A method for optimizing mechanical properties of a nanomodified bio-based composite material

Abstract

The application provides a kind of based on nanometer modified bio-based composite mechanical property optimization method, belongs to the technical field of polymer compound, the application is through the silane coupling agent modification pretreatment to wood powder, prepares functionalized nanoparticle dispersion liquid and uses graph theory model to evaluate dispersion, modified wood powder and polyethylene are proportioned in twin-screw extruder melt blending, simultaneously join functionalized nanoparticle dispersion liquid and utilize intelligent dispersion optimization model to carry out real-time control, through the grid performance index evaluation network structure integrity, uses multi-objective optimization game model to carry out comprehensive optimization to tensile strength, bending strength, impact strength, adjusts nanoparticle addition amount, mixing temperature, mixing time and other process parameters according to optimization result and repeats optimization until reaching preset target value, solves the technical problem that nanometer particle is not evenly dispersed, forms local stress concentration point, affects overall mechanical property.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of polymer compound composition, and particularly relates to a method for optimizing mechanical properties of a nano-modified bio-based composite material. BACKGROUND

[0002] A bio-based composite material is an environmentally friendly composite material filled with plant fibers in a thermoplastic polymer matrix, which has a wide application prospect in the fields of automobile manufacturing, construction engineering, packaging materials, etc. Traditional wood-plastic composites are mainly prepared by mechanically mixing modified wood powder and thermoplastic polymers such as polyethylene in a molten state, through physical coating and interfacial bonding. This method is simple and low in cost, and has been widely used in industrial production.

[0003] In the traditional preparation process, the mechanical properties of the composite material are limited due to the poor interfacial compatibility between the polymer matrix and the plant fibers, especially the tensile strength and impact toughness, which are difficult to meet the requirements of high-performance applications. In order to improve the mechanical properties of the composite material, the existing technology usually adopts the method of adding nano-particles for modification, which enhances the comprehensive performance of the composite material through the enhancement and interface adjustment of nano-particles.

[0004] In the current nano-modification technology, due to the high specific surface area and surface energy of nano-particles, agglomeration phenomenon easily occurs in the polyethylene matrix material, forming micro-scale aggregates. These aggregates form local defect areas in the composite material, which easily form stress concentration points under external force, thereby reducing the overall mechanical properties of the material. That is, the existing technology has the technical problem that nano-particles easily agglomerate in the polyethylene matrix material, causing uneven dispersion in the composite matrix, forming local stress concentration points, and affecting the overall mechanical properties. SUMMARY

[0005] Therefore, the present application provides a method for optimizing the mechanical properties of a nano-modified bio-based composite material, which can solve the technical problem that nano-particles easily agglomerate in the polyethylene matrix material, causing uneven dispersion in the composite matrix, forming local stress concentration points, and affecting the overall mechanical properties.

[0006] The application is implemented in the following manner: a method for optimizing the mechanical properties of a nano-modified bio-based composite material is provided, which comprises the following steps: pretreating wood powder, modifying the surface of the wood powder with a silane coupling agent to obtain modified wood powder; preparing a nano-particle dispersion liquid, performing surface functionalization treatment with ultrasonic treatment and the addition of a titanate coupling agent to obtain a functionalized nano-particle dispersion liquid; evaluating the uniformity of the nano-particles in the functionalized nano-particle dispersion liquid through dispersion degree detection, establishing a graph theory model of the nano-particle dispersion network, taking each nano-particle as a node in the graph and the interaction force between the particles as the edge weight, calculating the optimal dispersion path using the minimum spanning tree algorithm to ensure that the nano-particles are in a good dispersion state; mixing the modified wood powder with a polyethylene matrix material, performing melt blending in a twin-screw extruder to obtain a wood powder-polyethylene mixture; synchronously adding the functionalized nano-particle dispersion liquid during the melt blending process, using in-situ polymerization technology to form a three-dimensional network structure of the nano-particles in the polyethylene matrix material, and simultaneously using an intelligent dispersion optimization model to perform real-time regulation of the spatial distribution of the nano-particles to obtain a nano-modified bio-based composite material; testing the mechanical properties of the prepared nano-modified bio-based composite material, establishing a multi-objective optimization game model to optimize the process parameters; adjusting the nano-particle addition amount, mixing temperature and mixing time parameters according to the calculation results of the multi-objective optimization game model, and repeating the aforementioned steps until the tensile strength, bending strength and impact strength all reach the preset target values.

[0007] The wood powder is pretreated by controlling the particle size of the wood powder to be within the range of 40 to 80 mesh, drying the wood powder at a temperature of 80 to 100°C for 4 to 6 hours to reduce the water content of the wood powder to less than 2% by mass, and then modifying the surface of the wood powder with a silane coupling agent.

[0008] The nano-particle dispersion liquid is prepared by dispersing the nano-particles in ethanol solution at a mass fraction of 0.5% to 3%, continuously dispersing for 30 to 60 minutes using ultrasonic treatment at a power of 300 to 500W, and simultaneously performing surface functionalization treatment with the addition of a titanate coupling agent.

[0009] The dispersion degree is a dimensionless parameter that characterizes the uniform distribution of the nano-particles in the ethanol solution, with a numerical range of 0 to 1, and is calculated by taking the reciprocal of the ratio of the standard deviation of the distance between the nano-particles in a unit volume to the average distance, with a requirement that the numerical value of the dispersion degree be greater than 0.85.

[0010] The requirement that the nano-particles achieve a good dispersion state also includes a requirement for the aggregation degree, which is an evaluation index reflecting the degree of agglomeration of the nano-particles, with a numerical range of 0 to 1, and is determined by calculating the ratio of the number of agglomerates to the total number of particles, with a requirement that the numerical value of the aggregation degree be less than 0.15.

[0011] The step of mixing the modified wood powder with the polyethylene matrix material, specifically mixing the modified wood powder with the polyethylene matrix material in a mass ratio of 30-50:70-50, melt blending in a twin-screw extruder at a temperature of 160-180°C, with a screw speed of 50-80 revolutions per minute, and the mass ratio refers to the total mass fraction of the two being 100.

[0012] After the step of forming a three-dimensional network structure of nanoparticles in the polyethylene matrix material using in-situ polymerization technology, the method further comprises the step of evaluating the network structure integrity by grid performance index, requiring the grid performance index to be greater than 0.75, wherein the grid performance index is a comprehensive index for evaluating the integrity of the three-dimensional network structure of nanoparticles in the polyethylene matrix material, calculated by observing the network connectivity under a scanning electron microscope and combining the conductivity test data, with a value range of 0 to 1, and the grid performance index is required to be greater than 0.75.

[0013] The intelligent dispersion optimization model is a nanoparticle spatial distribution prediction and optimization system based on a graph convolution network architecture, and the model includes an input layer, a graph convolution layer, a message passing layer, and an output layer, wherein the number of message passing steps is dynamically adjusted according to the number of typical aggregation regions of nanoparticle clusters, and the model uses residual connection and batch normalization techniques to improve training stability and convergence speed.

[0014] In the step of testing the mechanical properties of the prepared nanomodified bio-based composite material, the method further comprises the step of calculating a performance gain index to evaluate the nanomodification effect, wherein the performance gain index is a quantitative evaluation parameter for the improvement effect of nanomodification on the mechanical properties of the nanomodified bio-based composite material, calculated by subtracting 1 from the ratio of the material performance after modification to the original material performance, and a positive value indicates performance improvement.

