Multi-component mixing optimization method of polypropylene composite material

Abstract

The invention provides a multi-component mixing optimization method of a polypropylene composite material, and relates to the technical field of material optimizing.The method comprises the steps that a multi-component formula is obtained, the multi-component formula is conveyed to a multi-layer co-extrusion device through independently-controlled feeding channels, real-time melt parameters are collected through an online monitoring module, and the melt parameters are obtained through a multi-layer co-extrusion device; real-time melt parameters are input into the intelligent control module, the real-time melt parameters and target performance parameters are compared and analyzed based on the intelligent control module, and according to a performance comparison result, a process optimization instruction is dynamically generated through a digital twinborn model, process parameters are adjusted in real time, and the gradient distribution structure of the polypropylene composite material is optimized. The technical problems that in the prior art, mixing of multi-component components is not accurate, the gradient structure of the polypropylene composite material is difficult to optimize, and then comprehensive performance is difficult to effectively improve are solved. The technical effects of accurately regulating and controlling the mixing of the multi-component components, optimizing the gradient structure of the polypropylene composite material and improving the comprehensive performance are achieved.

Description

Technical Field

[0001] This invention relates to the field of materials optimization technology, and specifically to a method for optimizing the mixing of multiple components in a polypropylene composite material. Background Technology

[0002] With the widespread application of high-performance plastic materials in the automotive, electronics, electrical appliances, and packaging industries, polypropylene composites have become important structural and functional materials in industrial production due to their lightweight, excellent processability, and functional modification capabilities. However, existing technologies often suffer from uneven mixing and poor dispersion of multiple components, leading to complex internal structures and uneven performance distribution, which limits the improvement of the composite's overall performance, such as conductivity, mechanical properties, and interfacial stability. Furthermore, the design and optimization of gradient composite structures are challenging. Traditional processes struggle to achieve precise spatial control of components such as conductive layers, reinforcing layers, and interfacial control layers, making it difficult to balance functional gradients and overall performance across different application requirements. In addition, the performance optimization of composite materials often relies on extensive experimental trial and error, lacking efficient theoretical guidance and process matching schemes, resulting in long development cycles, high costs, and poor repeatability. Especially in high-performance applications, it is impossible to guarantee synergistic optimization of multiple indicators such as conductivity, mechanical strengthening, and interfacial bonding.

[0003] Existing technologies suffer from technical problems such as inaccurate mixing of multi-component components and difficulty in optimizing the gradient structure of polypropylene composite materials, which in turn makes it difficult to effectively improve the overall performance. Summary of the Invention

[0004] The purpose of this application is to provide a method for optimizing the mixing of multiple components in polypropylene composites, which solves the technical problems of inaccurate mixing of multiple components and difficulty in optimizing the gradient structure of polypropylene composites, thus making it difficult to effectively improve the overall performance.

[0005] In view of the above problems, this application provides a method for optimizing the mixing of multiple components of polypropylene composite materials. The method includes: obtaining a multi-component formulation, wherein the multi-component formulation includes at least a conductive network layer component, a matrix reinforcement layer component, and an interface control layer component; conveying the multi-component materials corresponding to the multi-component formulation to a multilayer co-extrusion device through independently controlled feeding channels; collecting real-time melt parameters in the mixing channel of the multilayer co-extrusion device using an online monitoring module installed thereon, wherein the real-time melt parameters include at least component concentration, viscosity, and interlayer interface stress; inputting the real-time melt parameters into an intelligent control module, wherein the intelligent control module has a pre-trained digital twin model and target performance parameters built in; comparing and analyzing the real-time melt parameters with the target performance parameters based on the intelligent control module, and dynamically generating process optimization instructions through the digital twin model according to the performance comparison results; and adjusting the process parameters in real time according to the process optimization instructions to optimize the gradient distribution structure of the polypropylene composite material.

[0006] Optionally, the conductive network layer component includes conductive filler and dispersant, the matrix reinforcement layer component includes polypropylene substrate and toughening modifier, and the interface control layer component includes compatibilizer and peel modifier.

[0007] Optionally, the independently controlled feeding channel adopts temperature zone regulation, wherein the feeding temperature of the conductive network layer component is lower than that of the matrix reinforcement layer component, and the feeding temperature of the interface regulation layer component is between that of the conductive network layer component and the matrix reinforcement layer component.

[0008] Optionally, a training dataset containing historical process parameters, historical melt parameters, and historical finished product performance parameters is obtained. The historical process parameters include historical feed channel parameters, historical gradient allocation parameters, and historical temperature parameters. The historical melt parameters include historical component concentration data, historical viscosity data, and historical interlayer interface stress data. An initial digital twin model is constructed, comprising a melt parameter analysis module and a finished product performance evaluation module. The historical process parameters are input into the melt parameter analysis module, and the obtained predicted melt parameters are compared with the corresponding historical melt parameters to calculate a first prediction error. The predicted melt parameters are input into the finished product performance evaluation module, and the obtained predicted finished product performance parameters are compared with the corresponding historical finished product performance parameters to calculate a second prediction error. Based on the first and second prediction errors, reinforcement learning is performed on the initial digital twin model, and the prediction and comparison process is repeated until the combined error is less than a preset threshold. The trained digital twin model is then deployed to the intelligent control module.

[0009] Optionally, component concentration data, viscosity data, and interlayer interface stress data are extracted from the real-time melt parameters; the component concentration data, viscosity data, and interlayer interface stress data are compared with the corresponding indicators in the target performance parameters to calculate the differences, generating component concentration deviation, viscosity deviation, and interlayer interface stress deviation; based on the dynamic weights of the component concentration deviation, viscosity deviation, and interlayer interface stress deviation, a comprehensive performance comparison result including the degree of deviation of each parameter is generated.

[0010] Optionally, the comprehensive performance comparison results and the current process parameters are input into the digital twin model; backpropagation calculations are performed through the digital twin model to obtain the process parameter adjustment amount that can reduce performance deviation; the process parameter adjustment amount is converted into a process optimization instruction that includes feed channel adjustment parameters, gradient distribution adjustment parameters, and temperature adjustment parameters.

[0011] Optionally, the comprehensive performance comparison results are input into the melt parameter analysis module of the digital twin model; based on the deviation between the component concentration gradient change rate and the target gradient distribution in the real-time melt parameters, a melt flow-performance response surface model is established through the melt parameter analysis module; according to the melt flow-performance response surface model and combined with the finished product performance evaluation results of the finished product performance evaluation module, the feed channel adjustment parameters that make the component concentration gradient change rate approach the target gradient distribution are calculated; based on the deviation between the interlayer interface stress data and the target interface bonding strength in the real-time melt parameters, the gradient distribution adjustment parameters for optimizing interface bonding are calculated through the collaborative calculation of the melt parameter analysis module and the finished product performance evaluation module; based on the deviation between the viscosity data and the target flow performance in the real-time melt parameters and the dynamic weights in the comprehensive performance comparison results, the temperature adjustment parameters for optimizing melt flow are calculated.

