PC / GF composite material and preparation method thereof

Abstract

According to the PC / GF composite material and the preparation method thereof, working state parameters such as the temperature, the pressure and the rotating speed of a double-screw extruder are monitored and collected in real time through a sensor group, and a data processing and analysis algorithm is introduced at the rear end to carry out time sequence collaborative correlation analysis on the working state parameters; therefore, whether the current extrusion state is stable or not is judged. Therefore, the dynamic change of the working state of the double-screw extruder can be effectively monitored, and abnormal conditions can be found in time, so that a basis is provided for optimization and adjustment of the extrusion process, and in this way, real-time monitoring and automatic control of the extrusion process can be realized, so that stable operation of the extrusion process is ensured, and the production efficiency is improved. And the melt is ensured to have proper viscosity and flowability.

Description

Technical Field

[0001] The present application relates to the field of intelligent preparation, and more specifically, to a PC / GF composite material and a preparation method thereof. Background Art

[0002] PC / GF composite material is a composite material composed of polycarbonate (PC) and glass fiber (GF). PC has excellent mechanical properties, heat resistance and electrical insulation properties, while GF has high strength and rigidity. The combination of PC and GF can give full play to the advantages of both, and obtain high-performance engineering plastics with excellent mechanical properties, heat resistance and flame retardant properties, which are widely used in the fields of automobiles, aerospace, electronics, etc.

[0003] The key to preparing PC / GF composites is to ensure the full dispersion and combination between PC and GF, as well as the stability of the material's melt extrusion and molding process. However, during the preparation of PC / GF composites, the working state of the twin-screw extruder has an important influence on the quality and fluidity of the melt. If the extrusion state is unstable, it will lead to uneven viscosity and fluidity of the melt, thereby affecting the injection molding effect and the performance of the composite material. Therefore, how to monitor and control the working state of the twin-screw extruder in real time to ensure the stable operation of the extrusion process is a key issue in the preparation of PC / GF composites.

[0004] Therefore, a preparation scheme for PC / GF composite materials is desired. Summary of the invention

[0005] In order to solve the above-mentioned technical problems, the present application is proposed. The embodiment of the present application provides a PC / GF composite material and a preparation method thereof, which monitors and collects the working state parameters such as temperature, pressure, and rotation speed of a twin-screw extruder in real time through a sensor group, and introduces data processing and analysis algorithms at the back end to perform time-series collaborative correlation analysis of these working state parameters, so as to judge whether the current extrusion state is stable. In this way, the dynamic changes in the working state of the twin-screw extruder can be effectively monitored, and abnormal conditions can be discovered in time, thereby providing a basis for the optimization and adjustment of the extrusion process. In this way, real-time monitoring and automatic control of the extrusion process can be achieved to ensure the stable operation of the extrusion process and ensure that the melt has suitable viscosity and fluidity.

[0006] According to another aspect of the present application, a method for preparing a PC / GF composite material is provided, comprising:

[0007] The polycarbonate, the glass fiber, the grafted elastomer, the flame retardant and the anti-dripping agent are stirred and mixed in a high-speed mixer to obtain a PC / GF mixture;

[0008] Melting and extruding the PC / GF mixture through a twin-screw extruder to obtain a molten PC / GF mixture;

[0009] The molten PC / GF mixture is injection molded through a mold to obtain a PC / GF composite material.

[0010] Compared with the prior art, the PC / GF composite material and its preparation method provided by the present application monitor and collect the working state parameters such as temperature, pressure, and rotation speed of the twin-screw extruder in real time through a sensor group, and introduce data processing and analysis algorithms at the back end to perform time-series collaborative correlation analysis of these working state parameters, so as to judge whether the current extrusion state is stable. In this way, the dynamic changes of the working state of the twin-screw extruder can be effectively monitored, and abnormal conditions can be discovered in time, thereby providing a basis for optimizing and adjusting the extrusion process. In this way, real-time monitoring and automatic control of the extrusion process can be achieved to ensure the stable operation of the extrusion process and ensure that the melt has suitable viscosity and fluidity. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0012] Figure 1 is a flow chart of a method for preparing a PC / GF composite material according to an embodiment of the present application;

[0013] Figure 2 It is a system architecture diagram of a method for preparing a PC / GF composite material according to an embodiment of the present application;

[0014] Figure 3 Flow chart of sub-step S2 of the method for preparing the PC / GF composite material according to an embodiment of the present application;

[0015] Figure 4 Flow chart of sub-step S27 of the method for preparing the PC / GF composite material according to an embodiment of the present application. DETAILED DESCRIPTION

[0016] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0017] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0018] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0019] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0020] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0021] In the technical solution of the present application, a method for preparing a PC / GF composite material is proposed. Figure 1 Flow chart of the preparation method of PC / GF composite material according to the embodiment of the present application. Figure 1 As shown, the preparation method of the PC / GF composite material according to the embodiment of the present application includes the following steps: S1, stirring and mixing the polycarbonate, the glass fiber, the grafted elastomer, the flame retardant and the anti-dripping agent in a high-speed mixer to obtain a PC / GF mixture; S2, melt-extrude the PC / GF mixture through a twin-screw extruder to obtain a molten PC / GF mixture; S3, injection molding the molten PC / GF mixture through a mold to obtain a PC / GF composite material.