[0015] The multi-objective optimization game model is a mathematical optimization model that considers tensile strength, bending strength, impact strength, cost control, and processing performance, and uses the Pareto optimal theory to establish a balance relationship between the objective functions, and solves the optimal parameter combination through a genetic algorithm.

[0016] The typical aggregation region of nanoparticle clusters is a relatively stable spatial region formed by the aggregation of multiple nanoparticles in the polyethylene matrix material due to van der Waals forces, electrostatic interactions, and surface tension factors, and each typical aggregation region of nanoparticle clusters usually contains 5-50 nanoparticles, with a region diameter range of 100-500 nm.

[0017] The material rheological property parameter, specifically a physical quantity describing the flow and deformation behavior of the nano-modified bio-based composite material in the processing process, includes shear viscosity, storage modulus, loss modulus and complex viscosity, which is obtained by measuring with a rheometer under different temperatures and shear rates.

[0018] The process complexity parameter, specifically a comprehensive index quantifying the difficulty level of the preparation process of the nano-modified bio-based composite material, is calculated by the weighted average of the number of processing steps, the number of equipment types, the control precision requirement of process parameters and the number of quality detection items, with a value range of 1 to 10.

[0019] The nano-particles are selected from one or more combinations of nano-particles, nano-particles, nano-particles, nano-particles, nano-particles, carbon nanotubes, graphene nanosheets or nano-clay. The nano-particles are selected from one or more combinations of nano-particles, nano-particles, nano-particles, nano-particles, nano-particles, carbon nanotubes, graphene nanosheets or nano-clay. The nano-particles are selected from one or more combinations of nano-particles, nano-particles, nano-particles, nano-particles, nano-particles, carbon nanotubes, graphene nanosheets or nano-clay. The nano-particles are selected from one or more combinations of nano-particles, nano-particles, nano-particles, nano-particles, nano-particles, carbon nanotubes, graphene nanosheets or nano-clay. The nano-particles are selected from one or more combinations of nano-particles, nano-particles, nano-particles, nano-particles, nano-particles, carbon nanotubes, graphene nanosheets or nano-clay. The nano-particles are selected from one or more combinations of nano-particles, nano-particles, nano-particles, nano-particles, nano-particles, carbon nanotubes, graphene nanosheets or nano-clay.

[0020] The training data set of the intelligent dispersion optimization model, specifically, the spatial distribution images of the nano-particles under different dispersion conditions are obtained by scanning electron microscopy and atomic force microscopy, the images are converted into three-dimensional coordinate data and the position information of each nano-particle is labeled, the clustering algorithm is used to identify the agglomeration area and calculate the aggregation degree, and the corresponding ultrasonic treatment power, dispersion time, titanate coupling agent concentration process parameters are recorded.

[0021] The training step of the intelligent dispersion optimization model, specifically, the input data is first standardized and preprocessed, and a graph structure representation is constructed, the Adam optimizer is used to set the learning rate to 0.001, the batch size is set to 32, the mean square error loss function is used to measure the difference between the predicted dispersion and the true dispersion, and the training process is divided into three stages of pre-training, fine tuning and verification testing.

[0022] The present application establishes a nanoparticle dispersion evaluation system based on a graph theory model, uses a minimum spanning tree algorithm to calculate the optimal dispersion path between nanoparticles, controls the dispersion degree to be above 0.85 and the aggregation degree to be below 0.15, and effectively solves the agglomeration problem of nanoparticles in the preparation stage of the dispersion liquid. The present application introduces an intelligent dispersion optimization model, performs real-time prediction and regulation on the spatial distribution of nanoparticles based on a graph convolution network architecture, realizes the aggregation and propagation of neighboring particle information through a message passing mechanism, dynamically adjusts the number of message passing steps according to the number of typical aggregation zones of nanoparticle clusters, makes the nanoparticles form a uniform three-dimensional network structure in the polyethylene matrix material, and avoids the local stress concentration phenomenon caused by uneven dispersion in the traditional process. Through a multi-objective optimization game model, the balance relationship between the tensile strength, bending strength, impact strength, cost control and processing performance is comprehensively considered, the Pareto optimal theory and genetic algorithm are used to solve the optimal parameter combination, the accurate control of the nanoparticle addition amount, mixing temperature, mixing time and other key process parameters is realized, and the technical problems that the nanoparticles are easy to agglomerate in the polyethylene matrix material, which causes uneven dispersion in the composite material matrix, forms local stress concentration points and affects the overall mechanical properties are solved. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 A flowchart of the method of the present application.

[0024] Figure 2 A structural diagram of the intelligent dispersion optimization model involved in the present application. DETAILED DESCRIPTION

[0025] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0026] As shown in Figure 1 , it is a flowchart of a mechanical property optimization method of a bio-based composite material based on nano modification provided by the present application, and the method comprises the following steps:

[0027] S01, the wood powder is pretreated, the particle size of the wood powder is controlled in the range of 40 to 80 meshes, and the wood powder is dried at a temperature of 80 to 100 DEG C for 4 to 6 hours, so that the moisture content of the wood powder is reduced to less than 2% by mass fraction, and then the surface of the wood powder is modified by using a silane coupling agent to obtain modified wood powder;

[0028] S02, a nanoparticle dispersion liquid is prepared, the nanoparticles are dispersed in ethanol solution at a mass fraction of 0.5% to 3%, ultrasonic treatment power is 300 to 500W, and continuous dispersion is performed for 30 to 60 minutes, and a titanate coupling agent is added for surface functionalization treatment to obtain a functionalized nanoparticle dispersion liquid;

[0029] S03, evaluating the uniformity of the nanoparticles in the functionalized nanoparticle dispersion by dispersion degree detection, determining the particle size distribution by dynamic light scattering technology, establishing a graph theory model of the nanoparticle dispersion network, regarding each nanoparticle as a node in the graph and the interaction force between the nanoparticles as the edge weight, calculating the optimal dispersion path by the minimum spanning tree algorithm, requiring the dispersion degree value to be greater than 0.85 and the aggregation degree value to be less than 0.15, to ensure that the nanoparticles achieve a good dispersion state;

[0030] S04, mixing the modified wood powder and the polyethylene matrix material at a mass ratio of 30:70 to 50:50, melt blending in a twin-screw extruder at a temperature of 160 to 180℃, and controlling the screw rotation speed at 50 to 80 revolutions per minute to obtain a wood powder-polyethylene mixture;

[0031] S05, synchronously adding the functionalized nanoparticle dispersion during the melt blending process, using in-situ polymerization technology to form a three-dimensional network structure of the nanoparticles in the polyethylene matrix material, evaluating the network structure integrity by the grid performance index, requiring the grid performance index to be greater than 0.75, and simultaneously using the intelligent dispersion optimization model to real-time regulate the spatial distribution of the nanoparticles to obtain a nano-modified bio-based composite material;

[0032] S06, testing the mechanical properties of the prepared nano-modified bio-based composite material, determining the tensile strength, bending strength and impact strength, calculating the performance gain index to evaluate the nano-modification effect, establishing a multi-objective optimization game model to optimize the process parameters, and obtaining the mechanical property test results;

[0033] S07, adjusting the nanoparticle addition amount, mixing temperature and mixing time parameters according to the calculation results of the multi-objective optimization game model, repeating steps S04 to S06 until the tensile strength, bending strength and impact strength all reach the preset target values, and completing the performance optimization of the composite material.

[0034] wherein the dispersion degree is a dimensionless parameter specifically representing the uniform distribution degree of the nanoparticles in the ethanol solution, the numerical value range is 0 to 1, and the closer the numerical value is to 1, the more uniform the dispersion, which is calculated by the reciprocal of the ratio of the standard deviation of the distance between the nanoparticles in a unit volume to the average distance.

