediction results of the neural network in the test set in Example 1. Detailed Implementation

[0047] This invention provides a method, system, and terminal for predicting the bending stiffness of composite material pipe fittings based on deep learning. To make the objectives, technical solutions, and effects of this invention clearer and more explicit, the invention is further described in detail below. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0048] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0049] like Figure 1 As shown, this invention provides a method for predicting the bending stiffness of composite material pipe fittings based on deep learning, including the following steps:

[0050] Step S10: Construct the finite element model of the laminated composite pipe fitting;

[0051] Step S20: Test the bending performance of the laminated composite pipe under cantilever beam conditions using the finite element model, and build a basic dataset;

[0052] Step S30: Perform standardization preprocessing and segmentation on the basic dataset to obtain a standard dataset;

[0053] Step S40: Construct a neural network model and train the neural network model using the standard dataset to obtain a pipe fitting bending stiffness prediction model;

[0054] Step S50: Input the parameters of the laminated composite pipe into the pipe bending stiffness prediction model, and output the bending stiffness of the laminated composite pipe.

[0055] In this embodiment, data is obtained through simulation of composite material pipe fittings. The data is preprocessed, dividing it into layer layup data, pipe fitting geometry data, and bending stiffness. Layer layup data and pipe fitting geometry data are used as independent variables, and bending stiffness as the dependent variable to create a sample dataset and construct a pipe fitting bending stiffness prediction model. Thus, the bending stiffness of composite material pipe fittings can be calculated from basic data, achieving low-cost and rapid, accurate design of composite material pipe fittings. Compared to the traditional reverse design process for pipe fittings based on finite element analysis, the neural network model used in this invention can be reused after only one training iteration. Finite element analysis requires remodeling for each design, and the operators have a certain skill level. The neural network model constructed in this invention, once trained, can be easily called to achieve multiple reverse designs of pipe fittings, greatly improving design efficiency and reducing computational consumption.

[0056] Specifically, the deep learning-based method for predicting the bending stiffness of composite pipe fittings is applicable to layered pipe fittings constructed from anisotropic composite materials. By appropriately adjusting the training dataset and model parameters, it can effectively realize the design functions of various pipe fittings, making the deep learning-based method for predicting the bending stiffness of pipe fittings highly compatible.

[0057] In a preferred embodiment, the deep learning-based method for predicting the bending stiffness of composite pipe fittings is used in the reverse design of CFRP (Carbon Fiber Reforcement Polymer) pipe fittings.

[0058] In some embodiments, step S10, the step of constructing the finite element model of the laminated composite pipe fitting, includes:

[0059] Step S11: Set the basic properties of a single-layer material in the finite element software, construct a multi-layer material model, and set the angles of each layer;

[0060] Step S12: Set the geometric data of the pipe fitting, establish the composite material pipe fitting using the multilayer material model, divide the mesh elements, and complete the construction of the finite element model of the layered composite material pipe fitting.

[0061] In this embodiment, the basic properties of a single-layer material are first set in the finite element software, including Young's modulus, shear modulus, Poisson's ratio, and mass density. Then, a basic multilayer material model is established, and the angles of each layer are set. Next, the basic geometric properties of the composite material pipe fitting are set. The composite material pipe fitting is established using multilayer materials, and the mesh elements are divided to complete the establishment of the finite element model of the basic composite material layer pipe fitting.

[0062] Specifically, constructing the finite element model of the laminated composite pipe fitting includes: determining the laminated design variables and the pipe fitting geometric variables, wherein the length of the pipe fitting is set as L [mm], the radius of the intermediate layer (midpoint in the thickness direction) is R [mm], and the layup angle of each layer is set as θ. i The number of layers is n. When the fiber direction is set to the x-axis, the layup angle is 0, and the layup sequence corresponds to the radial direction of the pipe from the inner wall to the outer wall.

[0063] In some implementations, the basic properties include, but are not limited to, Young's modulus, shear modulus, Poisson's ratio, and mass density.