[0022] In particular, in S1, the polycarbonate, the glass fiber, the grafted elastomer, the flame retardant and the anti-dripping agent are stirred and mixed in a high-speed mixer to obtain a PC / GF mixture. It should be understood that during the stirring and mixing process, chemical reactions or physical interactions may occur between the various components, further changing the properties and characteristics of the mixture.

[0023] In particular, in S2, the PC / GF mixture is melt-extruded through a twin-screw extruder to obtain a molten PC / GF mixture. In particular, in a specific example of the present application, Figure 2 4 is a system architecture diagram of a method for preparing a PC / GF composite material according to an embodiment of the present application. Figure 3 FIG. 1 is a flow chart of sub-step S2 of the method for preparing the PC / GF composite material according to an embodiment of the present application. Figure 2 and Figure 3 As shown, the S2 includes: S21, obtaining the time series of the working state parameters of the twin-screw extruder collected by the sensor group within a predetermined time period, wherein the working state parameters include extrusion temperature value, extrusion pressure value and speed value; S22, performing data regularization on the time series of the working state parameters according to the parameter sample dimension to obtain an extrusion temperature time series input vector, an extrusion pressure time series input vector and a speed time series input vector; S23, calculating the sample covariance association matrix of the extrusion temperature time series input vector relative to the extrusion pressure time series input vector to obtain an extrusion temperature-extrusion pressure time series association matrix; S24, extracting the extrusion temperature-extrusion pressure time series association feature of the extrusion temperature-extrusion pressure by an extrusion temperature-extrusion pressure time series association feature extractor based on a first deep neural network model. S25, extracting features from the timing correlation matrix to obtain an extrusion temperature-extrusion pressure timing correlation feature diagram; S26, using a speed timing feature extractor based on a second deep neural network model to extract features from the speed timing input vector to obtain a speed timing correlation feature vector; S26, using a channel-by-channel responsive transfer association module to process the speed timing correlation feature vector and the extrusion temperature-extrusion pressure timing correlation feature diagram to obtain a channel-by-channel responsive multi-parameter timing correlation feature diagram; S27, using an essential feature extraction network to process the channel-by-channel responsive multi-parameter timing correlation feature diagram to obtain a channel-by-channel responsive multi-parameter timing correlation essential feature vector; S28, determining whether the current extrusion state is stable based on the channel-by-channel responsive multi-parameter timing correlation essential feature vector.

[0024] Specifically, the S21 obtains the time series of the working state parameters of the twin-screw extruder collected by the sensor group within a predetermined time period, wherein the working state parameters include extrusion temperature value, extrusion pressure value and speed value. It should be understood that by monitoring and recording the working state parameters of the twin-screw extruder, key parameters such as temperature, pressure, speed, etc. in the extrusion process can be monitored in real time. This helps to promptly discover and correct abnormal conditions in the production process and improve the quality control level of the product. In a specific example of the present application, the sensor group consists of a temperature sensor, a pressure sensor and a speed sensor. It is worth mentioning that the temperature sensor is a device for measuring the temperature of an environment or an object. It can convert temperature into an electrical signal or a digital signal for applications such as monitoring, control and data recording. A pressure sensor is a device for measuring pressure. It can convert pressure into an electrical signal or a digital signal for applications such as monitoring, control and data recording. A speed sensor is a device for measuring the rotational speed or speed of an object. It can convert the speed into an electrical signal or a digital signal for applications such as monitoring, control and data recording.

[0025] Specifically, in S22, the time series of the working state parameters are data regularized according to the parameter sample dimension to obtain the extrusion temperature timing input vector, the extrusion pressure timing input vector and the speed timing input vector. Considering that the working state parameters include the extrusion temperature value, the extrusion pressure value and the speed value, and these working state parameter items have a dynamic change law of timing in the time dimension, in order to be able to perform a timing collaborative analysis on the working state parameters of the twin-screw extruder to determine whether the current extrusion state is stable, it is necessary to perform data regularization on the time series of the working state parameters according to the parameter sample dimension to obtain the extrusion temperature timing input vector, the extrusion pressure timing input vector and the speed timing input vector.