[0012] Optionally, based on the feeding channel adjustment parameters in the process optimization instruction, the material conveying rate of each feeding channel is adjusted to control the input ratio of the multi-component material; based on the gradient distribution adjustment parameters in the process optimization instruction, the flow channel distribution ratio in the multi-layer co-extrusion device is adjusted to optimize the interlayer distribution of the multi-component material; based on the temperature adjustment parameters in the process optimization instruction, the temperature distribution of the mixing flow channel is adjusted to regulate the melting state and interfacial bonding of the multi-component material.

[0013] Optionally, the optimized real-time melt parameters are obtained through an online monitoring module; the component concentration data in the optimized real-time melt parameters are compared with the target gradient distribution for the second time; when the result of the second comparison shows that the component concentration data has not reached the target gradient distribution, a new process optimization instruction is generated; according to the new process optimization instruction, the process parameters are adjusted again until the optimized real-time melt parameters reach the target performance parameters.

[0014] Optionally, the multi-layer co-extrusion device is a variable configuration co-extrusion device with an adaptive flow channel topology, whose flow channel geometry parameters can be dynamically adjusted according to process requirements.

[0015] One or more technical solutions provided in this application have at least the following technical effects or advantages: The method provided in this application involves obtaining a multi-component formulation, which includes at least a conductive network layer component, a matrix reinforcement layer component, and an interface control layer component. The multi-component materials corresponding to the formulation are then fed into a multi-layer co-extrusion device through independently controlled feeding channels. Within the mixing channel of the multi-layer co-extrusion device, real-time melt parameters are collected using an online monitoring module. These real-time melt parameters include at least component concentration, viscosity, and interlayer interface stress. The real-time melt parameters are input into an intelligent control module, which has a pre-trained digital twin model and target performance parameters. Based on the intelligent control module, the real-time melt parameters are compared and analyzed with the target performance parameters. Based on the performance comparison results, process optimization instructions are dynamically generated through the digital twin model. According to the process optimization instructions, process parameters are adjusted in real-time to optimize the gradient distribution structure of the polypropylene composite material. This achieves the technical effect of precisely controlling the mixing of multi-components, optimizing the gradient structure of the polypropylene composite material, and improving its overall performance.

[0016] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating a multi-component mixing optimization method for polypropylene composite materials provided in this application.

[0019] Figure 2This is a schematic diagram of the structure for training a digital twin model in a multi-component mixing optimization method for polypropylene composite materials provided in this application. Detailed Implementation

[0020] This application provides a method for optimizing the mixing of multiple components in polypropylene composites, addressing the technical problems of inaccurate mixing of multiple components and difficulty in optimizing the gradient structure of polypropylene composites, which ultimately hinders the effective improvement of overall performance. The method achieves precise control of the mixing of multiple components, optimizes the gradient structure of polypropylene composites, and enhances overall performance.

[0021] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. It should also be noted that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, not all of them.

[0022] like Figure 1 , Figure 2 As shown, this application provides a method for optimizing the mixing of multiple components in a polypropylene composite material. The method includes: A multi-component formulation is obtained, wherein the multi-component formulation comprises at least a conductive network layer component, a matrix reinforcement layer component, and an interface regulation layer component.

[0023] Furthermore, the conductive network layer component includes conductive filler and dispersant, the matrix reinforcement layer component includes polypropylene substrate and toughening modifier, and the interface control layer component includes compatibilizer and peel modifier.

[0024] Specifically, based on the material design requirements and performance targets of polypropylene composite materials, a multi-component formulation of polypropylene composite materials is obtained. The multi-component formulation is divided into at least three categories according to the functional layering design logic: conductive network layer components, matrix reinforcement layer components, and interface regulation layer components.

[0025] The conductive network layer component is used to construct continuous or semi-continuous conductive pathways within the polypropylene composite material. It is specifically composed of conductive fillers and dispersing agents. The conductive fillers are preferably carbon black, carbon nanotubes, or graphene, used to form conductive pathways within the material. The dispersing agents reduce the tendency for agglomeration between fillers, improving the dispersion uniformity of the conductive fillers in the molten polypropylene system, thereby avoiding agglomeration and stabilizing the conductive network structure. The matrix reinforcement layer component uses the polypropylene matrix as the main framework of the composite material, providing basic mechanical and processing properties. It also improves the impact resistance and ductility of the polypropylene matrix by introducing toughening modifiers, such as ethylene-octene copolymers, which interact with the polypropylene matrix to form a more uniform structure and improve the material's impact resistance. An interface control layer is set between different functional layers. It contains compatibilizers and release modifiers. The compatibilizers enhance the interfacial bonding strength between the conductive layer and the substrate layer through chemical or physical actions. For example, maleic anhydride-grafted polypropylene can form good compatibility with polypropylene and at the same time control the interfacial adhesion strength with the insulating layer. The release modifiers are used to finely control the interlayer adhesion ability to avoid performance imbalance caused by excessively strong or weak interfaces, thereby achieving a controllable interfacial synergistic effect.

[0026] By accurately obtaining the multi-component formulation containing conductive network layer components, matrix reinforcement layer components, and interface regulation layer components, a clear material composition basis is provided for subsequent material transportation, mixing optimization, and the final acquisition of high-performance polypropylene composite materials, thereby achieving more efficient and precise performance optimization.

[0027] The multi-component materials corresponding to the multi-component formulation are respectively fed to the multi-layer co-extrusion unit through independently controlled feeding channels.

[0028] Furthermore, the independently controlled feeding channel adopts temperature zone regulation, wherein the feeding temperature of the conductive network layer component is lower than that of the matrix reinforcement layer component, and the feeding temperature of the interface regulation layer component is between that of the conductive network layer component and the matrix reinforcement layer component.

[0029] Furthermore, the multi-layer co-extrusion device is a variable configuration co-extrusion device with an adaptive flow channel topology, and its flow channel geometric parameters can be dynamically adjusted according to process requirements.