[0035] wherein the aggregation degree is an evaluation index specifically reflecting the agglomeration degree of the nanoparticles, the numerical value range is 0 to 1, and the smaller the numerical value, the less the agglomeration phenomenon, which is determined by calculating the ratio of the number of agglomerates to the total number of particles.

[0036] The grid performance index is a comprehensive index for evaluating the integrity of the three-dimensional network structure of the nanoparticles in the polyethylene matrix material, is calculated by observing the network connectivity through a scanning electron microscope and combining with the conductivity test data, and has a value ranging from 0 to 1.

[0037] The performance gain index is an evaluation parameter for quantifying the improvement effect of nano-modification on the mechanical properties of the nano-modified bio-based composite material, is calculated by subtracting 1 from the ratio of the performance of the modified material to the performance of the original material, and is a positive value indicating performance improvement.

[0038] The multi-objective optimization game model is a mathematical optimization model that comprehensively considers tensile strength, bending strength, impact strength, cost control and processing performance, uses the Pareto optimal theory to establish a balance relationship between the objective functions, and solves the optimal parameter combination through a genetic algorithm. The mechanical property optimization function is used to maximize the comprehensive mechanical strength of the nano-modified bio-based composite material, and the input includes the tensile strength, the bending strength, the impact strength and the dispersity, and the output is a comprehensive score of mechanical properties. The cost control function is used to minimize the production and preparation cost, and the input includes the unit price of modified wood powder, the addition amount of nanoparticles, the processing technology energy consumption and the equipment depreciation fee, and the output is the total cost per unit product. The processing performance function is used to optimize the processability and forming stability of the nano-modified bio-based composite material, and the input includes the melt flow rate, the processing temperature window width, the shrinkage rate and the warping deformation degree, and the output is a processing difficulty coefficient. The mechanical property optimization function and the cost control function form a negative correlation coupling relationship through the addition amount of nanoparticles. The mechanical property optimization function and the processing performance function form a positive correlation coupling relationship through the material rheological characteristic parameter. The cost control function and the processing performance function form a negative correlation coupling relationship through the process complexity parameter. The constraint conditions include that the addition amount of nanoparticles needs to be controlled within the range of 0.5% to 3% by mass fraction, the mixing temperature needs to be maintained within the interval of 160 to 180°C, the material density needs to be kept within the range of 0.9 to 1.2 g / cm3, and the environmental impact assessment index needs to be lower than the preset threshold value of 0.3.

[0039] The intelligent dispersion optimization model is a nano-particle spatial distribution prediction and optimization system based on a graph convolution network architecture. The model includes an input layer for receiving nano-particle position coordinates and interaction force data, a graph convolution layer for extracting spatial relationship features between particles, a message passing layer for aggregating and propagating information of adjacent particles, and an output layer for generating an optimal dispersion control strategy. The number of message passing steps is dynamically adjusted according to the number of typical aggregation zones of nano-particle clusters. When the number of typical aggregation zones of nano-particle clusters increases, the number of message passing steps is increased to expand the information propagation range and improve the dispersion effect. The model uses residual connection and batch normalization techniques to improve training stability and convergence speed.​

[0040] The training data set of the intelligent dispersion optimization model is established by scanning electron microscopy and atomic force microscopy to obtain the spatial distribution images of nanoparticles under different dispersion conditions, converting the images into three-dimensional coordinate data and labeling the position information of each nanoparticle, using a clustering algorithm to identify the agglomeration area and calculate the aggregation degree, while recording the corresponding ultrasonic treatment power, dispersion time, titanate coupling agent concentration process parameters, establishing a data set containing 50000 groups of samples, of which 70% is used for training, 15% for verification, and 15% for testing, each group of samples contains particle coordinate matrix, adjacency graph, process parameter vector and target dispersion degree label, the data set covers different nanoparticle types, different polyethylene matrix materials and different dispersion process conditions to ensure the generalization ability of the model.

[0041] The training of the intelligent dispersion optimization model includes first standardizing the input data and constructing a graph structure representation, setting the learning rate to 0.001 using the Adam optimizer, setting the batch size to 32, using the mean square error loss function to measure the difference between the predicted dispersion and the true dispersion, and dividing the training process into three stages, the first stage is pre-trained for 100 rounds to establish the basic feature extraction ability, the second stage is fine-tuned for 200 rounds to optimize the message passing mechanism parameters, and the third stage is tested on the validation set for 50 rounds and the hyperparameters are adjusted, the early stopping strategy is used during training to prevent overfitting, the training is stopped when the validation set loss does not decrease for 10 consecutive rounds, the final model has a dispersion prediction accuracy of more than 92% on the test set, and the adaptive adjustment mechanism of the message passing step number can dynamically optimize according to the number of nanoparticle cluster typical aggregation areas.

[0042] The nanoparticle cluster typical aggregation area is a relatively stable spatial area formed by the aggregation of multiple nanoparticles in the polyethylene matrix material due to van der Waals force, electrostatic interaction and surface tension factors, each nanoparticle cluster typical aggregation area usually contains 5 to 50 nanoparticles, the area diameter ranges from 100 to 500 nm, the inter-particle distance in the aggregation area is less than 10 nm, and the inter-area distance is greater than 200 nm, the formation of the nanoparticle cluster typical aggregation area will significantly affect the mechanical and electrical properties of the nanomodified bio-based composite material, the number distribution rule of the nanoparticle cluster typical aggregation area in the material is determined by statistical analysis, the number of nanoparticle cluster typical aggregation areas is an important indicator for evaluating the dispersion quality of nanoparticles, and is also a key basis for setting the number of message passing steps in the intelligent dispersion optimization model.

[0043] The material rheological property parameter is specifically a physical quantity describing the flow and deformation behavior of the nano-modified bio-based composite material in the processing process, including shear viscosity, storage modulus, loss modulus, and complex viscosity, which is obtained by measuring the rheometer under different temperature and shear rate conditions.

[0044] The process complexity parameter is specifically a comprehensive index quantifying the difficulty level of the nano-modified bio-based composite material preparation process, which is calculated by the weighted average of the number of processing steps, the number of equipment types, the control accuracy requirement of process parameters, and the number of quality detection items, with a numerical range of 1 to 10, and the larger the value, the more complex the process.

[0045] The nanoparticles are preferably one or more combinations of nano particles, nano particles, nano particles, nano particles, nano particles, carbon nanotubes, graphene nanosheets, or nano-clay, wherein: the nano particles have an average particle size of 20-50 nm and a specific surface area of 200-400 , mainly used to improve the tensile strength and wear resistance of the material; the nano particles have an average particle size of 15-30 nm, have good photocatalytic performance and anti-UV aging effect; the nano particles have an average particle size of 30-80 nm, can significantly improve the hardness and thermal stability of the material; the nano particles have an average particle size of 20-40 nm, have antibacterial properties and UV shielding function; the nano particles have an average particle size of 40-100 nm, have low cost and can improve the processing performance of the material; the carbon nanotubes have a length of 1-10 μm and a diameter of 10-50 nm, mainly used to enhance the electrical conductivity and mechanical strength of the material; the graphene nanosheet has a thickness of 1-5 layers and a lateral size of 0.5-5 μm, has excellent thermal and electrical conductivity; the nano-clay has a sheet thickness of 1 nm and an aspect ratio greater than 100, can form an effective barrier structure.