[0026] Specifically, the S23, calculates the sample covariance association matrix of the extrusion temperature time series input vector relative to the extrusion pressure time series input vector to obtain the extrusion temperature-extrusion pressure time series association matrix. It should be understood that in the process of melt extrusion, extrusion temperature and extrusion pressure are two important process parameters, which have an important impact on the performance and quality of the final product. The extrusion temperature determines the melting state and fluidity of PC and GF, while the extrusion pressure reflects the shear force and resistance in the extrusion process. Therefore, in order to analyze the time series correlation between extrusion temperature and extrusion pressure, to explore their interaction and influence in the extrusion process, in the technical solution of the present application, further calculate the sample covariance association matrix of the extrusion temperature time series input vector relative to the extrusion pressure time series input vector to obtain the extrusion temperature-extrusion pressure time series association matrix. Wherein, the covariance association matrix can reflect the time series correlation and change trend between the two variables of extrusion temperature and extrusion pressure. By analyzing this association matrix, the correlation degree, change law and possible mutual influence between extrusion temperature and extrusion pressure can be understood, and valuable information is provided for monitoring, control and optimization of extrusion process. More specifically, the sample covariance correlation matrix of the extrusion temperature timing input vector relative to the extrusion pressure timing input vector is calculated using the following sample covariance formula to obtain the extrusion temperature-extrusion pressure timing correlation matrix; wherein the sample covariance formula is:

[0027] M cov =W T XX T W

[0028] Wherein, W is the extrusion temperature timing input vector, X is the extrusion pressure timing input vector, M cov is the extrusion temperature-extrusion pressure time series correlation matrix.

[0029] Specifically, in S24, feature extraction is performed on the extrusion temperature-extrusion pressure time series correlation matrix through an extrusion temperature-extrusion pressure time series correlation feature extractor based on a first deep neural network model to obtain an extrusion temperature-extrusion pressure time series correlation feature graph. In particular, the first deep neural network model is a convolutional neural network model. That is, the extrusion temperature-extrusion pressure time series correlation matrix is ​​subjected to feature mining in an extrusion temperature-extrusion pressure time series correlation feature extractor based on a convolutional neural network model to extract the time series collaborative correlation feature information between the extrusion temperature and the extrusion pressure, thereby obtaining an extrusion temperature-extrusion pressure time series correlation feature graph. More specifically, each layer of the extrusion temperature-extrusion pressure timing correlation feature extractor based on the convolutional neural network model performs the following operations on the input data in the forward transfer of the layer: convolution processing on the input data to obtain a convolution feature map; pooling based on the local feature matrix on the convolution feature map to obtain a pooled feature map; and nonlinear activation on the pooled feature map to obtain an activation feature map; wherein the output of the last layer of the extrusion temperature-extrusion pressure timing correlation feature extractor based on the convolutional neural network model is the extrusion temperature-extrusion pressure timing correlation feature map, and the input of the first layer of the extrusion temperature-extrusion pressure timing correlation feature extractor based on the convolutional neural network model is the extrusion temperature-extrusion pressure timing correlation matrix.

[0030] Convolutional Neural Network (CNN) is a deep learning model that is specifically designed to process data with a grid structure, such as images and speech. The core idea of ​​CNN is to extract the features of the input data through convolution operations, and to perform high-level representation and abstraction of the features by stacking them layer by layer. The following are the basic components and working principles of CNN: Convolution layer: The convolution layer is the core component of CNN and is used to extract the features of the input data. It performs convolution operations on the input data by applying a set of learnable convolution kernels (filters). The convolution operation can capture local patterns and features in the input data and generate a series of feature maps; Activation function: After the convolution layer, a nonlinear activation function such as ReLU is usually applied. The activation function introduces nonlinear features, allowing the network to learn more complex patterns and representations; Pooling layer: The pooling layer is used to reduce the size and number of parameters of the feature map and extract the most important features. Common pooling operations include maximum pooling and average pooling; Fully connected layer: After a series of convolutional and pooling layers, some fully connected layers are usually added. The fully connected layer converts the feature map of the previous layer into an output result, such as classification or regression; Dropout: In order to prevent overfitting, Dropout technology is often used in CNN. Dropout randomly discards a part of neurons during the training process to reduce the dependency between neurons and improve the generalization ability of the model. Through the back propagation algorithm, CNN can automatically learn to extract features from the input data and optimize according to the training objectives. During the training process, CNN adjusts the network parameters by minimizing the loss function so that the output results are as close to the true labels as possible.