[0030] Specifically, for the conductive network layer component, matrix reinforcement layer component, and interface control layer component in the multi-component formulation, independently controlled feeding channels are set up respectively. Based on the multi-component formulation, the corresponding multi-component materials are respectively fed into the independently controlled feeding channels of the multi-layer co-extrusion device for conveying. Each feeding channel is structurally isolated from each other and independently controlled to avoid non-target mixing or performance interference between different functional components before entering the co-extrusion zone. Each independently controlled feeding channel is equipped with an independent temperature adjustment unit and adopts a temperature zone control method to set differentiated feeding temperature ranges according to the melting characteristics and functional requirements of different components. The conductive network layer component contains conductive fillers and dispersants. Excessive temperature can easily cause filler migration or network collapse, affecting its conductivity. Therefore, its feeding temperature is controlled in a relatively low range to maintain the spatial stability of the conductive filler, such as 160-180℃. The matrix reinforcement layer component contains polypropylene matrix and toughening modifier. It needs to be fully melted to ensure good flowability and mechanical continuity. Its feeding temperature is set in a higher range, such as 190-210℃. The interface control layer component contains compatibilizer and release regulator. It also undertakes the functions of interface wetting and stress buffering. Its feeding temperature is between that of the conductive network layer component and the matrix reinforcement layer component to ensure that the compatibilizer is activated without damaging the conductive structure, such as 180-190℃. By adopting temperature zoning control, it helps each component reach a suitable physical state before entering the co-extrusion unit.

[0031] Each component material is conveyed to the multi-layer co-extrusion unit through an independently controlled feeding channel. The multi-layer co-extrusion unit is a variable configuration co-extrusion unit with an adaptive flow channel topology. The adaptive flow channel topology means that the flow channel can automatically adjust its shape and size according to the characteristics of different component materials and process requirements. The variable configuration emphasizes its flexibility. The geometric parameters such as the width, length and flow angle of the internal flow channel can be adjusted in real time by the actuator, thereby dynamically changing the layer thickness ratio, contact sequence and interface morphology of each component melt according to process requirements, ensuring the uniformity of material flow in the flow channel and avoiding material accumulation or poor flow.

[0032] By combining independent feeding with zoned temperature control, it is ensured that different functional components are in their optimal melting and rheological state before entering the multi-layer co-extrusion zone, thereby achieving precise control of polypropylene composites during the multi-component mixing process and improving the effectiveness and stability of multi-component mixing optimization of polypropylene composites.

[0033] Within the mixing channel of the multilayer co-extrusion device, real-time melt parameters are collected using an online monitoring module installed thereon. These real-time melt parameters include at least component concentration, viscosity, and interlayer interface stress.

[0034] Specifically, the mixing channel of the multi-layer co-extrusion device is the area where the various component materials are fully integrated to form a composite material. An online monitoring module is integrated in the mixing channel. The online monitoring module is an embedded sensor used to continuously, non-contactly or micro-invasively detect the multi-component materials in the molten state and obtain real-time melt parameters.

[0035] The online monitoring module includes at least an infrared spectral sensor, an online viscometer, and an interface stress sensing unit. The infrared spectral sensor achieves online inversion of the relative content of different components in the flow channel cross section by real-time analysis of the intensity of characteristic absorption peaks, thereby obtaining component concentration data. The online viscometer calculates the apparent viscosity in real time by monitoring the flow response of the melt under controlled shear or vibration conditions, which is used to characterize the rheological stability and processability of the melt. The interlayer interface stress is obtained by mechanical sensors deployed in the interlayer contact area, which work based on the piezoelectric effect or strain gauge principle, to reflect the intensity of the interfacial interaction caused by the difference in flow rate and viscosity of different functional layer melts during co-extrusion.

[0036] Furthermore, an online monitoring module installed within the mixing channel collects real-time melt parameters, including at least component concentration, viscosity, and interlayer interface stress. By collecting these melt parameters in real time, the dynamic changes of the material within the mixing channel can be monitored promptly. If any parameters deviate from the preset range, the process parameters of the co-extrusion unit, such as temperature, pressure, and flow rate, can be quickly adjusted to ensure uniform mixing of all components and good interlayer bonding, thereby guaranteeing the stable performance and high quality of the final polypropylene composite material.

[0037] The real-time melt parameters are input into the intelligent control module, which has a pre-trained digital twin model and target performance parameters built in.

[0038] Furthermore, the training of the digital twin model includes: acquiring a training dataset containing historical process parameters, historical melt parameters, and historical finished product performance parameters, wherein the historical process parameters include historical feed channel parameters, historical gradient allocation parameters, and historical temperature parameters, and the historical melt parameters include historical component concentration data, historical viscosity data, and historical interlayer interface stress data; constructing an initial digital twin model, the initial digital twin model including a melt parameter analysis module and a finished product performance evaluation module; inputting the historical process parameters into the melt parameter analysis module, and comparing the obtained predicted melt parameters with the corresponding historical melt parameters to calculate a first prediction error; inputting the predicted melt parameters into the finished product performance evaluation module, and comparing the obtained predicted finished product performance parameters with the corresponding historical finished product performance parameters to calculate a second prediction error; based on the first prediction error and the second prediction error, performing reinforcement learning on the initial digital twin model, and repeatedly executing the prediction and comparison process until the combined error is less than a preset threshold, and deploying the trained digital twin model to the intelligent control module.

[0039] Specifically, the melt parameters will be input into the intelligent control module in real time by the online monitoring module. The intelligent control module has a pre-trained digital twin model and target performance parameters corresponding to the application requirements. The digital twin model is used to reproduce the material flow, interface evolution and performance formation mechanism in the actual co-extrusion process in virtual space with high fidelity. The target performance parameters serve as evaluation benchmarks to constrain and guide the direction of process optimization.

[0040] The training process of the digital twin model is driven by multi-source historical data. First, long-term operating data of existing multi-layer co-extrusion production lines are collected and organized to obtain a training dataset containing historical process parameters, historical melt parameters, and historical finished product performance parameters. During the historical production process, the equipment control system automatically records the feeding channel parameters, gradient distribution parameters, and temperature parameters of each batch, forming traceable historical process parameter data. At the same time, the online monitoring module deployed in the mixing channel continuously collects and stores historical melt parameters such as component concentration, viscosity, and interlayer interface stress under the corresponding operating conditions. After the product is formed, the finished product performance parameters of the corresponding batch are obtained through offline detection methods, such as mechanical property testing, electrical property testing, and interface peel strength testing. These parameters are then timestamped and batch number-associated with the process parameters and melt parameters of the current batch to construct a training dataset containing historical process parameters, historical melt parameters, and historical finished product performance parameters. The historical process parameters specifically include historical feeding channel parameters, historical gradient distribution parameters, and historical temperature parameters. The historical feeding channel parameters specifically refer to parameters such as material conveying rate, instantaneous feeding ratio, and feeding pressure of each independent feeding channel in historical production batches. The historical gradient distribution parameters refer to parameters such as the target layer thickness ratio, actual layer thickness distribution, flow channel distribution coefficient, and interlayer sequence configuration of different component melts in each layer within the multilayer co-extrusion unit. The historical temperature parameters include data such as the temperature of each feeding channel temperature zone, used to characterize different process conditions.

