Fiber composite material compression strength estimation method based on 3D printing

Abstract

A fiber composite material compression strength estimation method based on 3D printing comprises the following steps: collecting sample data, determining process parameters, preparing a sample, and compressing, testing and recording data; preprocessing the original data to form a data set; secondly, constructing a recurrent neural network model, capturing a long-term dependency relationship by utilizing the sequence data processing capability of the recurrent neural network model to estimate the strength, training through a training set, optimizing the performance by using a verification set, and evaluating through a test set to determine a final model; and finally, inputting new sample material characteristics and printing parameters into the model to obtain an intensity estimation value. The method has an effective evaluation verification and updating mechanism, and can adapt to different environments and material characteristic changes.

Description

Technical Field

[0001] The present invention relates to the field of 3D printing, and in particular to a method for estimating the compressive strength of fiber composite materials based on 3D printing. Background Art

[0002] With the rapid development of 3D printing technology, fiber-reinforced composite materials have been widely used in aerospace, automobile manufacturing, building structures, biomedicine and other fields due to their excellent mechanical properties and designability. 3D printing technology can realize the integrated manufacturing of complex structures, and the addition of fiber composite materials further improves the performance of 3D printed products. However, since the performance of fiber composite materials is affected by many factors, such as fiber arrangement, matrix material, printing parameters, etc., the accurate estimation of key mechanical properties such as compressive strength has become a technical problem.

[0003] Traditional methods for estimating compressive strength usually rely on experimental testing, which is not only time-consuming and labor-intensive, but also costly and difficult to adapt to the needs of rapid iteration of designs. In addition, experimental testing can often only be performed on specific samples, making it difficult to cover all possible combinations of printing parameters, and therefore difficult to fully reflect the actual situation of material performance. Summary of the invention

[0004] To solve the above problems, the present invention proposes a method for estimating the compressive strength of fiber composite materials based on 3D printing. By collecting and analyzing a large amount of process parameter and material performance data and using advanced machine learning algorithms, the method can achieve efficient and accurate prediction of the compressive strength of fiber composite materials. The specific steps are as follows, which are characterized by:

[0005] Step 1: Sample data collection, determine different process parameter combinations, use a 3D printer to prepare fiber composite material samples according to the determined process parameters, perform compression tests on the prepared samples, and record relevant information of the samples;

[0006] Step 2: Data preprocessing: sorting the collected raw data, sorting the material properties, printing parameters and compression strength data of the samples accordingly to form a data set;

[0007] Step 3, construct a recurrent neural network, process the sequence data through the recurrent neural network, capture the long-term dependencies in the data to estimate the compressive strength of the fiber composite material;

[0008] Step 4: Model training optimization: Use the training set to train the constructed recurrent neural network model. During the training process, use the validation set to verify the performance of the model.

[0009] Step 5: Model evaluation and verification: Use the test set to evaluate and verify the constructed recurrent neural network model, and release the final version of the fiber composite material compressive strength estimation model;

[0010] Step 6, compressive strength estimation, when estimating the compressive strength of a new 3D printed fiber composite material sample, the sample material properties and printing parameters are used as input, and the trained model is used to predict the input sample information to obtain the estimated value of the sample's compressive strength.

[0011] The present invention provides a method for estimating the compressive strength of fiber composite materials based on 3D printing, and the beneficial effects are as follows:

[0012] 1. The present invention collects samples and conducts tests by combining multiple process parameters, ensuring comprehensive data coverage. The constructed data set contains rich information and provides solid data support for the accurate estimation of compressive strength.

[0013] 2. The recurrent neural network model used in the present invention has the ability to process sequence data and can effectively capture the long-term dependencies in the data. This advanced model can deeply explore the inherent laws of the data, thereby achieving high-precision prediction of the compressive strength of the material.