[0031] Specifically, the S25, through the speed time series feature extractor based on the second deep neural network model, the speed time series input vector is feature extracted to obtain the speed time series associated feature vector. In particular, the speed time series feature extractor based on the second deep neural network model is a speed time series feature extractor based on a one-dimensional convolution layer. Considering that in the extrusion process of preparing PC / GF composite materials, the speed is another important process parameter. The speed of the extruder directly affects the shear force, mixing effect and material fluidity during the extrusion process. Therefore, the change in speed can have an important impact on the stability of the extrusion process and the quality of the material. At the same time, the speed value also has a dynamic change law in the time series. In order to explore the time series change characteristics of the speed value, in order to coordinate other working state parameters to comprehensively monitor the current extrusion state, in the technical solution of the present application, the speed time series input vector is further subjected to feature mining in the speed time series feature extractor based on the one-dimensional convolution layer to extract the time series associated feature information of the speed value in the time dimension, thereby obtaining the speed time series associated feature vector.

[0032] The one-dimensional convolutional layer is a basic layer type in convolutional neural networks, which is used to process data with temporal correlation. It applies convolution operations on the one-dimensional space of the input data to extract temporal features. The following is a step-by-step expansion of the one-dimensional convolutional layer: Define the convolution kernel: The convolution kernel is a small weight matrix that slides on the input data and performs a convolution operation; Padding: You can choose to pad the input data. Padding can add zero values ​​on both sides of the input sequence to control the length of the output sequence; Convolution operation: Apply the convolution kernel to each position on the input sequence and perform a convolution operation. In a one-dimensional convolution, the convolution operation is performed by multiplying the convolution kernel with a part of the input sequence element by element and summing the results to obtain an element of the output sequence; Non-linear activation function: Usually, after the convolution operation, a non-linear activation function such as ReLU is applied to introduce non-linear features; Downsampling: You can choose to apply a downsampling operation to reduce the length of the output sequence and extract more important features. Common downsampling operations are maximum pooling and average pooling; Output: After the convolution operation and possible downsampling operations, the output sequence is obtained.

[0033] Specifically, the S26 uses a channel-by-channel responsive transfer association module to process the speed timing association feature vector and the extrusion temperature-extrusion pressure timing association feature diagram to obtain a channel-by-channel responsive multi-parameter timing association feature diagram. It should be understood that in the actual process of extrusion monitoring, there is a complex correlation between the extrusion temperature, extrusion pressure and the speed. The speed timing association feature vector and the extrusion temperature-extrusion pressure timing association feature diagram respectively represent different aspects of information in the extrusion process, wherein the speed timing association feature vector reflects the timing dynamic characteristics of the speed, and the extrusion temperature-extrusion pressure timing association feature diagram reflects the timing synergistic association feature information of the extrusion temperature and the extrusion pressure in the time dimension. Therefore, in order to comprehensively consider the timing correlation between multiple working state parameters, so as to obtain a more comprehensive and accurate feature representation, in the technical solution of the present application, the speed timing association feature vector and the extrusion temperature-extrusion pressure timing association feature diagram are further processed using a channel-by-channel responsive transfer association module to obtain a channel-by-channel responsive multi-parameter timing association feature diagram. By processing through the channel-by-channel responsive transfer association module, the expression capability of the working state characteristics of the extruder can be enhanced by interacting and integrating the speed time series association feature vector and the extrusion temperature-extrusion pressure time series association feature graph channel by channel. This helps to extract more distinguishing and discriminating features, thereby better describing the characteristics and changes of the extrusion process and improving the ability to accurately identify and classify the extrusion process state. More specifically, the extrusion temperature-extrusion pressure timing association feature map is processed in the convolutional neural network model of the parameter-guided cross-modal encoder to obtain the extrusion temperature-extrusion pressure timing association enhanced semantic feature map; the speed timing association feature vector is passed through the first convolution layer of the parameter-guided cross-modal encoder and then linearly corrected by the ReLU function to obtain the linearly corrected speed timing association feature vector; the linearly corrected speed timing association feature vector is passed through the second convolution layer of the parameter-guided cross-modal encoder and then processed by the Sigmoid function to obtain the normalized speed timing association feature vector; the extrusion temperature-extrusion pressure timing association enhanced semantic feature map is weightedly fused along the channel dimension with the normalized speed timing association feature vector to obtain the channel-by-channel response multi-parameter timing association feature map.