[0041] The historical melt parameters include historical component concentration data, historical viscosity data, and historical interlayer interface stress data. The historical component concentration data refers to the actual mass fraction, volume fraction, and distribution changes along the flow direction of each functional component within the flow channel cross-section or layer, obtained through online spectroscopy or component inversion techniques. The historical viscosity data refers to the apparent viscosity or viscosity fluctuation amplitude of the melt under different shear rates and temperature zones. The historical interlayer interface stress data refers to the stress change characteristics generated during the co-extrusion and parallel flow processes of different functional layer melts, used to reflect the process state. The historical finished product performance parameters reflect the conductivity, mechanical properties, and interfacial stability of the final material, mainly obtained through offline testing. The conductivity parameters include volume resistivity, surface resistivity, and conductive network stability decay index, used to evaluate the continuity and effectiveness of the conductive network layer. The mechanical property parameters include tensile strength, tensile modulus, elongation at break, and notched impact strength, used to reflect the contribution of the matrix reinforcement layer to the structural load-bearing capacity. The interfacial stability parameters include interlayer peel strength, interlayer crack propagation energy, and interface retention rate under thermal aging or cyclic loading, used to comprehensively evaluate the reliability of multilayer structures under long-term service conditions.

[0042] An initial digital twin model is constructed on an industrial simulation and algorithm platform with multiphysics modeling and data-driven modeling capabilities, such as COMSOL, ANSYS Polyflow combined with TensorFlow or PyTorch. The construction process is based on the principle of combining physical mechanism constraints with data-driven fitting. The specific steps include: establishing a virtual process framework consistent with the actual equipment structure in the simulation platform according to the multi-layer co-extrusion process flow; modeling key units such as independently controlled feeding channels, mixing channels, and co-extrusion dies to provide space and boundary conditions for the simulation of melt flow and component distribution; setting historical process parameters as the input interface of the initial digital twin model; setting historical melt parameters and historical finished product performance parameters as supervisory outputs; and introducing neural networks or regression networks as core mapping units on the basis of the virtual process to achieve multi-layer nonlinear fitting from process parameters to melt behavior and finished product performance, thereby forming the initial digital twin model. The digital twin model includes a melt parameter analysis module and a finished product performance evaluation module. The melt parameter analysis module uses historical feed channel parameters, historical gradient distribution parameters, and historical temperature parameters as input vectors to predict the melt state through mechanistic constraint modeling and data-driven compensation. Specifically, it includes: establishing a virtual flow channel model based on a multi-layer co-extrusion process; simulating the transport, mixing, and diffusion behavior of each component in the mixing channel; considering viscosity differences, temperature gradients, and flow resistance of different components during the simulation; predicting the concentration distribution of each component within the flow channel using computational fluid dynamics methods; and evaluating interfacial stresses between different functional layers, including interlayer shear stress and normal stress, to determine the interlayer bonding state. A non-Newtonian flow model is used to predict local viscosity changes in the melt, and the rheological properties of the melt are calculated in conjunction with the temperature distribution within the flow channel. To improve prediction accuracy, the prediction results of the mechanism model are compared with historical online monitoring data, and a residual neural network is constructed for correction. The residual neural network is a multi-layer feedforward network structure, including an input layer, two to three hidden layers, each with 32 to 64 nodes, the activation function is ReLU, and an output layer. The input is the joint vector of the mechanism prediction output and historical process parameters, and the output is the prediction residual. By superimposing the residual and the mechanism prediction value, a melt parameter prediction result that is closer to the actual working condition is obtained.

[0043] The finished product performance evaluation module takes the predicted melt parameters output by the melt parameter analysis module as input, extracts features from them, and forms a comprehensive feature vector. This comprehensive feature vector includes key indicators such as the average concentration gradient of each component, melt viscosity changes and fluctuations, and peak and average interlayer interface stress. Based on this comprehensive feature vector, the module predicts the key performance indicators of the final molded material, including conductivity, mechanical properties, and interface stability. Conductivity includes volume resistivity; mechanical properties include tensile strength, elongation at break, and impact strength; and interface stability includes interlayer peel strength. To achieve high-precision prediction, the finished product performance evaluation module uses a multi-output neural network for training. The multi-output neural network structure includes an input layer for inputting the predicted melt parameters, two to three hidden layers (each with 32 to 64 nodes), and the ReLU activation function. The output layer corresponds to each performance indicator. The training process uses historical finished product performance data as supervision information, comparing the predicted output with the actual performance data. The loss function is defined as the weighted sum of the mean square errors of each performance indicator. The network weights are iteratively updated using gradient descent or Adam optimization algorithms until the prediction error reaches the preset accuracy.

[0044] Historical process parameters are batch-input into the melt parameter analysis module, which generates corresponding predicted melt parameters, including component concentration distribution, melt viscosity, and interlayer interface stress. These predicted melt parameters are then compared with historical melt parameters collected under the same operating conditions. By calculating the deviation between the predicted and measured values, a first prediction error, characterizing the accuracy of the melt state prediction, is obtained. Next, the predicted melt parameters are passed as intermediate input to the finished product performance evaluation module. This module outputs corresponding predicted finished product performance parameters and compares them with historical finished product performance parameters to calculate a second prediction error reflecting the accuracy of the finished product performance prediction.

[0045] The first and second prediction errors are normalized, for example, by using z-score standardization, to eliminate the influence of differences in the dimensions and numerical ranges of different physical quantities on the training process. Then, the two types of errors are weighted and fused according to preset weights to construct a unified combined error index, which is used to comprehensively characterize the overall deviation level of the digital twin model in both process prediction and performance prediction. The preset weights can be set according to actual needs, such as setting the weights of the first and second prediction errors to 0.4 and 0.6, respectively. The combined error index is used as a reward / penalty signal to enhance iterative optimization. When the combined error decreases, a positive reward is given, and when the combined error increases, a penalty signal is applied. This signal is then passed back to the model parameter update units corresponding to the melt parameter analysis module and the finished product performance evaluation module. The network weights of the melt parameter analysis module and the finished product performance evaluation module are jointly adjusted through gradient descent or policy gradient. After each round of weight update, the historical process parameter input, melt parameter prediction, and finished product performance prediction processes are re-executed to continuously monitor the changing trend of the combined error, forming a closed-loop iterative training process, thereby guiding the digital twin model to gradually approach the real working conditions in multiple rounds of iteration. When the combined error is less than the preset threshold in multiple consecutive iterations, the digital twin model is determined to have reached a stable convergence state, and the trained digital twin model is solidified and deployed to the intelligent control module. The preset threshold is set according to actual production needs and accuracy requirements, for example, the combined error is controlled within ±1%.