[0046] The selection of the nanoparticles should be optimized according to the application requirements and performance requirements of the target composite material, and a single nanoparticle or a plurality of nanoparticles can be used to achieve a synergistic enhancement effect, wherein the mass ratio of each component nanoparticle in the multi-component complex system should be controlled within the range of 1:1 to 5:1, and the total addition amount still needs to meet the process requirement of 0.5% to 3% by mass fraction.

[0047] The specific embodiments of the above steps are described in detail below.

[0048] The specific implementation of step S01 is to first screen the particle size of the raw wood powder, and control the particle size of the wood powder in the range of 40 to 80 meshes by using a standard screen, which can ensure that the wood powder has a suitable specific surface area and packing density. Too small particle size will cause serious agglomeration, and too large particle size will affect the uniformity of dispersion. Then the screened wood powder is placed in a constant temperature drying oven and dried at a temperature of 80 to 100°C for 4 to 6 hours. The drying process adopts a gradient heating method, first preheating at 60°C for 1 hour, and then heating to the target temperature. By controlling the drying time and temperature, the moisture content of the wood powder is reduced to less than 2% by mass fraction. Too high moisture content will affect the treatment effect of the subsequent silane coupling agent. Next, the surface of the wood powder is modified by using a silane coupling agent. The silane coupling agent is prepared into a modification solution by mixing with anhydrous ethanol at a mass fraction of 1% to 3%. The dried wood powder is immersed in the modification solution at room temperature, and a mechanical stirrer is used for continuous stirring at a speed of 200 to 300 revolutions per minute for 2 to 4 hours. The silane coupling agent forms a covalent bond with the hydroxyl groups on the surface of the wood powder through chemical bonding, which significantly improves the interfacial compatibility of the wood powder and the polymer matrix. Finally, the modified wood powder is dried at 60°C for 2 hours to remove the residual solvent, and the surface-modified wood powder is obtained.

[0049] The specific implementation of step S02 is to accurately weigh the selected nanoparticles at a mass fraction of 0.5% to 3%, disperse them in anhydrous ethanol solution to prepare an initial dispersion liquid, and calculate the volume of the ethanol solution according to the mass and target concentration of the nanoparticles. An ultrasonic dispersion device is used to treat the initial dispersion liquid, with an ultrasonic power of 300 to 500W and a treatment time of 30 to 60 minutes. Ultrasonic dispersion uses micro-jets and shock waves generated by cavitation effect to destroy the agglomeration structure between nanoparticles. The temperature of the dispersion liquid should be controlled not to exceed 40°C during ultrasonic treatment to prevent ethanol evaporation. At the same time, titanate coupling agent is added at a mass fraction of 0.1% to 0.5% during ultrasonic dispersion. The titanate coupling agent can form a coordination bond with the hydroxyl or oxygen atoms on the surface of the nanoparticles through the titanium atoms in its molecular structure, and its organic segment has good compatibility with the polymer matrix, achieving functional modification of the nanoparticle surface. An intermittent ultrasonic method is used during dispersion, i.e. stopping for 5 minutes after 10 minutes of ultrasonic treatment, to avoid local overheating and agglomeration. Finally, functionalized nanoparticle dispersion liquid is obtained.

[0050] The specific implementation of step S03 is to measure the particle size distribution in the functionalized nanoparticle dispersion liquid by dynamic light scattering technology. Dynamic light scattering is based on the principle of Brownian motion, and the particle size distribution is calculated by detecting the fluctuation of scattered light intensity caused by Brownian motion of particles. Before testing, the dispersion liquid needs to be diluted to an appropriate concentration to avoid the interference of multiple scattering effects. A graph theory model of nanoparticle dispersion network is established, each nanoparticle in the detection area is regarded as a node in the graph, and the interaction force between particles is represented as the weight of the edge. The interaction force includes van der Waals force, electrostatic interaction and steric hindrance force, and the weight value is calculated by the distance between particles and surface characteristics. The minimum spanning tree algorithm is used to calculate the optimal dispersion path, which can find a tree structure connecting all nodes with the minimum total weight, thereby determining the optimal spatial distribution state of the nanoparticles. The dispersion degree is calculated by the reciprocal of the ratio of the standard deviation of the distance between nanoparticles in a unit volume to the average distance. The dispersion degree value is required to be greater than 0.85, indicating that the particle distribution is relatively uniform. At the same time, the aggregation degree is calculated by the ratio of the number of aggregates to the total number of particles. The aggregation degree value is required to be less than 0.15, indicating that the aggregation phenomenon is slight. Only when both the dispersion degree and the aggregation degree meet the requirements can the nanoparticles achieve good dispersion state.

[0051] The specific implementation of step S04 is to accurately proportion the modified wood powder and the polyethylene matrix material at a mass ratio of 30:70 to 50:50. The selection of the proportioning range is based on the filler critical volume fraction theory. Too low wood powder content cannot fully play the reinforcing effect, and too high wood powder content will lead to processing difficulty and mechanical property decline. The proportioned material is put into a twin-screw extruder for melt blending. The twin-screw extruder adopts a co-rotating design, with a screw diameter of 25 mm and a length-diameter ratio of 40:1. The screw configuration includes a conveying section, a compression section, a shearing section and a homogenizing section. The melt blending temperature is controlled at 160 to 180°C, which can ensure that the polyethylene is fully melted and the wood powder is not thermally degraded. The screw speed is controlled at 50 to 80 revolutions per minute. Too low speed will lead to insufficient mixing, and too high speed will cause excessive shearing and lead to wood powder fracture. During the extrusion process, the temperature gradient is controlled by adjusting the temperature of each heating zone. The feeding section temperature is set to 140°C, the melting section temperature is set to 170°C, the homogenizing section temperature is set to 165°C, and the die temperature is set to 160°C. The temperature gradient design is beneficial to the gradual plasticization and uniform mixing of the material, and finally the wood powder polyethylene mixture is obtained.

[0052] The specific implementation of step S05 is to synchronously add functionalized nanoparticle dispersion liquid through a side feeding port during the melt blending process of the twin-screw extruder, and the addition position is selected at two-thirds of the screw length. At this position, the wood powder and polyethylene have been fully mixed, but the material is still in a high shear state, which is conducive to the further dispersion of the nanoparticles. The in-situ polymerization technology is used to form a three-dimensional network structure of the nanoparticles in the polyethylene matrix material. During the in-situ polymerization process, the ethanol solvent volatilizes at high temperature, and the nanoparticles form a continuous three-dimensional network through physical entanglement and chemical bonding between the surface functionalized groups and the polyethylene molecular chains. The network connectivity is observed by scanning electron microscopy, and the grid performance index is calculated based on the conductivity test data. The grid performance index is calculated based on the percolation theory, and the required value is greater than 0.75 to ensure the integrity of the network structure. At the same time, the intelligent dispersion optimization model is used to real-time control the spatial distribution of the nanoparticles. The model is based on the graph convolution network architecture, the input layer receives the nanoparticle position coordinates and interaction force data, the graph convolution layer extracts the spatial relationship features between the particles through convolution operation, the message passing layer realizes the aggregation and propagation of the information of the adjacent particles, the number of message passing steps is dynamically adjusted according to the number of typical aggregation areas of the nanoparticle groups detected, and when the number of aggregation areas increases, the number of message passing steps is increased to expand the information propagation range, and the output layer generates the optimal dispersion control strategy to guide the adjustment of the processing parameters, and finally the nano-modified bio-based composite material is obtained.