[0034] Specifically, in S27, the channel-by-channel response multi-parameter temporal correlation feature map is processed using an essential feature extraction network to obtain a channel-by-channel response multi-parameter temporal correlation essential feature vector. In particular, in a specific example of the present application, Figure 4As shown, S27 includes: S271, optimizing the feature distribution of the channel-by-channel response multi-parameter timing correlation feature map to obtain an optimized channel-by-channel response multi-parameter timing correlation feature map; S272, using an essential feature extraction network to process the optimized channel-by-channel response multi-parameter timing correlation feature map to obtain the channel-by-channel response multi-parameter timing correlation essential feature vector.

[0035] More specifically, the S271 performs feature distribution optimization on the channel-by-channel response multi-parameter timing correlation feature diagram to obtain an optimized channel-by-channel response multi-parameter timing correlation feature diagram. In the above technical solution, each feature matrix of the extrusion temperature-extrusion pressure timing correlation feature diagram expresses the local high-order correlation characteristics of the full time domain covariance correlation of the extrusion temperature and the extrusion pressure, and the channel distribution of the convolutional neural network model is followed between each feature matrix. In this way, when the speed timing correlation feature vector and the extrusion temperature-extrusion pressure timing correlation feature diagram are processed using the channel-by-channel responsiveness transfer association module, the extrusion temperature-extrusion pressure timing correlation feature diagram is transferred by channel-by-channel responsiveness distribution using the local timing correlation characteristics of the speed value expressed by the speed timing correlation feature vector, so that the feature distribution between each feature matrix of the channel-by-channel response multi-parameter timing correlation feature diagram deviates from the channel distribution of the convolutional neural network model, affecting the integrity of the channel distribution.

[0036] Therefore, when using the essential feature extraction network to process the channel-by-channel response multi-parameter temporal correlation feature map, it is expected to improve the distribution integrity between the various feature matrices of the channel-by-channel response multi-parameter temporal correlation feature map, so as to improve the expression effect of the channel-by-channel response multi-parameter temporal correlation essential feature vector.

[0037] Based on this, in the technical solution of the present application, before using the essential feature extraction network to process the channel-by-channel response multi-parameter time series correlation feature map, the channel-by-channel response multi-parameter time series correlation feature map is subjected to feature distribution optimization, and the feature distribution optimization process includes:

[0038] Determine the characteristic mean and characteristic variance of the channel-by-channel response multi-parameter time series association characteristic graph, and divide the characteristic mean by the characteristic variance to obtain a channel-by-channel response multi-parameter time series association probability statistic value:

[0039]

[0040] Among them, μ and σ 2 They respectively represent the characteristic mean and characteristic variance of the channel-by-channel response multi-parameter time series association characteristic graph, and ρ represents the probability statistic value of the channel-by-channel response multi-parameter time series association;

[0041] The channel-by-channel response multi-parameter temporal association feature map is point-multiplied with the inverse of the maximum eigenvalue in the channel-by-channel response multi-parameter temporal association feature map to obtain a channel-by-channel response multi-parameter temporal association probability constraint map:

[0042] F1=F⊙f max -1

[0043] Wherein, F represents the channel-by-channel response multi-parameter temporal correlation characteristic map, f max represents the maximum eigenvalue in the channel-by-channel response multi-parameter time series association feature map, ⊙ represents the position point multiplication, and F1 represents the channel-by-channel response multi-parameter time series association probability constraint map;

[0044] After performing point addition on the channel-by-channel response multi-parameter time series association probability constraint graph and the channel-by-channel response multi-parameter time series association probability statistical value, a logarithmic value with base 2 is calculated to obtain a channel-by-channel response multi-parameter time series association information interaction graph:

[0045]

[0046] Among them, F1 represents the probability constraint diagram of multi-parameter temporal association of each channel response, ρ represents the probability statistic of multi-parameter temporal association of each channel response, It means adding by position point, and F2 means the interactive diagram of multi-parameter time series correlation information of channel-by-channel response;

[0047] After performing point subtraction on the channel-by-channel response multi-parameter time series association probability constraint diagram using the channel-by-channel response multi-parameter time series association probability statistics, the inverse of each eigenvalue is calculated to obtain the channel-by-channel response multi-parameter time series association sequence constraint diagram:

[0048]

[0049] Among them, F1 represents the probability constraint diagram of multi-parameter temporal association of each channel response, ρ represents the probability statistic of multi-parameter temporal association of each channel response, It means subtraction by position point, and F3 means the multi-parameter timing correlation sequence constraint diagram of channel-by-channel response;

[0050] The channel-by-channel response multi-parameter timing association information interaction diagram and the channel-by-channel response multi-parameter timing association sequence constraint diagram are interpolated to obtain an optimized channel-by-channel response multi-parameter timing association feature diagram.