[0014] 3. The present invention only needs to input the relevant parameters of the new sample to quickly obtain the estimated value of the compressive strength, and the operation process is simple and convenient. In addition, the model has an evaluation verification and update mechanism, which can flexibly adapt to different application environments and changes in material properties, enhancing its practicality and adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a flow chart of the present invention;

[0016] Figure 2 This is a structural diagram of the network model of the present invention. DETAILED DESCRIPTION

[0017] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments:

[0018] The present invention proposes a method for estimating the compressive strength of fiber composite materials based on 3D printing. The invention flow chart is as follows: Figure 1 As shown, the steps of the present invention are described in detail below.

[0019] Step 1: Sample data collection, determine different process parameter combinations, use a 3D printer to prepare fiber composite material samples according to the determined process parameters, perform compression tests on the prepared samples, and record relevant information of the samples;

[0020] Step 1.1, determine different combinations of process parameters, including printing temperature T, printing speed v, layer thickness h, fiber type F, fiber content C, and matrix material M, set the printing temperature between 180℃-250℃, the printing speed between 20mm / s-100mm / s, and the layer thickness between 0.1mm-0.5mm;

[0021] Step 1.2: According to the determined process parameters, use a 3D printer to prepare fiber composite material samples to ensure the consistency and reliability of the samples. Let the sample number be i. For the i-th sample, the process parameters in its preparation process are T i 、v i 、h i 、F i , C i 、M i ;

[0022] Step 1.3, perform compression test on the prepared samples, obtain the compression strength data of each sample through the electronic tensile machine, and record the relevant information of the test samples; let the compression strength of the i-th sample be σ i During the compression test, the compression strength data of each sample, the corresponding process parameters and sample number information are recorded. The compression strength σ i It can be calculated by the formula:

[0023]

[0024] Among them, F max is the maximum load that the sample is subjected to during compression, and A is the cross-sectional area of ​​the sample.

[0025] Step 2: Data preprocessing: sorting the collected raw data, sorting the material properties, printing parameters and compression strength data of the samples accordingly to form a data set;

[0026] Step 2.1: Sort the collected raw data. Let the data set D be:

[0027] D={(x1,σ1),(x2,σ2),…(x n ,σ n )}

[0028] Where n is the number of samples, σ i represents the compression strength of the i-th sample, x i Represents the feature vector of the i-th sample, including material properties and printing parameters:

[0029] x i =[T i ,v i ,h i ,Fi ,C i ,M i ]

[0030] Among them, T i is the printing temperature of the ith sample, v i is the printing speed of the ith sample, h i is the layer thickness of the ith sample, F i is the fiber type of the i-th sample, C i is the fiber content of the ith sample, M i is the matrix material of the ith sample;

[0031] Step 2.2, normalize the feature data in the data set, and use the minimum-maximum normalization method to normalize the data.

[0032] Step 3, construct a recurrent neural network, process the sequence data through the recurrent neural network, capture the long-term dependencies in the data to estimate the compressive strength of the fiber composite material;

[0033] Step 3.1, construct the input layer of the network. The number of nodes in the input layer corresponds to the number of features of the sample, which is set to 6;

[0034] Step 3.2, the hidden layer has m LSTM units. There are multiple gate structures inside the LSTM unit to control the flow of information. The formula of the forget gate is as follows:

[0035] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0036] Among them, f t is the output of the forget gate at time t, W f is the weight matrix of the forget gate, h t-1 is the hidden state at the previous moment, x t is the input at the current moment, b f is the bias vector of the forget gate, σ is the activation function, and t is the time step index;

[0037] The input gate formula is as follows:

[0038] i t =σ(W i ·[h t-1 ,x t ]+b i )

[0039] g t =σ(W g ·[ht-1 ,x t ]+b g )

[0040] Among them, i t is the output of the input gate at time t, g t is the output of the input node after the activation function at time t, W i , W g is the weight matrix of the input gate and input node, b i 、b g is the bias vector;

[0041] The cell state update formula is as follows:

[0042] C t =f t ⊙C t-1 +i t ⊙g t

[0043] Among them, C t is the cell state at time t, ⊙ represents element-wise multiplication;

[0044] The output gate formula is as follows:

[0045] O t =σ(W o ·[h t-1 ,x t ]+b o )