[0053] The specific implementation of step S06 is to test the standard mechanical properties of the prepared nano-modified bio-based composite material. Standard test samples are prepared according to relevant national standards, and tensile strength, bending strength and impact strength tests are performed. The tensile strength test uses a universal testing machine, the sample size is 150 mm in length, 25 mm in width and 4 mm in thickness, the tensile speed is set to 5 mm per minute, and the test temperature is room temperature 23°C. The bending strength test uses a three-point bending method, the sample size is 80 mm in length, 10 mm in width and 4 mm in thickness, the span is set to 64 mm, and the loading speed is 2 mm per minute. The impact strength test uses a pendulum impact testing machine, the sample is prepared as a standard notched sample, the impact energy is 1 joule, and the test result is expressed in terms of unit area energy absorption. The performance gain index is calculated by reducing the ratio of the performance of the modified material to the performance of the original material by 1, a positive value indicates performance improvement, and a negative value indicates performance decline. A multi-objective optimization game model is established to optimize the process parameters. The model uses the Pareto optimal theory to establish a balance relationship between the mechanical property optimization function, the cost control function and the processing performance function, and solves the optimal parameter combination by genetic algorithm. The genetic algorithm includes basic operations such as selection, crossover and mutation, the population size is set to 100, the evolution number is set to 500, the crossover probability is set to 0.8, and the mutation probability is set to 0.1. Finally, the mechanical property test results and optimized parameter suggestions are obtained.

[0054] The specific implementation of step S07 is to accurately adjust the key process parameters such as the amount of nanoparticles added, the mixing temperature and the mixing time according to the calculation results of the multi-objective optimization game model. The parameter adjustment range needs to be within the constraint condition. The amount of nanoparticles added is controlled within the range of 0.5% to 3% by mass fraction, the mixing temperature is maintained within the interval of 160 to 180°C, and the material density is maintained within the range of 0.9 to 1.2 The parameter adjustment process adopts an iterative optimization strategy. After each adjustment, steps S04 to S06 are repeated to determine the adjustment direction and adjustment amplitude of the next round by comparing the difference between the test results and the preset target value. The adjustment amplitude is determined according to the gradient information of the objective function. When the tensile strength, bending strength and impact strength reach the preset target value at the same time, the iteration process stops. The preset target value is determined according to the specific application requirements. Generally, the tensile strength is required to be improved by more than 20%, the bending strength is required to be improved by more than 15%, and the impact strength is required to be improved by more than 25%. The entire optimization process is automatically adjusted through the feedback control principle, a closed-loop control system is established to ensure the stability and consistency of the final product quality, the performance optimization of the composite material is completed, and the nano-modified bio-based composite material product that meets the application requirements is obtained.

[0055] The intelligent dispersion optimization model adopts a deep learning framework based on graph convolution network architecture and is designed for the prediction and optimization control of the spatial distribution of nanoparticles in the polymer matrix. The input layer of the model is responsible for receiving multi-dimensional nanoparticle state information, including the three-dimensional spatial coordinates of each nanoparticle, surface charge density, particle size, surface functional group type, and physical quantity data such as van der Waals force, electrostatic interaction force, and spatial steric hindrance force between adjacent particles. After standardization preprocessing, these input data form a node feature matrix and an adjacency relationship graph. The graph convolution layer is the core component of the model, which adopts a multi-layer stacked graph convolution neural network structure. Each layer can extract the spatial relationship features between nanoparticles from the local neighborhood through convolution operation and map these features to a high-dimensional feature space, enabling the model to capture complex inter-particle interaction patterns and dispersion behavior rules. The message passing layer implements a dynamic aggregation and propagation mechanism for information between nanoparticles. Through the designed message passing function, the state information of each particle is propagated to its adjacent particles, while receiving state update information from adjacent particles. This bidirectional information exchange mechanism enables the model to globally understand the dynamic evolution process of the entire nanoparticle network. The adaptive adjustment mechanism of the number of message passing steps is an important innovation of the model. The system dynamically adjusts the number of message passing iterations by real-time detection of the number of typical aggregation zones of nanoparticle clusters. When the number of aggregation zones increases, the model increases the number of message passing steps to expand the information propagation range, thereby more effectively identifying and eliminating agglomeration phenomena. This adaptive mechanism ensures the dynamic adjustment capability of dispersion optimization effect. The output layer adopts a multi-head attention mechanism to generate the optimal dispersion control strategy, including ultrasonic power adjustment suggestions, titanium acid ester coupling agent concentration optimization schemes, temperature control parameters, and mechanical stirring intensity adjustments, etc. Specific process parameter guidance. The model also integrates residual connection and batch normalization technology to improve the stability and convergence speed of the training process. Residual connection effectively alleviates the gradient vanishing problem in deep networks, while batch normalization ensures the stability of the input data distribution of each layer, significantly improving the training efficiency and generalization ability of the model.

[0056] The establishment process of the intelligent dispersion optimization model training data set adopts a comprehensive method of multi-modal data acquisition and labeling. First, high-resolution imaging of the spatial distribution of nanoparticles under different dispersion conditions is performed using a scanning electron microscope and an atomic force microscope. The accelerating voltage of the scanning electron microscope is set to 15-20 kV, and the magnification is controlled at 50,000-200,000 times to obtain detailed topography and distribution information of the nanoparticles. The atomic force microscope is scanned in tapping mode with a scanning range of 5 μm x 5 μm to 20 μm x 20 μm, which can provide three-dimensional surface topography and height information of the nanoparticles. The image acquisition process covers different nanoparticle types including nanosilica, nanotitanium dioxide, nanometer alumina, nanometer zinc oxide, nanometer calcium carbonate, carbon nanotubes, graphene nanosheets, and nanoclay, as well as different polyethylene matrix materials including high-density polyethylene, low-density polyethylene, and linear low-density polyethylene, ensuring the diversity and representativeness of the data set. The post-processing of image data includes noise filtering, contrast enhancement, edge detection, and particle recognition. Computer vision algorithms are used to convert two-dimensional image information into three-dimensional coordinate data. Each identified nanoparticle is assigned precise spatial coordinates, particle size information, and shape parameters. A clustering algorithm is used to analyze particle distribution, identify agglomeration regions, and calculate corresponding aggregation degree values. The simultaneous recording of process parameters is a key step in data set establishment, including different set values of ultrasonic treatment power from 300 W to 500 W, dispersion time from 30 min to 60 min, titanium ester coupling agent concentration from 0.1% to 0.5% (mass fraction), treatment temperature from room temperature to 40°C, and initial concentration of nanoparticles in ethanol solution from 0.5% to 3% (mass fraction). These process parameters and corresponding particle distribution images form a complete input-output data pair. The data set finally contains 50,000 high-quality samples, each containing detailed particle coordinate matrices, adjacency relationship graphs based on spatial proximity, complete process parameter vectors, and expert-labeled target dispersion degree labels. The data set is divided into 70% for model training, 15% for verification and parameter adjustment, and 15% for final testing. A strict data quality control system is established, including data consistency check, outlier detection, and label accuracy verification, to ensure the reliability and effectiveness of the training data, laying a solid data foundation for the successful training and application of the intelligent dispersion optimization model.

[0057] It should be noted that the nanoparticle dispersion network optimization technology based on the graph theory model is one of the innovations of the present application. The traditional nanoparticle dispersion method mainly relies on empirical process parameter adjustment, and lacks accurate characterization and prediction ability of the spatial distribution state of particles. The present application regards each nanoparticle as a node in the graph, and the interaction force between the particles as the edge weight, and calculates the optimal dispersion path through the minimum spanning tree algorithm, realizing the fundamental change from qualitative experience to quantitative science. This method can accurately quantify the dispersion and aggregation parameters, providing theoretical guidance for the optimization of the spatial distribution of nanoparticles, and significantly improving the controllability and reproducibility of the dispersion effect.