[0051] Accordingly, the probability statistics of the channel-by-channel response multi-parameter time series correlation feature map are used to simulate the mesoscale interaction structure under probability constraints between the feature value scale and the feature map scale of the channel-by-channel response multi-parameter time series correlation feature map, so as to construct a bidirectional hidden variable motif based on the mesoscale short sequence relative to the probability statistics value to perform mesoscale bidirectional migration, and to perform posterior recovery based on the sequence constraints on the interactive information to improve the integrity of the feature representation of the channel-by-channel response multi-parameter time series correlation feature map, thereby improving the expression effect of the channel-by-channel response multi-parameter time series correlation essential feature vector obtained by processing the channel-by-channel response multi-parameter time series correlation feature map using the essential feature extraction network, and improving the accuracy of the classification result obtained by the classifier. In this way, it is possible to more accurately judge whether the current extrusion state is stable, so as to effectively monitor the dynamic changes in the working state of the twin-screw extruder, and timely discover abnormal conditions, providing a basis for the optimization and adjustment of the extrusion process.

[0052] More specifically, in S272, the optimized channel-by-channel response multi-parameter time-series correlation feature map is processed using an essential feature extraction network to obtain the channel-by-channel response multi-parameter time-series correlation essential feature vector. In a complex extrusion process, the multi-parameter time-series correlation feature map may contain a large amount of redundant information and noise, which may not be efficient and accurate enough for subsequent analysis and extrusion state stability judgment. In addition, considering the essential characteristics of the extrusion state, it is of great significance to the state and performance of the extrusion process. Therefore, it is necessary to extract the most representative and discriminative essential features from the multi-parameter time-series correlation feature map by using an essential feature extraction network. Based on this, in order to reduce the impact of deviation from the essential semantic features of the extrusion state on the prototype calculation, in the technical solution of the present application, the channel-by-channel response multi-parameter time-series correlation feature map is further processed using an essential feature extraction network to obtain a channel-by-channel response multi-parameter time-series correlation essential feature vector. By using the essential feature extraction network to process the channel-by-channel response multi-parameter time-series correlation feature map, the time-series synergistic correlation essential feature information between multiple working state parameters in the working process of the twin-screw extruder can be extracted, and the most representative and discriminative key essential features can be automatically learned and extracted. These key essential features can be local patterns, important change points or key time periods in the time series correlation feature graph. In this way, the influence of redundant information and noise can be reduced, the expressiveness and discrimination ability of the features can be improved, and the accurate judgment of the stability of the current extrusion state can be achieved. In a specific example, the essential feature extraction network is used to process the set of reference feature vectors of the surface features of the finished plush toy with the following essential feature extraction formula to obtain the reference essential feature vector of the surface features of the finished plush toy; wherein the essential feature extraction formula is:

[0053]

[0054] Among them, v i and v j are the i-th and j-th optimized channel-by-channel response multi-parameter temporal correlation feature vectors in each feature matrix along the channel dimension of the optimized channel-by-channel response multi-parameter temporal correlation feature map, V k is the set of the optimized channel-by-channel response multi-parameter time series associated feature vectors, ||·||1 is the vector norm, M is the number of vectors in the set of the optimized channel-by-channel response multi-parameter time series associated feature vectors - 1, D i is the eigenvalue of each position in the channel-by-channel multi-parameter temporal correlation essential eigenvector, N s is the length of the channel-by-channel response multi-parameter time series correlation essential characteristic vector, V is the channel-by-channel response multi-parameter time series correlation essential characteristic vector, and exp(·) is an exponential operation.

[0055] Specifically, the S28 determines whether the current extrusion state is stable based on the channel-by-channel response multi-parameter time-series associated essential feature vector. That is, in a specific example of the present application, the channel-by-channel response multi-parameter time-series associated essential feature vector is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether the current extrusion state is stable. In other words, the time-series collaborative associated essential feature information between multiple working state parameters of the twin-screw extruder is used for classification processing to determine whether the current extrusion state is stable. In this way, the dynamic changes in the working state of the twin-screw extruder can be effectively monitored, and abnormal conditions can be discovered in time, thereby providing a basis for the optimization and adjustment of the extrusion process. In this way, real-time monitoring and automatic control of the extrusion process can be achieved to ensure the stable operation of the extrusion process and ensure that the melt has suitable viscosity and fluidity. More specifically, the channel-by-channel response multi-parameter time-series associated essential feature vector is fully connected and encoded using multiple fully connected layers of the classifier to obtain an encoded classification feature vector; and the encoded classification feature vector is passed through the Softmax classification function of the classifier to obtain the classification result.