[0046] h t =O t ⊙tanh(C t )

[0047] Among them, O t is the output of the output gate at time t, W o is the weight matrix of the output gate, b o is the bias vector, h t is the hidden state at the current moment;

[0048] Step 3.3, the number of output layer nodes is 1, and the predicted compression strength is output;

[0049] Step 3.4, select the activation function. The activation function inside the LSTM unit uses the ReLU function, and the output layer uses the linear activation function;

[0050] Step 3.5, divide the data set, divide the preprocessed data set into training set, validation set and test set according to a certain ratio, use 70% of the data set as training set, 10% of the data set as validation set, and the remaining data as test set.

[0051] Step 4: Model training optimization: Use the training set to train the constructed recurrent neural network model. The network model structure diagram is as follows: Figure 2 As shown, during the training process, the validation set is used to verify the performance of the model;

[0052] Step 4.1, use the training set to train the constructed recurrent neural network model. For the recurrent neural network, at each time step t, calculate the error between the predicted output and the actual output, and use the mean square error as the loss function. The loss function is:

[0053]

[0054] in, is the model's predicted compression strength value for the i-th training sample, y i is the true compression strength value of the i-th training sample, N train is the number of samples in the training set after partitioning;

[0055] Step 4.2, during the training process, adjust the weights and biases of the model through the back-propagation algorithm;

[0056] In step 4.3, the performance of the model is verified using the validation set. According to the results of the validation set, the hyperparameters of the model are adjusted, including the learning rate, the number of hidden layer nodes, and the regularization parameters.

[0057] Step 5: Model evaluation and verification: Use the test set to evaluate and verify the constructed recurrent neural network model, and release the final version of the fiber composite material compressive strength estimation model;

[0058] Step 6, compressive strength estimation, when estimating the compressive strength of a new 3D printed fiber composite material sample, the sample material properties and printing parameters are used as input, and the trained model is used to predict the input sample information to obtain the estimated value of the sample's compressive strength.

[0059] The above description is only a preferred embodiment of the present invention and does not constitute any other form of limitation to the present invention. Any modification or equivalent change made based on the technical essence of the present invention still falls within the scope of protection required by the present invention.

Claims

1. A method for estimating the compressive strength of fiber composite materials based on 3D printing, comprising the following specific steps, characterized in that: Step 1: Sample data collection, determine different process parameter combinations, use a 3D printer to prepare fiber composite material samples according to the determined process parameters, perform compression tests on the prepared samples, and record relevant information of the samples; Step 2: Data preprocessing: sorting the collected raw data, sorting the material properties, printing parameters and compression strength data of the samples accordingly to form a data set; Step 3, construct a recurrent neural network, process the sequence data through the recurrent neural network, capture the long-term dependencies in the data to estimate the compressive strength of the fiber composite material; Step 4: Model training optimization: Use the training set to train the constructed recurrent neural network model. During the training process, use the validation set to verify the performance of the model. Step 5: Model evaluation and verification: Use the test set to evaluate and verify the constructed recurrent neural network model, and release the final version of the fiber composite material compressive strength estimation model; Step 6, compressive strength estimation, when estimating the compressive strength of a new 3D printed fiber composite material sample, the sample material properties and printing parameters are used as input, and the trained model is used to predict the input sample information to obtain the estimated value of the sample's compressive strength.

2. The method for estimating the compressive strength of fiber composite materials based on 3D printing according to claim 1, characterized in that: The process of sample data collection in step 1 is expressed as: Step 1.1, determine different combinations of process parameters, including printing temperature T, printing speed v, layer thickness h, fiber type F, fiber content C, and matrix material M, set the printing temperature between 180℃-250℃, the printing speed between 20mm / s-100mm / s, and the layer thickness between 0.1mm-0.5mm; Step 1.2: According to the determined process parameters, use a 3D printer to prepare fiber composite material samples to ensure the consistency and reliability of the samples. Let the sample number be i. For the i-th sample, the process parameters in its preparation process are T i 、v i 、h i 、F i , C i 、M i ; Step 1.3, perform compression test on the prepared samples, obtain the compression strength data of each sample through the electronic tensile machine, and record the relevant information of the test samples; let the compression strength of the i-th sample be σ i During the compression test, the compression strength data of each sample, the corresponding process parameters and sample number information are recorded. The compression strength σ i It can be calculated by the formula: Among them, F max is the maximum load that the sample is subjected to during compression, and A is the cross-sectional area of ​​the sample.