[0064] In some embodiments, step S20, the step of testing the bending performance of the laminated composite pipe fitting under cantilever beam conditions using the finite element model and building a basic dataset, includes:

[0065] Step S21: Combine the layer data and pipe fitting geometry data, and scan each parameter range at a predetermined step size to obtain multiple sets of independent variable data about the layer data and pipe fitting geometry data;

[0066] Step S22: Add a fixed constraint to one end of the laminated composite pipe fitting and apply a static load to the other end. Perform finite element simulation on each set of independent variable data, record the maximum deflection of the laminated composite pipe fitting corresponding to each set of independent variable data, and build a basic dataset containing the laminated composite pipe fitting layup data, pipe fitting geometry data and bending stiffness.

[0067] Specifically, the finite element model is used to analyze the end displacement of the pipe under static load in the cantilever beam model. Then, the basic design data of the pipe is changed and the finite element analysis is repeated. All test data are recorded to form a basic dataset.

[0068] In some implementations, a fixed constraint is added to one end of the laminated composite pipe, and a static load is applied to the other end to simulate the bending characteristics of the rod as a cantilever beam structure. Then, by performing a finite element simulation with loading, the displacement before and after applying the static load to the end of the pipe is recorded, which yields the maximum deflection of the composite pipe.

[0069] Specifically, deflection can be expressed by the formula The bending stiffness is converted into the bending stiffness of the pipe; where EI represents the bending stiffness of the pipe, F corresponds to the static load at the end of the pipe, L represents the length of the pipe, and y represents the deflection of the pipe obtained by finite element simulation. This formula can be used to convert the deflection predicted by the model into the bending stiffness of the pipe.

[0070] In some implementations, step S21 specifically involves: combining the layup data and pipe fitting geometry data within a certain design range, scanning each parameter range with a certain step size to obtain multiple sets of independent variable data about the layup data and pipe fitting geometry characteristics, performing a finite element simulation for each set accordingly, recording all data, i.e. changing the basic design and repeating step S22 to form a complete dataset containing composite material pipe fitting layup design, geometry design and bending stiffness.

[0071] In some implementations, the expression for standardizing and preprocessing the basic dataset is as follows: Where X represents the input data, X mean X represents the mean of the data. std This represents the variance of the data.

[0072] In some implementations, the input data X includes the sequence data of the layering, the length of the pipe, the radius of the pipe, the thickness of the pipe, etc., all of which are standardized and preprocessed in their respective data spaces.

[0073] In some implementations, the segmentation process involves dividing the standardized preprocessed basic dataset into a training set, a test set, and a validation set; specifically, the standardized preprocessed basic dataset is randomly shuffled and selected proportionally as the training set, validation set, and test set.

[0074] In some implementations, the standard dataset may consist of 70% training set, 15% validation set, and 15% test set, or it may consist of 80% training set, 10% validation set, and 10% test set, but is not limited thereto.

[0075] In some implementations, the step of constructing the neural network model includes: constructing a long short-term memory model combined with a fully connected neural network model, and setting training parameters; wherein the training parameters include batch size, training epochs, validation frequency, and learning rate.

[0076] Specifically, a neural network prediction model based on LSTM (Long Short-Term Memory) combined with FCNN (Full Connected Neural Network) was constructed, and training parameters were set, including batch size, training epochs, and validation frequency. The optimizer was set to Adam, and the learning rate was set to 0.0001.

[0077] In some embodiments, step S40, which involves training the neural network model using the standard dataset to obtain a pipe fitting bending stiffness prediction model, includes:

[0078] The input data of the neural network model is x = [x1, x2], where x1 corresponds to the layup data of the composite material and x2 corresponds to the geometric data of the pipe fitting.

[0079] If x1 passes through the LSTM layer, then the formula for inputting x1 into the processing layer is L(x1) = Lstm(x1, m), where Lstm(x1, m) represents the forward propagation function of the LSTM network and m represents the number of kernels in the LSTM layer.

[0080] If x2 passes through a fully connected layer, the formula for inputting x2 into the processing layer is F1(x2) = Dense(x2; w1, b1), where w1 and b1 are the weights and bias parameters of the fully connected layer, respectively.