[0046] Through a well-trained digital twin model, the intelligent control module can accurately predict the real-time melt parameters of the multi-layer co-extrusion melt in the runner, and based on the real-time melt parameters, quickly and accurately predict the finished product performance parameters that may be generated under the current process conditions, providing reliable and accurate data support for real-time process optimization, thereby improving the effectiveness and accuracy of multi-component mixing optimization of polypropylene composite materials.

[0047] Based on the intelligent control module, the real-time melt parameters are compared and analyzed with the target performance parameters, and process optimization instructions are dynamically generated through the digital twin model according to the performance comparison results.

[0048] Furthermore, based on the intelligent control module, the real-time melt parameters are compared and analyzed with the target performance parameters, including: extracting component concentration data, viscosity data, and interlayer interface stress data from the real-time melt parameters; calculating the differences between the component concentration data, viscosity data, and interlayer interface stress data and the corresponding indicators in the target performance parameters to generate component concentration deviation, viscosity deviation, and interlayer interface stress deviation; and generating a comprehensive performance comparison result including the degree of deviation of each parameter based on the dynamic weights of the component concentration deviation, viscosity deviation, and interlayer interface stress deviation.

[0049] Specifically, the real-time melt parameters obtained by the online monitoring module are input into the intelligent control module, and key indicators that can directly reflect the formation mechanism of material structure and performance are extracted from the real-time melt parameters, including component concentration data for characterizing the spatial distribution of multi-component components, viscosity data for reflecting melt flow and processability, and interlayer interface stress data for characterizing the synergistic state of different functional layers.

[0050] The extracted component concentration data, viscosity data, and interlayer interface stress data are compared and analyzed with the corresponding indices in the pre-set target performance parameters to identify the degree of deviation between the current operating condition and the target state. These target performance parameters are determined comprehensively based on material application requirements, historical optimal process data, and mechanistic constraints. First, according to the final application scenario of the polypropylene composite material, such as conductivity level, mechanical strength requirements, and long-term interface stability requirements, the product technical specifications or industry standards provide the lower limit and optimal range of performance. Then, stable operating conditions that simultaneously meet the requirements of conductivity, mechanical properties, and interface stability in actual operation are screened from historical production data. The corresponding melt parameters are statistically analyzed to extract the component concentration gradient range, viscosity range, and interlayer interface stress safety threshold as empirical optimal references. Simultaneously, combined with production process capabilities, the empirical optimal references are boundary-checked and their rationality corrected, eliminating parameter ranges that, while meeting a single performance requirement, pose a risk of processing instability or interface failure. The final target performance parameters are stored in the intelligent control module in the form of a multi-index target vector, serving as a benchmark for real-time melt parameter comparison and analysis.

[0051] The component concentration data is compared with the target component concentration data in the target performance parameters to obtain the component concentration deviation, which characterizes the degree of deviation of the component distribution. That is, component concentration deviation = actual component concentration - target component concentration. Similarly, the viscosity data is compared with the target viscosity data to obtain the viscosity deviation, and the interlayer interface stress data is compared with the target interface stress data to obtain the interlayer interface stress deviation. Based on the sensitivity of different deviations to the final performance, weights are dynamically allocated to weight and fuse the component concentration deviation, viscosity deviation, and interlayer interface stress deviation, generating a comprehensive performance comparison result that includes the degree of deviation of each parameter. The comprehensive performance comparison result = w1 × component concentration deviation + w2 × viscosity deviation + w3 × interlayer interface stress deviation, where w1, w2, and w3 are the weights of component concentration deviation, viscosity deviation, and interlayer interface stress deviation, respectively, and are set according to actual needs. For example, when producing composite materials with extremely high strength requirements, the weight of component concentration deviation may be set higher, with w1 set to 0.5 and w2 and w3 to 0.25 respectively. When producing materials with high requirements for molding uniformity, the weight of viscosity deviation is set even higher. Through dynamic weight allocation, a comprehensive performance comparison result that includes the degree of deviation of each parameter is generated, which can more comprehensively and accurately reflect the difference between the current melt state and the target state, thereby improving the consistency and stability of the multilayer structure of polypropylene composite materials through targeted adjustments.

[0052] Furthermore, based on the performance comparison results, process optimization instructions are dynamically generated through the digital twin model, including: inputting the comprehensive performance comparison results and the current process parameters into the digital twin model; performing backpropagation calculations through the digital twin model to obtain process parameter adjustment amounts that can reduce performance deviations; and converting the process parameter adjustment amounts into process optimization instructions that include feed channel adjustment parameters, gradient distribution adjustment parameters, and temperature adjustment parameters.

[0053] Furthermore, by performing backpropagation calculations through the digital twin model, process parameter adjustments that can reduce performance deviations are obtained, including: inputting the comprehensive performance comparison results into the melt parameter analysis module of the digital twin model; establishing a melt flow-performance response surface model through the melt parameter analysis module based on the deviation between the component concentration gradient change rate and the target gradient distribution in the real-time melt parameters; calculating the feed channel adjustment parameters that make the component concentration gradient change rate approach the target gradient distribution based on the melt flow-performance response surface model and the finished product performance evaluation results from the finished product performance evaluation module; calculating the gradient distribution adjustment parameters that optimize interface bonding through collaborative calculations by the melt parameter analysis module and the finished product performance evaluation module based on the deviation between the interlayer interface stress data and the target interface bonding strength in the real-time melt parameters; and calculating the temperature adjustment parameters that optimize melt flow based on the deviation between the viscosity data and the target flow performance in the real-time melt parameters and the dynamic weights in the comprehensive performance comparison results.

[0054] Specifically, the comprehensive performance comparison results and the current process parameters are input into the digital twin model. The comprehensive performance comparison results include key information such as component concentration deviation, viscosity deviation and interlayer interface stress deviation, reflecting the gap between the current melt state and the target performance parameters. The current process parameters include the feeding rate of each feeding channel, the flow channel distribution ratio of the gradient distributor in the multi-layer co-extrusion unit and the set temperature of each temperature zone, which are the control variables in the actual production process.