[0058] The dynamic regulation technology of the intelligent dispersion optimization model represents the deep integration of artificial intelligence and material preparation process. The dispersion state of nanoparticles in the traditional preparation process is difficult to monitor and adjust in real time, often leading to instability of the performance of the final product. The intelligent model based on the graph convolution network architecture constructed in the present application realizes the aggregation and propagation of the information of adjacent particles through the message passing mechanism, and dynamically adjusts the number of message passing steps according to the number of typical aggregation zones of the nanoparticle clusters, realizing the adaptive optimization of the dispersion process. This intelligent regulation strategy can adjust the process parameters in real time according to the actual dispersion state, ensuring that the nanoparticles form an ideal three-dimensional network structure in the polymer matrix.

[0059] The parameter coordination technology of the multi-objective optimization game model solves the problem of mutual restriction of multiple performance indicators in the preparation process of composite materials. The traditional optimization method often only focuses on a single performance indicator, ignoring the complex coupling relationship between mechanical properties, cost control and processing performance. The present application uses the Pareto optimal theory to establish the balance relationship between multiple objective functions, and solves the optimal parameter combination through genetic algorithm, realizing the comprehensive optimization goal of meeting the performance requirements while considering economy and process feasibility.

[0060] The synergistic effect of these three key technology ideas produces significant technical advantages. The graph theory model provides an accurate mathematical description basis for intelligent optimization, the intelligent model provides real-time feedback information for multi-objective optimization, and the multi-objective optimization provides a global optimal solution for the entire preparation process. The three form a complete technical chain from micro-particle distribution to macro-performance regulation, realizing the intelligentization, precision and high efficiency of the preparation process of nano-modified bio-based composite materials, and providing important technical support for the industrial production of new materials.

[0061] Specifically, the principle of the present application is that the technical solution adopted by the present application can solve the problem of nanoparticle agglomeration. The fundamental principle of solving the problem of nanoparticle agglomeration is to establish a whole-process control system from dispersion liquid preparation to composite material forming. In the dispersion liquid preparation stage, the surface chemical properties of the nanoparticles are changed by surface functionalization treatment of silane coupling agents and titanate coupling agents, the van der Waals force and electrostatic interaction between the particles are reduced, and at the same time, external mechanical force is provided by ultrasonic treatment to destroy the agglomerates that have been formed. By establishing a graph theory model, the nanoparticles are regarded as nodes in the graph, the interaction force between the particles is regarded as the edge weight, and the optimal spatial distribution between the particles is calculated by using the minimum spanning tree algorithm, which provides a theoretical guidance for subsequent dispersion control.

[0062] In the composite material preparation stage, the intelligent dispersion optimization model based on the graph convolution network architecture can capture the complex spatial relationship between the nanoparticles, realize the propagation of local information to the global through the message passing layer, and when the number of typical agglomeration areas of the nanoparticles is detected to increase, the model automatically increases the number of message passing steps and expands the information propagation range, thereby guiding the parameter adjustment in the dispersion process. This adaptive mechanism ensures that the nanoparticles can form a uniform distribution network structure in three-dimensional space, avoiding local aggregation phenomenon.

[0063] The multi-objective optimization game model converts the complex process optimization problem into a solvable mathematical problem by mathematical modeling, uses the Pareto optimal theory to process multiple mutually restrictive objective functions, and finds the optimal solution set by using the global search ability of the genetic algorithm. The model considers the negative correlation coupling relationship between the mechanical property optimization function and the cost control function through the nanoparticle addition amount, and the positive correlation coupling relationship between the processing performance function and the material rheological characteristic parameter through the nanoparticle addition amount, ensuring the rationality and feasibility of the process parameters. Through repeated iteration optimization, the dispersion state of the nanoparticles in the polyethylene matrix gradually tends to be ideal, and finally the mechanical properties of the composite material are significantly improved. This systematic optimization method avoids the nanoparticle agglomeration phenomenon from the root cause, ensures the uniformity of the stress distribution in the composite material, and effectively improves the overall mechanical properties of the material.

[0064] A specific embodiment 1 of the present application is provided below, and the specific implementation manner of each step in the embodiment 1 is described in detail as follows.

[0065] The specific embodiment of step S01 is to first screen the particle size of the raw wood powder, control the particle size of the wood powder in the range of 40 to 80 meshes by using a standard screen, and then place the screened wood powder in a constant temperature drying oven for drying treatment at a temperature of 80 to 100°C for 4 to 6 hours, so that the moisture content of the wood powder is reduced to less than 2% by mass. Next, the surface of the wood powder is modified by using a silane coupling agent, the silane coupling agent is prepared into a modified solution in a proportion of 1% to 3% by mass with anhydrous ethanol, the dried wood powder is immersed in the modified solution at room temperature, a mechanical stirrer is used for continuous stirring at a speed of 200 to 300 revolutions per minute for 2 to 4 hours, and finally the modified wood powder is dried at 60°C for 2 hours to remove residual solvents, and the surface-modified wood powder is obtained.

[0066] The specific embodiment of step S02 is to accurately weigh the selected nanoparticles in a proportion of 0.5% to 3% by mass, and disperse them in anhydrous ethanol solution to prepare an initial dispersion liquid. An ultrasonic dispersion device is used to treat the initial dispersion liquid, the ultrasonic power is set to 300 to 500W, and the treatment time is 30 to 60 minutes. While ultrasonic dispersion, a titanate coupling agent is added in a proportion of 0.1% to 0.5% by mass. An intermittent ultrasonic method is used during dispersion, i.e. ultrasonic treatment for 10 minutes followed by 5 minutes of stopping. Finally, a functionalized nanoparticle dispersion liquid is obtained.

[0067] The specific embodiment of step S03 is to use dynamic light scattering technology to measure the particle size distribution of the functionalized nanoparticle dispersion liquid, establish a graph theory model of the nanoparticle dispersion network, regard each nanoparticle in the detection area as a node in the graph, represent the interaction force between the particles as the weight of the edge, and calculate the optimal dispersion path by using the minimum spanning tree algorithm. The calculation expression of the dispersion degree is:

[0068] ;

[0069] In the formula, is the dispersion degree; is the standard deviation of the distance between nanoparticles in unit volume, in nm, which is obtained by statistical analysis of the standard deviation of the distance between all nanoparticles in unit volume; is the average distance between nanoparticles in unit volume, in nm, which is obtained by calculating the arithmetic mean of the distance between all nanoparticles in unit volume. The calculation expression of the aggregation degree is:

[0070] ;

[0071] In the formula, is the aggregation degree; is the number of aggregates, which is obtained by identifying the aggregation area by clustering algorithm and statistical analysis; The total particle number is obtained by image recognition and particle counting. The dispersion value is required to be greater than 0.85, and the aggregation value is required to be less than 0.15, to ensure that the nanoparticles are in a good dispersion state.

[0072] The specific implementation of step S04 is to accurately proportion the modified wood powder and the polyethylene matrix material at a mass ratio of 30:70 to 50:50, and then put the proportioned material into a twin-screw extruder for melt blending. The melt blending temperature is controlled at 160 to 180°C, and the screw rotation speed is controlled at 50 to 80 revolutions per minute. The temperature gradient is controlled by adjusting the temperature of each heating zone. The feeding section temperature is set to 140°C, the melting section temperature is set to 170°C, the homogenization section temperature is set to 165°C, and the die temperature is set to 160°C. Finally, a wood powder polyethylene mixture is obtained.