[0081] x1 and x2 are concatenated according to the feature dimension, and the formula is C(x) = Concat(x1, x2), where Concat represents the concatenation processing function;

[0082] The spliced ​​data is input into two fully connected layers for processing to obtain the pipe fitting bending stiffness prediction model; the formula for inputting the spliced ​​data into two fully connected layers for processing is F2(x)=Dense(Dense(x;w2,b2);w3,b3), where w2 and w3 represent the weights of the two fully connected layers, and b2 and b3 represent the biases of the two fully connected layers, respectively.

[0083] The loss function for training materials in a neural network is Where y, These are represented as predicted and true values, respectively, and n represents the total number of training samples.

[0084] In this embodiment, compared to the traditional reverse design of pipe fittings based on finite element analysis, the neural network model used in this invention can be reused after only one training, while finite element analysis requires remodeling for each design and has a certain skill threshold for operators. Once the neural network model constructed in this invention is trained, it can be used to perform reverse design of pipe fittings multiple times simply by calling it, which greatly improves design efficiency and reduces computing power consumption.

[0085] In some embodiments, after obtaining the pipe fitting bending stiffness prediction model, the method further includes evaluating the pipe fitting bending stiffness prediction model; the step of evaluating the pipe fitting bending stiffness prediction model includes: using the mean square error, root mean square error and mean absolute error of the model prediction results and actual results in the test set to evaluate the quality of the pipe fitting bending stiffness prediction model.

[0086] Specifically, the test set is input into the pipe fitting bending stiffness prediction model to calculate the predicted bending stiffness value. Then, the mean square error, root mean square error, and mean absolute error of both are calculated. These are used to evaluate the model's performance, and the training and validation loss graphs during the training process are examined to verify whether the training was performed well. Furthermore, hyperparameters such as the optimizer, training epochs, and loss function can be adjusted to optimize the network and improve its fitting ability.

[0087] In some implementations, a basic dataset can be constructed using deflection values, layup data of the composite material pipe fittings, and pipe fitting geometry data. In this case, the output of the pipe fitting bending stiffness prediction model is the deflection value. Then, the deflection value is converted into bending stiffness using the relationship between bending stiffness, deflection, and geometric parameters, and finally, the bending stiffness value is obtained. Alternatively, the deflection can be converted first using the relationship between bending stiffness, deflection, and geometric parameters, with the layup data of the composite material pipe fittings, pipe fitting geometry data, and bending stiffness as the basic dataset.

[0088] In some embodiments, the parameters of the laminated composite pipe fitting include length, radius, thickness, and layup angle.

[0089] Specifically, the length, radius, thickness, and ply angle of the pipe fittings are used as input data, and then the mean X of the original data stored in the standardized preprocessing process is used. mean and variance X std The input data is standardized, and then the standardized data is input into the pipe fitting bending stiffness prediction model. The model obtains standardized output values, which are then inversely standardized. Specifically, the inverse standardization process is: x*X sts +X mean Where x is the value to be inversely standardized, the predicted value y of the pipe deflection is obtained. Since the test is for a cantilever beam, the relationship between its bending stiffness, deflection, and geometric parameters is as follows:

[0090]

[0091] Where EI represents the bending stiffness of the pipe fitting, F corresponds to the static load at the end of the pipe fitting, and L represents the length of the pipe fitting. Using this formula, the deflection predicted by the model can be converted into the bending stiffness of the pipe fitting.

[0092] Specifically, such as Figure 2As shown, the method for predicting the stiffness and strength of pipe fittings based on a neural network model includes the following steps: establishing a finite element model of the CFRP pipe fitting, determining the layer design variables and the geometric variables of the pipe fitting; applying a fixed constraint to one end of the pipe fitting and a static load to the other end to obtain the displacement at the end of the fitting; selecting several sets of different layer data and geometric data of the fitting, repeating the previous step to obtain several sets of CFRP pipe fitting bending simulation data, i.e., the basic dataset; preprocessing the original dataset, classifying the input data into layer sequence data and pipe fitting geometric data, inputting them into the neural network model and training it to obtain a pipe fitting bending stiffness predictor; using the mean square error, root mean square error, and mean absolute error of the model prediction results and actual results in the test set to evaluate the quality of the model; given the length, radius, thickness, and layer angle of the pipe fitting, using the trained neural network to predict the bending stiffness of the pipe fitting.