[0055] The digital twin model performs backpropagation calculations through its trained network structure to optimize performance by reducing overall performance deviation. It determines the direction and magnitude of adjustments to the process parameters most sensitive to performance. Specifically, the overall performance comparison results are first input to the melt parameter analysis module of the digital twin model. Based on the deviation relationship between the component concentration gradient change rate and the target gradient distribution in the real-time melt parameters, the melt parameter analysis module constructs a melt flow-performance response surface model reflecting the relationship between feeding conditions, channel structure, and melt distribution response. This melt flow-performance response surface model can be represented as a multidimensional mapping relationship: G=f(F1,F2…Fi,R1,R2…Ri), where G represents the component concentration gradient change rate, Fi represents the feeding rate of each independent feeding channel, and Ri represents the distribution ratio parameter of different channels in the multi-layer co-extrusion unit. The melt flow-performance response surface model is approximated by the trained neural network in the digital twin model, enabling it to reflect the influence of changes in feeding conditions and channel structure on the melt component distribution in real time. Based on the melt flow-performance response surface model, the difference between the currently calculated component concentration gradient change rate and the target gradient change rate corresponding to the target gradient distribution is obtained to obtain the gradient deviation. This deviation is then used as the optimization target input to the melt flow-performance response surface model. By taking the partial derivative of the response surface model with respect to the rate of each feeding channel, the sensitivity of the gradient change rate to the parameters of each feeding channel is obtained. Combined with the performance prediction results output by the finished product performance evaluation module, an iterative optimization method is used to assign higher weights to the feeding channels that contribute more to the gradient improvement. This allows the calculation of the feeding channel adjustment parameters that gradually bring the component concentration gradient change rate closer to the target gradient distribution. The finished product performance evaluation results provide quantitative feedback on the actual performance of the product, which is used to more accurately determine the adjustment direction and magnitude of the feeding channel parameters.

[0056] Simultaneously, for the interlayer interface stress data in the real-time melt parameters, the melt parameter analysis module first compares the current interface stress distribution with the stress safety range corresponding to the target interface bonding strength to obtain the interface stress deviation distribution characteristics. Based on this, combined with the flow channel topology parameters of the multi-layer co-extrusion unit, a correlation model between interface stress and flow channel allocation ratio is established to describe the stress response characteristics of different functional layer melts at the flow channel confluence location. The finished product performance evaluation module simultaneously provides interface stability prediction results to determine the impact of interface stress changes on the final material's delamination risk and long-term stability. Through collaborative calculations between the melt parameter analysis module and the finished product performance evaluation module, the contribution of interface stress deviation at different flow channel locations is analyzed in reverse, identifying the flow channel allocation parameters that have the most significant impact on interface stress. Based on this, the corresponding flow channel allocation ratio adjustment is calculated, forming gradient allocation adjustment parameters to improve the synergistic state of different functional layer interfaces, gradually reducing the interface stress back to the safety range corresponding to the target bonding strength. Furthermore, for the viscosity data in the real-time melt parameters, the melt parameter analysis module compares the current melt viscosity value with the target viscosity range corresponding to the target flow performance, calculates the viscosity deviation, and establishes the response relationship between viscosity and temperature range setting by combining the influence law of temperature on melt flow. Simultaneously, the intelligent control module dynamically calculates the weight coefficients of each deviation index based on the relative magnitudes of component concentration deviation, viscosity deviation, and interlayer interface stress deviation in the comprehensive performance comparison results. This reflects the priority of different performance targets under the current operating conditions. Specifically, the component concentration deviation, viscosity deviation, and interlayer interface stress deviation in the comprehensive performance comparison results are first normalized to eliminate the influence of differences in the dimensions and numerical ranges of different parameters on the weight calculation. Then, based on the relative deviation degree of each deviation index under the current operating conditions, a dynamic weight allocation rule is constructed. This involves combining each normalized deviation value with its sensitivity coefficient under historical stable operating conditions to calculate the corresponding weight coefficient. The larger the deviation amplitude and the higher the sensitivity of the index to the final product performance, the larger its weight coefficient. When the proportion of viscosity deviation in the overall deviation increases, the intelligent control module automatically increases the weighting coefficient corresponding to viscosity deviation, thereby amplifying the effect of temperature regulation on overall performance improvement in subsequent optimization calculations. When component concentration deviation or interlayer interface stress deviation dominates, the weighting coefficient of viscosity deviation is correspondingly reduced, and a constraint threshold is set for the temperature adjustment range to avoid adverse effects of temperature changes on gradient distribution or interfacial bonding. By determining the weighting coefficient based on the relative magnitude of the deviation and performance sensitivity, adaptive priority switching of different performance targets under different operating conditions is achieved.

[0057] The weighting coefficient of viscosity deviation under the current operating conditions is determined based on a dynamic weighting mechanism, and this weighting coefficient is introduced into the melt viscosity-temperature response model to characterize the effective contribution of temperature change to the improvement of melt flowability. The melt viscosity-temperature response model refers to a mathematical or data-driven mapping relationship established by combining experimental data with a mechanistic model. It is used to characterize the influence of different temperature setting zones on melt viscosity. First, melt viscosity data is collected under different temperature conditions through multi-point melt rheological experiments or online monitoring modules to obtain temperature-viscosity curves and local flow responses in each temperature zone. At the same time, the rheological mechanism of polypropylene and its composite components, such as the generalized Carreau model or Arrhenius-type temperature dependence, is combined to fit or regress the experimental data to obtain a response function that outputs viscosity with temperature change. In the digital twin framework, the melt viscosity-temperature response model is used to predict the melt viscosity change at any temperature setting zone and provides a basis for the sensitivity analysis and optimization of temperature adjustment parameters. Using the set temperature of each temperature zone as the independent variable, the melt viscosity-temperature response model is subjected to local perturbation or partial derivative calculations to obtain the sensitivity distribution of melt viscosity changes to temperature variations in different temperature zones. Combining the sensitivity results with dynamic weighting coefficients, the intelligent control module prioritizes temperature adjustments for each temperature zone and, under the premise of meeting equipment safety and process stability constraints, calculates the optimal temperature adjustment amount for each zone. This ensures that the weighted viscosity deviation gradually decreases without triggering secondary fluctuations in component distribution and interfacial stress. Finally, the temperature adjustment amount is output as a temperature adjustment parameter, achieving synergistic optimization of melt flowability and processing stability, and ensuring that the temperature adjustment process is coordinated with feeding and flow channel adjustments to avoid mutual interference between process parameters. Finally, the feeding channel adjustment parameters, gradient distribution adjustment parameters, and temperature adjustment parameters are integrated to form a process optimization command that includes feeding rate adjustment, gradient distributor valve opening adjustment within the multi-layer co-extrusion unit, and extrusion temperature setting adjustment.

[0058] Based on the comparison of real-time melt parameters and target performance, the intelligent control module dynamically generates process optimization instructions through backpropagation calculation using a digital twin model. This enables real-time monitoring and intelligent optimization of the production process, allowing for timely adjustment of process parameters to adapt to changes in the production process and improving the performance and production efficiency of polypropylene composite materials.