[0056] That is, in the technical solution of the present application, the label of the classifier includes whether the current extrusion state is stable (first label), and whether the current extrusion state is unstable (second label), wherein the classifier determines to which classification label the channel-by-channel response multi-parameter time series correlation essential feature vector belongs through a soft maximum function. It is worth noting that the first label p1 and the second label p2 here do not contain artificially set concepts. In fact, during the training process, the computer model does not have the concept of "whether the current extrusion state is stable". It only has two classification labels and the probability of the output feature under these two classification labels, that is, the sum of p1 and p2 is one. Therefore, the classification result of whether the current extrusion state is stable is actually converted into a binary class probability distribution that conforms to the laws of nature through the classification label. In essence, the physical meaning of the natural probability distribution of the label is used, rather than the linguistic text meaning of "whether the current extrusion state is stable".

[0057] A classifier is a machine learning model or algorithm that is used to classify input data into different categories or labels. Classifiers are part of supervised learning and perform classification tasks by learning the mapping relationship from input data to output categories.

[0058] A fully connected layer is a common type of layer in neural networks. In a fully connected layer, each neuron is connected to all neurons in the previous layer, and each connection has a weight. This means that each neuron in the fully connected layer receives inputs from all neurons in the previous layer, and performs a weighted sum of these inputs through the weights, and then passes the result to the next layer.

[0059] The Softmax classification function is a commonly used activation function for multi-classification problems. It converts each element of the input vector into a probability value between 0 and 1, and the sum of these probability values ​​is equal to 1. The Softmax function is often used in the output layer of a neural network, especially for multi-classification problems, because it can map the network output to the probability distribution of each category. During the training process, the output of the Softmax function can be used to calculate the loss function and update the network parameters through the back propagation algorithm. It is worth noting that the output of the Softmax function does not change the relative size relationship between the elements, but only normalizes them. Therefore, the Softmax function does not change the characteristics of the input vector, but only converts it into a probability distribution form.

[0060] In particular, in S3, the molten PC / GF mixture is molded by injection molding to obtain a PC / GF composite material. It should be understood that mold injection molding is an efficient production method that can quickly and continuously produce PC / GF composite products.

[0061] In summary, the preparation method of the PC / GF composite material according to the embodiment of the present application is explained, which uses a sensor group to monitor and collect the working state parameters such as temperature, pressure, and speed of the twin-screw extruder in real time, and introduces data processing and analysis algorithms at the back end to perform time-series collaborative correlation analysis of these working state parameters to determine whether the current extrusion state is stable. In this way, the dynamic changes in the working state of the twin-screw extruder can be effectively monitored, and abnormal conditions can be discovered in time, thereby providing a basis for optimizing the adjustment of the extrusion process. In this way, real-time monitoring and automatic control of the extrusion process can be achieved to ensure stable operation of the extrusion process and ensure that the melt has suitable viscosity and fluidity.

[0062] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A PC / GF composite material, characterized in that: include: Polycarbonate, glass fiber, grafted elastomer, flame retardant and anti-drip agent.

2. A method for preparing a PC / GF composite material, characterized in that: include: The polycarbonate, the glass fiber, the grafted elastomer, the flame retardant and the anti-dripping agent are stirred and mixed in a high-speed mixer to obtain a PC / GF mixture; Melting and extruding the PC / GF mixture through a twin-screw extruder to obtain a molten PC / GF mixture; The molten PC / GF mixture is injection molded through a mold to obtain a PC / GF composite material.

3. The method for preparing the PC / GF composite material according to claim 2, characterized in that: The PC / GF mixture is melt-extruded through a twin-screw extruder to obtain a molten PC / GF mixture, comprising: Acquire a time series of working state parameters of the twin-screw extruder within a predetermined time period collected by a sensor group, wherein the working state parameters include an extrusion temperature value, an extrusion pressure value, and a rotation speed value; Regularizing the time series of the working state parameters according to the parameter sample dimensions to obtain an extrusion temperature timing input vector, an extrusion pressure timing input vector, and a rotation speed timing input vector; Calculating a sample covariance correlation matrix of the extrusion temperature time series input vector relative to the extrusion pressure time series input vector to obtain an extrusion temperature-extrusion pressure time series correlation matrix; Extracting features from the extrusion temperature-extrusion pressure time series correlation matrix using an extrusion temperature-extrusion pressure time series correlation feature extractor based on a first deep neural network model to obtain an extrusion temperature-extrusion pressure time series correlation feature graph; Extracting features from the speed time series input vector using a speed time series feature extractor based on a second deep neural network model to obtain a speed time series associated feature vector; Processing the rotation speed timing correlation feature vector and the extrusion temperature-extrusion pressure timing correlation feature diagram using a channel-by-channel responsive transfer correlation module to obtain a channel-by-channel responsive multi-parameter timing correlation feature diagram; Processing the channel-by-channel response multi-parameter temporal correlation feature map using an essential feature extraction network to obtain a channel-by-channel response multi-parameter temporal correlation essential feature vector; Based on the channel-by-channel response multi-parameter time series correlation essential characteristic vector, it is determined whether the current extrusion state is stable.