[0093] In addition, the present invention also provides a composite material pipe fitting bending stiffness prediction system for implementing a deep learning-based method for predicting the bending stiffness of composite material pipe fittings, comprising:

[0094] The first model building module is used to build the finite element model of the laminated composite pipe fitting;

[0095] The data construction module is used to test the bending performance of the laminated composite pipe fittings and build a basic dataset.

[0096] A data processing module is used to process the basic dataset;

[0097] The second model building module is used to build and train neural network models;

[0098] The data output module is used to output the bending stiffness of the composite material pipe fittings.

[0099] In addition, the present invention also provides a terminal, the terminal including a memory, a processor and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, it implements the steps of a deep learning-based method for predicting the bending stiffness of composite material pipe fittings.

[0100] In some embodiments, the memory may be an internal storage unit of the terminal, such as a hard drive or RAM. In other embodiments, the memory may be an external storage device of the terminal, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Further, the memory may include both internal and external storage units. The memory is used to store application software and various types of data installed on the terminal, such as the program code installed on the terminal. The memory can also be used to temporarily store data that has been output or will be output. In one embodiment, the memory stores a composite material pipe bending stiffness prediction program, which can be executed by a processor to implement the composite material pipe bending stiffness prediction method of the present invention.

[0101] In some embodiments, the processor may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory or process data, such as executing the method for predicting the bending stiffness of the composite material pipe fitting.

[0102] In addition, the present invention also provides a computer-readable storage medium having a computer program stored thereon, the computer-readable storage medium storing a composite material pipe fitting bending stiffness prediction program, the composite material pipe fitting bending stiffness prediction program being executed by a processor to implement the steps of the composite material pipe fitting bending stiffness prediction method.

[0103] The following examples further illustrate the present invention in detail. It should also be understood that the following examples are only for further explanation of the present invention and should not be construed as limiting the scope of protection of the present invention. Any non-essential improvements and adjustments made by those skilled in the art based on the above description of the present invention are within the scope of protection of the present invention.

[0104] Example 1

[0105] This embodiment first establishes a basic finite element analysis model, then uses finite element analysis to construct a dataset, and after preprocessing the dataset, constructs a neural network model, specifically including the following steps:

[0106] Step 1: Construct a pipe fitting made of multi-layer unidirectional carbon fiber reinforced composite prepreg (hereinafter referred to as CFRP pipe fitting), such as... Figure 3 As shown, the thickness of each layer of composite material is 0.2 mm, the length of the pipe is L [mm], and the center radius is R.

[0107] The mechanical properties of single-layer CFRP materials include:

[0108] Young's modulus: (E1, E2, E3) = (121 GPa, 8.6 GPa, 8.6 GPa);

[0109] Shear modulus: (G1, G2, G3) = (4.7 GPa, 3.1 GPa, 4.7 GPa);

[0110] Poisson's ratio: (ν1,ν2,ν3)=(0.27,0.4,0.27), where E1,E2,E3 represent Young's modulus in the x, y, and z directions, respectively; G1,G2,G3 represent shear modulus in the x, y, and z directions, respectively; ν1,ν2,ν3 represent Poisson's ratio in the x, y, and z directions, respectively; and the mass density of the composite material is 1490 g / cm³. 3 .

[0111] Step 2: Apply a fixed constraint to the left end of the pipe fitting and a static load F of 100N to the right end. Perform a loading analysis and record the displacement ω in the z-direction of the pipe fitting end. Figure 4 As shown.

[0112] Step 3: Change the basic design data of the pipe fitting, repeat step 2, select the pipe fitting length L∈[300mm,1200mm], the pipe fitting radius R∈[30mm,50mm], set the step size of the two variables to 20mm and 2mm respectively, combine them to construct 503 sets of pipe fitting geometric data, and further select multiple different ply combinations.