[0059] According to the process optimization instructions, the process parameters are adjusted in real time to optimize the gradient distribution structure of the polypropylene composite material.

[0060] Furthermore, based on the process optimization instructions, process parameters are adjusted in real time to optimize the gradient distribution structure of the polypropylene composite material, including: adjusting the material conveying rate of each feeding channel based on the feeding channel adjustment parameters in the process optimization instructions to control the input ratio of the multi-component materials; adjusting the flow channel distribution ratio in the multi-layer co-extrusion device based on the gradient distribution adjustment parameters in the process optimization instructions to optimize the interlayer distribution of the multi-component materials; and adjusting the temperature distribution of the mixing flow channel based on the temperature adjustment parameters in the process optimization instructions to regulate the melting state and interfacial bonding of the multi-component materials.

[0061] Specifically, based on the feeding channel adjustment parameters in the process optimization instructions, the material conveying rate of each feeding channel is adjusted. That is, by controlling the input ratio of the conductive network layer component, the matrix reinforcement layer component, and the interface control layer component, the precise supply of multi-component materials and the initial gradient construction are achieved. The fine adjustment range of the feeding channel rate can be controlled within ±2% to 5% to balance flow stability and gradient response speed. The material conveying rate of each feeding channel is adjusted to the set value, thereby accurately controlling the input ratio of multi-component materials.

[0062] Based on the gradient distribution adjustment parameters in the process optimization instructions, the distribution ratio of each flow channel in the multilayer co-extrusion unit is adjusted. This is achieved by adjusting the interlayer distribution of different functional layer materials in the mixing channel through variable configuration flow channels or distributor valve openings. This optimizes the relative thickness and uniformity of the conductive network layer, matrix reinforcement layer, and interface control layer, thereby controlling the gradient structure and interfacial bonding state of the final material. The adjustment range can be ±3% to 10% of the flow distribution variation. Based on the temperature adjustment parameters in the process optimization instructions, the set temperatures of each temperature zone in the mixing channel are adjusted to achieve precise control over the melting state, viscosity, and interfacial bonding force of the multi-component materials. This ensures a balance between melt flowability and interfacial stability. The temperature fine-tuning range can be set to ±5~10°C to quickly respond to melt flow deviations, ensuring that the temperature in the mixing channel quickly and accurately reaches the set value. This allows the multi-component materials to fully melt at the appropriate temperature and form a good interfacial bond, further improving the performance of the composite material.

[0063] Adjusting process parameters such as feeding channels, flow channel distribution, and temperature distribution according to process optimization instructions can effectively optimize the gradient distribution structure of polypropylene composites. This gradient distribution structure refers to the continuous or graded distribution of physical and chemical properties of different functional layer components, including conductive network layer components, matrix reinforcement layer components, and interface control layer components, along the thickness or transverse direction during multi-layer co-extrusion of polypropylene composites. Specifically, it manifests as the concentration gradient, viscosity gradient, and interfacial bonding strength gradient of each layer component. By controlling the material input ratio of each feeding channel, the flow channel distribution ratio, and the mixing flow channel temperature, the conductive filler, matrix polypropylene, and compatibilizer can be continuously arranged in the final material according to predetermined proportions, positions, and thicknesses. This ensures the controllability and uniformity of the overall performance of the polypropylene composite while improving its comprehensive properties, such as conductivity, mechanical properties, and interfacial bonding performance, further enhancing the production efficiency and product quality stability of polypropylene composites.

[0064] Furthermore, the method also includes: acquiring optimized real-time melt parameters through an online monitoring module; performing a secondary comparison between the component concentration data in the optimized real-time melt parameters and the target gradient distribution; when the secondary comparison result shows that the component concentration data has not reached the target gradient distribution, regenerating a new process optimization instruction; and adjusting the process parameters again according to the new process optimization instruction until the optimized real-time melt parameters reach the target performance parameters.

[0065] Specifically, after initial process optimization and adjustment, an online monitoring module installed in the mixing channel collects real-time melt parameters to obtain optimized real-time melt parameters, including key parameters such as component concentration, viscosity, and interlayer interface stress. After obtaining the real-time melt parameters, the component concentration data is compared a second time with a pre-set target gradient distribution. The target gradient distribution, based on the performance requirements and application scenarios of polypropylene composite materials, is a pre-set concentration or proportion distribution range that each functional layer component should achieve in the thickness or transverse direction, determined through extensive experiments and theoretical analysis, and is used to guide subsequent process corrections.

[0066] By calculating the deviation between the component concentration data in the optimized real-time melt parameters and the target gradient distribution, the degree of conformity between the current melt state and the target gradient is determined. When the deviation exceeds the allowable range, the deviation information obtained from the secondary comparison is used as input. Combined with the digital twin model, the real-time melt parameters are predicted and analyzed, and new process optimization instructions are regenerated, including adjustment of the feed channel conveying rate, optimization of the flow channel allocation ratio, and correction of temperature distribution. The allowable range refers to the tolerance range in which the deviation between the optimized real-time melt parameters and the target gradient distribution is judged to be acceptable during the secondary comparison process. Through experiments and simulations, the influence of different component concentration deviations on conductivity, mechanical properties, and interface stability is studied to identify the upper and lower limits of the deviation's impact on the final performance within the acceptable range. Alternatively, by calculating the mean and standard deviation of historical concentration data, the allowable deviation is set to the mean ± k times the standard deviation, such as 2 times the standard deviation or ±5%, to balance process controllability and performance stability.

[0067] Based on the new process optimization instructions, the process parameters are adjusted again to further refine the gradient distribution structure. This secondary closed-loop operation is repeated until the optimized real-time melt parameters are essentially matched with the target performance parameters, ensuring that the gradient distribution structure meets the design requirements in the final product. Through the secondary comparison and iterative optimization mechanism, the process parameters can be continuously adjusted until the optimized real-time melt parameters reach the target performance parameters, ensuring that the produced polypropylene composite material meets the design requirements and improving the product's pass rate and stability.

[0068] In summary, the multi-component mixing optimization method for polypropylene composite materials provided in this application has the following technical effects: By acquiring a multi-component formulation including a conductive network layer, a matrix reinforcement layer, and an interface control layer, and feeding it to a multi-layer co-extrusion unit via an independent feeding channel, real-time melt parameters such as component concentration, viscosity, and interlayer interface stress are collected using an online monitoring module. These parameters are then input into a pre-trained digital twin model for comparative analysis, thereby dynamically generating process optimization instructions. Based on these instructions, process parameters are adjusted in real time to achieve precise control over the multi-component mixing process and the gradient distribution structure of the polypropylene composite material. This improves the overall performance of the polypropylene composite material, including its conductivity, mechanical properties, and interfacial bonding properties, as well as the product qualification rate and production stability.