4. The method for preparing the PC / GF composite material according to claim 3, characterized in that: Calculating the sample covariance correlation matrix of the extrusion temperature timing input vector relative to the extrusion pressure timing input vector to obtain the extrusion temperature-extrusion pressure timing correlation matrix, comprising: calculating the sample covariance correlation matrix of the extrusion temperature timing input vector relative to the extrusion pressure timing input vector using the following sample covariance formula to obtain the extrusion temperature-extrusion pressure timing correlation matrix; wherein the sample covariance formula is: M cov =W T XX T W Wherein, W is the extrusion temperature timing input vector, X is the extrusion pressure timing input vector, M cov is the extrusion temperature-extrusion pressure time series correlation matrix.

5. The method for preparing the PC / GF composite material according to claim 4, characterized in that: The first deep neural network model is a convolutional neural network model, and the speed time series feature extractor based on the second deep neural network model is a speed time series feature extractor based on a one-dimensional convolutional layer.

6. The method for preparing the PC / GF composite material according to claim 5, characterized in that: The speed timing correlation feature vector and the extrusion temperature-extrusion pressure timing correlation feature diagram are processed using a channel-by-channel responsive transfer correlation module to obtain a channel-by-channel responsive multi-parameter timing correlation feature diagram, including: Processing the extrusion temperature-extrusion pressure temporal correlation feature map in a convolutional neural network model of the parameter-guided cross-modal encoder to obtain an extrusion temperature-extrusion pressure temporal correlation enhanced semantic feature map; The speed time series associated feature vector is passed through the first convolution layer of the parameter-guided cross-modal encoder and then linearly corrected by the ReLU function to obtain a linearly corrected speed time series associated feature vector; The linearly corrected speed time series associated feature vector is passed through the second convolution layer of the parameter-guided cross-modal encoder and then processed by a Sigmoid function to obtain a normalized speed time series associated feature vector; The normalized speed time series correlation feature vector is used to perform weighted fusion along the channel dimension on the extrusion temperature-extrusion pressure time series correlation enhanced semantic feature map to obtain the channel-by-channel response multi-parameter time series correlation feature map.

7. The method for preparing the PC / GF composite material according to claim 6, characterized in that: The channel-by-channel response multi-parameter temporal correlation feature map is processed using an essential feature extraction network to obtain a channel-by-channel response multi-parameter temporal correlation essential feature vector, including: Performing feature distribution optimization on the channel-by-channel response multi-parameter time series correlation feature map to obtain an optimized channel-by-channel response multi-parameter time series correlation feature map; The optimized channel-by-channel response multi-parameter temporal correlation feature map is processed using an essential feature extraction network to obtain the channel-by-channel response multi-parameter temporal correlation essential feature vector.

8. The method for preparing the PC / GF composite material according to claim 7, characterized in that: The method uses an essential feature extraction network to process the optimized channel-by-channel response multi-parameter time series correlation feature map to obtain the channel-by-channel response multi-parameter time series correlation essential feature vector, including: using the essential feature extraction network to process the optimized channel-by-channel response multi-parameter time series correlation feature map using the following essential feature extraction formula to obtain the channel-by-channel response multi-parameter time series correlation essential feature vector; Among them, the essential feature extraction formula is: Among them, v i and v j are the i-th and j-th optimized channel-by-channel response multi-parameter temporal correlation feature vectors in each feature matrix along the channel dimension of the optimized channel-by-channel response multi-parameter temporal correlation feature map, V k is the set of the optimized channel-by-channel response multi-parameter time series associated feature vectors, ||·||1 is the vector norm, M is the number of vectors in the set of the optimized channel-by-channel response multi-parameter time series associated feature vectors - 1, D i is the eigenvalue of each position in the channel-by-channel multi-parameter temporal correlation essential eigenvector, N s is the length of the channel-by-channel response multi-parameter time series correlation essential characteristic vector, V is the channel-by-channel response multi-parameter time series correlation essential characteristic vector, and exp(·) is an exponential operation.

9. The method for preparing the PC / GF composite material according to claim 8, characterized in that: Based on the channel-by-channel response multi-parameter time-series correlation essential characteristic vector, determining whether the current extrusion state is stable includes: passing the channel-by-channel response multi-parameter time-series correlation essential characteristic vector through a classifier to obtain a classification result, and the classification result is used to indicate whether the current extrusion state is stable.

Images