[0073] The specific implementation of step S05 is to synchronously add the functionalized nanoparticle dispersion liquid through the side feeding port during the melt blending process in the twin-screw extruder, and to use in-situ polymerization technology to form a three-dimensional network structure of the nanoparticles in the polyethylene matrix material. The calculation expression of the grid performance index is:

[0074] ;

[0075] In the formula, is the grid performance index; is the network connectivity coefficient, which is obtained by observation with a scanning electron microscope, and the value range is 0 to 1; is the conductivity parameter, and the unit is ; is the network density parameter, and the unit is , which is obtained by counting the number of network nodes in a unit volume; is the weight coefficient, which satisfies The default values are 0.4, 0.3, and 0.3, respectively. At the same time, an intelligent dispersion optimization model is used to real-time regulate and control the spatial distribution of the nanoparticles. The message passing step number is dynamically adjusted according to the number of typical aggregation zones of the nanoparticle clusters, and the adjustment formula is:

[0076] ;

[0077] In the formula, is the message passing step number; is the basic message passing step number, and the default value is 3; is the adjustment coefficient, and the default value is 0.5; is the number of typical aggregation zones of the nanoparticle clusters, which is obtained by image analysis and clustering algorithm recognition.

[0078] The specific implementation of step S06 is to perform standardized mechanical property tests on the prepared nano-modified bio-based composite material, including tensile strength, bending strength, and impact strength tests. The performance gain index is calculated as follows:

[0079]

[0080] wherein, is the performance gain index; is the performance parameter of the modified material, including tensile strength, bending strength, or impact strength, with units of MPa, MPa, obtained through standardized mechanical property tests; is the corresponding performance parameter of the original material, with the same units as obtained through the same test method as a comparison benchmark. A multi-objective optimization game model is established to optimize the process parameters, and the mechanical property optimization function expression is as follows:

[0081]

[0082] wherein, is the comprehensive score of mechanical properties; is the tensile strength, with units of MPa; is the bending strength, with units of MPa; is the impact strength, with units of ; is the dispersity; is the weight coefficient, with default values of 0.3, 0.25, 0.25, and 0.2, respectively, determined through the analytic hierarchy process or expert scoring method. The cost control function expression is as follows:

[0083]

[0084] wherein, is the total cost per unit product, with units of yuan; is the unit price of modified wood powder, with units of yuan / kg, obtained through market research; is the mass of modified wood powder, with units of kg, calculated according to the formula; is the unit price of nanoparticles, with units of yuan / kg, obtained through supplier quotes; is the addition amount of nanoparticles, with units of kg, calculated according to the addition ratio; is the processing process energy consumption, with units of kW, obtained through equipment power calculation; is the processing time, with units of h, determined according to the production process; is the equipment depreciation cost, with units of yuan, calculated by apportioning according to the service life of the equipment. The processing performance function expression is as follows:

[0085] ​​​ ;

[0086] wherein, is the processing difficulty coefficient; is the melt flow rate, unit is g / 10min; ; is the processing temperature window width, unit is ℃; is the shrinkage, dimensionless; is the warpage deformation degree, unit is mm.

[0087] The specific implementation of step S07 is to accurately adjust the key process parameters such as the amount of nanoparticle addition, mixing temperature and mixing time according to the calculation results of the multi-objective optimization game model. The parameter adjustment adopts an iterative optimization strategy, and the gradient calculation expression of the objective function is:

[0088] ;

[0089] wherein, is the gradient vector of the objective function; is the objective function of the multi-objective optimization game model, which includes the comprehensive evaluation of mechanical properties, cost control and processing performance; are the amount of nanoparticle addition, mixing temperature and mixing time, respectively, with units of mass fraction %, ℃ and min; is the corresponding unit vector; is the partial derivative of the objective function to each parameter, which is obtained by numerical differentiation method. When the tensile strength, bending strength and impact strength reach the preset target value at the same time, the iteration process is stopped. The constraint conditions include that the amount of nanoparticle addition is controlled within the range of 0.5% to 3% of mass fraction, the mixing temperature is maintained within the interval of 160 to 180 ℃, the material density is kept within the range of 0.9 to 1.2 , and the environmental impact assessment index needs to be lower than the preset threshold value 0.3, so as to complete the performance optimization of the composite material.

[0090] wherein, is measured by a universal testing machine, the sample size is 150 mm in length, 25 mm in width and 4 mm in thickness, the tensile speed is set to 5 mm / min, and the test temperature is room temperature 23 ℃. is obtained by three-point bending test, the sample size is 80 mm in length, 10 mm in width and 4 mm in thickness, the span is set to 64 mm, and the loading speed is 2 mm / min. is measured by a pendulum impact testing machine, the sample is prepared as a standard notched sample, and the impact energy is 1 J. is obtained by observing the network connectivity by a scanning electron microscope, the observation magnification is 50,000 to 200,000 times. The conductivity of the material is measured by a four-probe method. The melt flow rate is determined at 190°C under a load of 2.16 kg. The material processing temperature window is determined by differential scanning calorimetry. The dimensional change before and after molding of the product is calculated by measurement. The degree of surface warping deformation of the product is determined by a three-coordinate measuring instrument. The environmental impact assessment index is calculated by a life cycle assessment method, including environmental impact assessment of each link of raw material mining, processing and manufacturing, use stage and waste disposal.

[0091] In order to better understand and implement the present application, the following provides an embodiment 2 of a specific application scenario of the present application: a research team undertook a new type of bio-based composite material development task, which needs to prepare a wood powder polyethylene composite material with high strength and good processing performance, which is used to replace traditional pure polyethylene material to manufacture automobile interior parts. The traditional wood powder polyethylene composite material has technical problems such as poor interfacial compatibility, unstable mechanical properties, and uneven dispersion of nanoparticles, which leads to large fluctuations in product quality and cannot meet the strict requirements of the automobile industry. The research team decided to use the mechanical property optimization method of the nano-modified bio-based composite material of the present application to solve these technical problems.

[0092] The research team first pretreated the raw material wood powder, selected poplar powder as the filler, and controlled the particle size of the wood powder to 60 mesh through a standard screen to ensure uniform particle size distribution. The screened wood powder was placed in a constant temperature drying oven and dried at a temperature of 90°C for 5 hours to reduce the moisture content of the wood powder to 1.8%. Then, the surface of the wood powder was modified by using silane coupling agent KH550, and the silane coupling agent was prepared into a modification solution with anhydrous ethanol at a mass fraction of 2%. The dried wood powder was immersed in the modification solution at room temperature, and a mechanical stirrer was used to continuously stir at a speed of 250 revolutions per minute for 3 hours. After the treatment, the modified wood powder was dried at 60°C for 2 hours to remove the residual solvent, and the wood powder with good surface modification was obtained.

[0093] During the preparation of the nanoparticle dispersion liquid, the research team selected nanoparticles as the reinforcing phase, with an average particle size of 30 nm and a specific surface area of 300 / g. The nanoparticles were dispersed in the polyethylene matrix by using a high-speed homogenizer, and the dispersion liquid was prepared. ​The particles were dispersed at a mass fraction of 2% in anhydrous ethanol solution to prepare an initial dispersion, with the ethanol solution volume being 500 ml. The initial dispersion was treated using an ultrasonic dispersion device with an ultrasonic power set to 400 W for 45 minutes. Simultaneously with ultrasonic dispersion, a titanate coupling agent NDZ201 was added at a mass fraction of 0.3%. The dispersion process employed intermittent ultrasonication, i.e., ultrasonication for 10 minutes followed by a 5-minute pause, controlling the dispersion temperature to not exceed 35°C, ultimately yielding a functionalized nanoparticle dispersion.