[0069] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0070] Obviously, those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for optimizing the mixing of multiple components in a polypropylene composite material, characterized in that, The method includes: A multi-component formulation is obtained, wherein the multi-component formulation comprises at least a conductive network layer component, a matrix reinforcement layer component, and an interface regulation layer component; The multi-component materials corresponding to the multi-component formulation are respectively fed to the multi-layer co-extrusion unit through independently controlled feeding channels; In the mixing channel of the multilayer co-extrusion device, real-time melt parameters are collected using an online monitoring module installed thereon. The real-time melt parameters include at least component concentration, viscosity, and interlayer interface stress. The real-time melt parameters are input into the intelligent control module, which has a pre-trained digital twin model and target performance parameters built in. Based on the intelligent control module, the real-time melt parameters are compared and analyzed with the target performance parameters, and process optimization instructions are dynamically generated through the digital twin model according to the performance comparison results. According to the process optimization instructions, the process parameters are adjusted in real time to optimize the gradient distribution structure of the polypropylene composite material.

2. The method for optimizing the mixing of multiple components in a polypropylene composite material as described in claim 1, characterized in that, The conductive network layer component includes conductive fillers and dispersants; the matrix reinforcement layer component includes a polypropylene substrate and a toughening modifier; and the interface control layer component includes a compatibilizer and a release modifier.

3. The method for optimizing the mixing of multiple components in a polypropylene composite material as described in claim 1, characterized in that, The independently controlled feeding channel adopts temperature zone regulation, wherein the feeding temperature of the conductive network layer component is lower than that of the matrix reinforcement layer component, and the feeding temperature of the interface regulation layer component is between that of the conductive network layer component and the matrix reinforcement layer component.

4. The method for optimizing the mixing of multiple components in a polypropylene composite material as described in claim 1, characterized in that, The training of the digital twin model includes: Obtain a training dataset containing historical process parameters, historical melt parameters, and historical finished product performance parameters. The historical process parameters include historical feed channel parameters, historical gradient allocation parameters, and historical temperature parameters. The historical melt parameters include historical component concentration data, historical viscosity data, and historical interlayer interface stress data. Construct an initial digital twin model, which includes a melt parameter analysis module and a finished product performance evaluation module; The historical process parameters are input into the melt parameter analysis module, and the obtained predicted melt parameters are compared with the corresponding historical melt parameters to calculate the first prediction error. The predicted melt parameters are input into the finished product performance evaluation module, and the obtained predicted finished product performance parameters are compared with the corresponding historical finished product performance parameters to calculate the second prediction error. Based on the first prediction error and the second prediction error, the initial digital twin model is reinforced and the prediction and comparison process is repeated until the combined error is less than a preset threshold. The trained digital twin model is then deployed to the intelligent control module.

5. The method for optimizing the mixing of multiple components in a polypropylene composite material as described in claim 4, characterized in that, Based on the intelligent control module, the real-time melt parameters are compared and analyzed with the target performance parameters, including: Extract component concentration data, viscosity data, and interlayer interface stress data from the real-time melt parameters; The component concentration data, viscosity data and interlayer interface stress data are respectively compared with the corresponding indicators in the target performance parameters to calculate the differences, generating component concentration deviation, viscosity deviation and interlayer interface stress deviation. Based on the dynamic weights of the component concentration deviation, viscosity deviation, and interlayer interface stress deviation, a comprehensive performance comparison result including the degree of deviation of each parameter is generated.

6. The method for optimizing the mixing of multiple components in a polypropylene composite material as described in claim 5, characterized in that, Based on the performance comparison results, process optimization instructions are dynamically generated using the digital twin model, including: The comprehensive performance comparison results and the current process parameters are input into the digital twin model; By performing backpropagation calculations using the digital twin model, the amount of process parameter adjustment that can reduce performance deviation can be obtained. The process parameter adjustment amounts are converted into process optimization instructions that include feed channel adjustment parameters, gradient distribution adjustment parameters, and temperature adjustment parameters.

7. The method for optimizing the mixing of multiple components in a polypropylene composite material as described in claim 6, characterized in that, Backpropagation calculations are performed using the digital twin model to obtain process parameter adjustments that can reduce performance deviations, including: The comprehensive performance comparison results are input into the melt parameter analysis module of the digital twin model; Based on the deviation between the component concentration gradient change rate and the target gradient distribution in the real-time melt parameters, a melt flow-performance response surface model is established through the melt parameter analysis module. Based on the melt flow-performance response surface model and the finished product performance evaluation results from the finished product performance evaluation module, the feed channel adjustment parameters that make the component concentration gradient change rate approach the target gradient distribution are calculated. Based on the deviation between the interlayer interface stress data and the target interface bonding strength in the real-time melt parameters, the gradient allocation adjustment parameters for optimizing interface bonding are calculated through the collaborative calculation of the melt parameter analysis module and the finished product performance evaluation module. Based on the deviation between the viscosity data in the real-time melt parameters and the target flow performance, and taking into account the dynamic weights in the comprehensive performance comparison results, the temperature adjustment parameters for optimizing melt flow are calculated.

8. The method for optimizing the mixing of multiple components in a polypropylene composite material as described in claim 1, characterized in that, According to the process optimization instructions, process parameters are adjusted in real time to optimize the gradient distribution structure of the polypropylene composite material, including: Based on the feeding channel adjustment parameters in the process optimization instruction, the material conveying rate of each feeding channel is adjusted to control the input ratio of the multi-component materials; Based on the gradient distribution adjustment parameters in the process optimization instruction, the flow channel distribution ratio in the multi-layer co-extrusion unit is adjusted to optimize the interlayer distribution of the multi-component material; Based on the temperature adjustment parameters in the process optimization instructions, the temperature distribution of the mixing channel is adjusted to regulate the melting state and interfacial bonding of the multi-component materials.

9. The method for optimizing the mixing of multiple components in a polypropylene composite material as described in claim 1, characterized in that, The method further includes: Optimized real-time melt parameters are obtained through an online monitoring module; The component concentration data in the optimized real-time melt parameters are compared with the target gradient distribution in a second step. When the secondary comparison results show that the component concentration data has not reached the target gradient distribution, a new process optimization instruction is regenerated. According to the new process optimization instructions, the process parameters are adjusted again until the optimized real-time melt parameters reach the target performance parameters.

10. The method for optimizing the mixing of multiple components in a polypropylene composite material as described in claim 1, characterized in that, The multi-layer co-extrusion device is a variable configuration co-extrusion device with an adaptive flow channel topology, and its flow channel geometric parameters can be dynamically adjusted according to process requirements.

Images