[0113] In this example, the following ply combinations are tested:

[0114] [0,45,0,90,0,45,0,90,0,45], [0,90,0,90,0,90,0,90], [-45,45,0,-45,45,0,-45,45,0,0], [30,60,30,60,30,60,30,60,30,60], [90,90,0,90,90,90,90,0,90,90] [10,20,30,40,50,50,40,30,20,10]、[0,90,90,0,0,90,90,0,0,90,90,0]、[45,90,-45,90,45,90,-45,90,45,90,-45,90]、[-45,45,-45,45,-45,45,-45,45,-45,45]、[45,0,-45,0,90,0,45,0,-45,0]、[45,0,-45,0,90,0,45,0,-45,0] [90,0], [90,45,0,45,90,0,90,45,0,45,90,0], [30,45,60,90,60,45,30,45,60,90,60,45], [90,45,-45,90,90,45,-45,90,90,45,-45,90,90], where the unit is angle value, representing the layup angle of the composite prepreg from the first layer to the last layer. A total of 6072 sets of simulation data were obtained, which constitute the basic dataset.

[0115] Step 4: Preprocess the data. In this embodiment, the geometric parameters, displacement parameters and ply angles are standardized, and the dimension of all ply angle data is set to 12. Then, the intermediate values ​​of the standardization, namely the mean and variance of each parameter, are saved for use in the subsequent destandardization process.

[0116] Step 5: Construct a neural network model based on LSTM and FCNN architectures (e.g., Figure 5 As shown in the figure, the dataset from step 4 is divided into 80% training set, 10% validation set, and 10% test set. The loss function is set to Mse, the optimization solver is Adam, and the training is performed for 200 epochs with a learning rate of 0.0001. The training results are shown in the figure. Figure 6 As shown in the figure, the mean squared error is 0.000186, the root mean square error is 0.013623, and the mean absolute error is 0.010115, indicating that the model training results fit the data very well. The prediction performance was tested on the test set, and the prediction results are plotted as shown in the figure. Figure 7 As shown.

[0117] Step 6: Use the trained model to predict the bending stiffness of the pipe fitting. The layup angle of the pipe fitting is [0, 90, 60, 90, 0, 0, 90, 60, 90, 0], the length of the pipe fitting is 1100 mm, and the radius is 35 mm. Use the mean and variance of each parameter saved in Step 4 to uniformly process the above design parameters, and then input them into the model to obtain the prediction results. After inverse standardization, the bending stiffness prediction of the laminated composite pipe fitting can be completed.

[0118] In summary, this invention provides a method, system, and terminal for predicting the bending stiffness of composite material pipe fittings based on deep learning. The method includes the following steps: constructing a finite element model of the laminated composite material pipe fitting; testing the bending performance of the laminated composite material pipe fitting under cantilever beam conditions using the finite element model to build a basic dataset; performing standardization preprocessing and segmentation on the basic dataset to obtain a standard dataset; constructing a neural network model and training the neural network model using the standard dataset to obtain a pipe fitting bending stiffness prediction model; inputting the laminated composite material pipe fitting parameters into the pipe fitting bending stiffness prediction model and outputting the bending stiffness of the laminated composite material pipe fitting. This invention acquires relevant data through simulation of composite material pipe fittings, preprocesses the data, and divides the preprocessed data into layer layup data, pipe fitting geometry data, and bending stiffness. Layer layup data and pipe fitting geometry data are used as independent variables, and bending stiffness is used as the dependent variable to create a sample dataset and construct a pipe fitting bending stiffness prediction model. In this way, the bending stiffness of composite material pipe fittings can be calculated from basic data, achieving low-cost and rapid accurate design of composite material pipe fittings. Compared to the traditional reverse design process of pipe fittings based on finite element analysis, the neural network model used in this invention can be reused after only one training, while finite element analysis requires remodeling for each design and has a certain skill level. The neural network model constructed in this invention, once trained, can be easily called to achieve multiple reverse designs of pipe fittings, greatly improving design efficiency and reducing computational consumption.