3. The method for estimating the compressive strength of fiber composite materials based on 3D printing according to claim 1, characterized in that: The data preprocessing process in step 2 is as follows: Step 2.1: Sort the collected raw data. Let the data set D be: D={(x1,σ1),(x2,σ2),…(x n ,s n )} Where n is the number of samples, σ i represents the compression strength of the i-th sample, x i Represents the feature vector of the i-th sample, including material properties and printing parameters: x i =[T i ,v i ,h i ,F i ,C i ,M i ] Among them, T i is the printing temperature of the ith sample, v i is the printing speed of the ith sample, h i is the layer thickness of the ith sample, F i is the fiber type of the i-th sample, C i is the fiber content of the ith sample, M i is the matrix material of the ith sample; Step 2.2, normalize the feature data in the data set, and use the minimum-maximum normalization method to normalize the data.

4. The method for estimating the compressive strength of fiber composite materials based on 3D printing according to claim 1, characterized in that: The process of building a recurrent neural network in step 3 is expressed as follows: Step 3.1, construct the input layer of the network. The number of nodes in the input layer corresponds to the number of features of the sample, which is set to 6; Step 3.2, the hidden layer has m LSTM units. There are multiple gate structures inside the LSTM unit to control the flow of information. The formula of the forget gate is as follows: f t =σ(W f ·[h t-1 ,x t ]+b f ) Among them, f t is the output of the forget gate at time t, W f is the weight matrix of the forget gate, h t-1 is the hidden state at the previous moment, x t is the input at the current moment, b f is the bias vector of the forget gate, σ is the activation function, and t is the time step index; The input gate formula is as follows: i t =σ(W i ·[h t-1 ,x t ]+b i ) g t =σ(W g ·[h t-1 ,x t ]+b g ) Among them, i t is the output of the input gate at time t, g t is the output of the input node after the activation function at time t, W i , W g is the weight matrix of the input gate and input node, b i 、b g is the bias vector; The cell state update formula is as follows: C t =f t ⊙C t-1 +i t ⊙g t Among them, C t is the cell state at time t, ⊙ represents element-wise multiplication; The output gate formula is as follows: The t =σ(W o ·[h t-1 ,x t ]+b o ) h t =O t ⊙tanh(C t ) Among them, O t is the output of the output gate at time t, W o is the weight matrix of the output gate, b o is the bias vector, h t is the hidden state at the current moment; Step 3.3, the number of output layer nodes is 1, and the predicted compression strength is output; Step 3.4, select the activation function. The activation function inside the LSTM unit uses the ReLU function, and the output layer uses the linear activation function; Step 3.5, divide the data set, divide the preprocessed data set into training set, validation set and test set according to a certain ratio, use 70% of the data set as training set, 10% of the data set as validation set, and the remaining data as test set.

5. The method for estimating the compressive strength of fiber composite materials based on 3D printing according to claim 1, characterized in that: The process of model training optimization in step 4 is as follows: Step 4.1, use the training set to train the constructed recurrent neural network model. For the recurrent neural network, at each time step t, calculate the error between the predicted output and the actual output, and use the mean square error as the loss function. The loss function is: in, is the model's predicted compression strength value for the i-th training sample, y i is the true compression strength value of the i-th training sample, N train is the number of samples in the training set after partitioning; Step 4.2, during the training process, adjust the weights and biases of the model through the back-propagation algorithm; In step 4.3, the performance of the model is verified using the validation set. According to the results of the validation set, the hyperparameters of the model are adjusted, including the learning rate, the number of hidden layer nodes, and the regularization parameters.

Images