[0119] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A method for predicting the bending stiffness of a composite pipe fitting based on deep learning, characterized by, The method comprises the steps of: constructing a finite element model of the layer-ply composite pipe; testing the bending resistance of the layer-ply composite pipe under cantilever beam conditions using the finite element model to build a basic data set; standardizing and segmenting the basic data set to obtain a standard data set; constructing a neural network model and training the neural network model using the standard data set to obtain a pipe bending stiffness prediction model; inputting layer-ply composite pipe parameters into the pipe bending stiffness prediction model to output the bending stiffness of the layer-ply composite pipe; the step of testing the bending resistance of the layer-ply composite pipe under cantilever beam conditions using the finite element model to build a basic data set comprises: combining layer-ply data and pipe geometry data to scan each parameter interval at a predetermined step size to obtain multiple sets of independent variable data about the layer-ply data and pipe geometry data; adding a fixed constraint to one end of the layer-ply composite pipe and applying a static load to the other end to perform finite element simulation on each set of independent variable data and record the maximum deflection of the layer-ply composite pipe corresponding to each set of independent variable data to build a basic data set containing layer-ply composite pipe layer-ply data, pipe geometry data, and bending stiffness; the step of constructing a neural network model comprises constructing a long short-term memory model combined with a fully connected neural network model and setting training parameters; wherein the training parameters include batch size, training rounds, validation frequency, and learning rate; the step of training the neural network model using the standard data set to obtain a pipe bending stiffness prediction model comprises: Input data of the neural network model ], wherein x1 corresponds to layer layup data of the composite material, and x2 corresponds to geometric data of the pipe fitting; x1 is processed by the LSTM layer, then the formula of x1 input processing layer is wherein represents the forward propagation function of the LSTM network, and m represents the number of kernels of the LSTM layer; x2 is passed through a fully connected layer, then the formula for x2 input to the processing layer is wherein are the weight and bias parameters of the fully connected layer, respectively will be described below. and Concatenate by feature dimension, formula where Concat represents a concatenation processing function. The spliced data is input into two full connection layers for processing to obtain a pipe bending stiffness prediction model; wherein the formula for inputting the spliced data into two full connection layers for processing is , , respectively represent the weights of two full connections, , respectively represent the biases of two full connections; The loss function of the material for neural network training is wherein , respectively represent the predicted value and the true value, and n represents the total amount of training samples.

2. The deep learning-based composite pipe fitting bending stiffness prediction method according to claim 1, characterized by, the step of constructing a finite element model of the layer-ply composite pipe comprises: setting basic properties of single-layer materials in a finite element software, constructing a multi-layer material model, and setting the angles of each layer; setting the geometry data of the pipe, establishing a composite pipe using the multi-layer material model, and dividing the grid elements to complete the construction of the finite element model of the layer-ply composite pipe.

3. The deep learning-based composite pipe fitting bending stiffness prediction method according to claim 2, characterized by, The basic properties include Young's modulus, shear modulus, Poisson's ratio, and mass density.

4. The deep learning-based composite pipe fitting bending stiffness prediction method according to claim 1, characterized by, The expression for normalizing the basic data set is ; where X represents input data, represents a mean value of data, represents a variance of data.

5. The deep learning-based composite pipe fitting bending stiffness prediction method according to claim 1, characterized by, The layer-ply composite pipe parameters include length, radius, thickness, and layer-ply angle.

6. A composite pipe flexural stiffness prediction system for implementing the deep learning-based composite pipe flexural stiffness prediction method according to any one of claims 1-5, characterized in that, The method comprises the steps of: a first model construction module for constructing a finite element model of the layer-ply composite pipe; a data building module for testing the bending resistance of the layer-ply composite pipe and building a basic data set; a data processing module for processing the basic data set; a second model construction module for constructing a neural network model and training the model; a data output module for outputting the bending stiffness of the layer-ply composite pipe.

7. A terminal, characterized by comprising: The terminal comprises a memory, a processor, and a program stored on the memory and executable on the processor, and the program, when executed by the processor, implements the steps of the composite pipe bending stiffness prediction method based on deep learning according to any one of claims 1-5.

Images