[0094] The research team employed dynamic light scattering technology to evaluate the dispersion of functionalized nanoparticle dispersions. They established a graph theory model of the nanoparticle dispersion network, treating each nanoparticle within the detection area as a node in the graph, with the interaction forces between particles represented by edge weights. By statistically analyzing the distance data between all nanoparticles per unit volume, they calculated the standard deviation of the distances. The average spacing is 8.5nm. The value is 125 nm, according to the dispersion calculation formula. The dispersion was 14.7, which translates to a standardized value of 0.88, meeting the requirement of being greater than 0.85. Simultaneously, clustering algorithms were used to identify clustered regions, and the number of clusters was statistically determined. The total number of particles is 127. There are 1856 clusters, according to the clustering degree calculation formula. The aggregation degree was 0.068, which meets the requirement of being less than 0.15.

[0095] During the melt blending stage, the research team precisely proportioned modified wood flour and high-density polyethylene at a mass ratio of 40:60. The polyethylene melt flow rate was 0.3 g / 10 min, and the density was 0.96 g / min. The formulated materials were fed into a twin-screw extruder for melt blending. The extruder screw diameter was 25 mm, and the length-to-diameter ratio was 40:1. The melt blending temperature was controlled at 170℃, and the screw speed was controlled at 65 rpm. The temperatures of each heating zone were set as follows: feed section 140℃, melting section 170℃, homogenization section 165℃, and die section 160℃. Functionalized nanoparticle dispersion was simultaneously added through a side feed inlet at two-thirds of the screw length, at a rate of 50 ml / min.

[0096] The three-dimensional network structure formed by nanoparticles in a polyethylene matrix was observed using scanning electron microscopy. Network connectivity was observed at 50,000x magnification, and the network connectivity coefficient was measured. The value was 0.82. The conductivity of the material was determined using the four-probe method, and the conductivity parameters were obtained. for S / m. Network density parameters were measured through image analysis and statistical analysis of network density. For 156 The intelligent dispersion optimization model detects the number of typical aggregation areas of nanoparticle clusters For 18, according to the adjustment formula The number of message passing steps is calculated to be 12 steps, and the optimization model realizes further optimization of the spatial distribution of nanoparticles after running. As shown in Table 1, the calculation parameters and results of the grid performance index detail the numerical values of each key indicator.

[0097] Table 1 Calculation parameters of grid performance index

[0098]

[0099] According to the grid performance index calculation formula , the normalized weights of each parameter are 0.4, 0.3, and 0.3, respectively, and the calculated grid performance index is 0.78, which meets the requirement of greater than 0.75, ensuring the integrity of the three-dimensional network structure of the nanoparticles.

[0100] The research team conducted standardized mechanical property tests on the prepared nano-modified bio-based composite materials, and prepared standard samples according to national standards. The tensile strength test used a universal testing machine, with a sample size of length 150 mm, width 25 mm, and thickness 4 mm, a tensile speed of 5 mm / min, and a test temperature of 23°C. The measured tensile strength was 32.5 MPa. The bending strength test used a three-point bending method, with a sample size of length 80 mm, width 10 mm, and thickness 4 mm, a span of 64 mm, and a loading speed of 2 mm / min. The measured bending strength was 48.7 MPa. The impact strength test used a pendulum impact testing machine, with a standard notched sample, an impact energy of 1 J, and a measured impact strength of 25.3 Compared with the mechanical property data of the original wood flour polyethylene composite material, the tensile strength was 28.2 MPa, the bending strength was 42.1 MPa, and the impact strength was 21.8 As shown in Table 2, the comparison data of the mechanical properties of the material before and after modification clearly shows the effect of nano-modification.

[0101] Table 2 Comparison table of mechanical properties of material before and after modification

[0102]

[0103] According to the performance gain index calculation formula , the tensile strength performance gain index is 0.152, the bending strength performance gain index is 0.157, and the impact strength performance gain index is 0.161, all achieving significant performance improvement.

[0104] The research team established a multi-objective optimization game model to further optimize the process parameters, with the weighting coefficients in the mechanical performance optimization function set as follows: =0.3, =0.25, =0.25, =0.2, the comprehensive mechanical performance score is calculated. The unit price of modified wood flour in the cost control function is 35.8. The price is 2.8 yuan / kg, the unit price of nanoparticles. The cost is 85 yuan / kg. The total cost per unit product is calculated based on the formula. It is 12.6 yuan. The melt flow rate in the processing performance function. The processing temperature window width is 0.3g / 10min. At 15℃, the shrinkage rate The value is 0.008, indicating the degree of warping deformation. The machining difficulty coefficient is calculated to be 0.15mm. The value is 3.5. As shown in Table 3, the calculation parameters and results of the multi-objective optimization game model provide a quantitative basis for process optimization.

[0105] Table 3 Calculation Results of Multi-Objective Optimization Game Model

[0106]

[0107] Based on the calculation results of the multi-objective optimization game model, the research team fine-tuned and optimized the key process parameters, adjusting the nanoparticle addition from 2% to 2.2%, the mixing temperature from 170℃ to 168℃, and extending the mixing time to 50 minutes. After the second round of preparation with optimized parameters, the tensile strength of the material increased to 33.1 MPa, the flexural strength increased to 49.8 MPa, and the impact strength increased to 26.1 MPa. This further validated the effectiveness of the optimization model.

[0108] Traditional wood flour-polyethylene composite material preparation processes primarily employ simple mechanical mixing methods. The addition of nanoparticles typically relies on experience to determine the dosage and dispersion process, lacking systematic methods for evaluating and optimizing dispersion. This results in uneven nanoparticle dispersion and significant fluctuations in material properties. Traditional methods mainly rely on mechanical stirring and simple ultrasonic treatment for nanoparticle dispersion, which cannot achieve precise dispersion control and leads to severe agglomeration. Process parameter optimization is primarily conducted through single-factor experiments or orthogonal experiments, resulting in low optimization efficiency and difficulty in achieving multi-objective synergistic optimization.

[0109] The present application brings significant technical progress compared with traditional methods, mainly reflected in the establishment of quantitative evaluation system of dispersity and aggregation, the dispersion state of nanoparticles is accurately controlled through mathematical model, the dispersity is improved from 0.73 of traditional method to 0.88, and the aggregation is reduced from 0.25 to 0.068. The introduction of intelligent dispersion optimization model realizes the adaptive regulation of the spatial distribution of nanoparticles, compared with the traditional fixed parameter dispersion process, the dispersion effect is improved by 18%. The establishment of multi-objective optimization game model makes the mechanical properties, cost control and processing performance realize the synergistic optimization, avoids the performance imbalance problem caused by single target optimization in traditional method. The comprehensive evaluation method of grid performance index provides a quantitative standard for the integrity of nano network structure, compared with the traditional qualitative observation method, the evaluation accuracy is improved by 15%. The standardized calculation of performance gain index eliminates the influence of different base materials on the evaluation of modification effect, so that the modification effect has comparability and reproducibility. Through the optimization of process parameters, the tensile strength, bending strength and impact strength of the finally prepared nano modified bio-based composite material are improved by 15.2%, 15.7% and 16.1% respectively compared with the traditional method, which significantly improves the comprehensive mechanical properties of the material.

[0110] It should be noted that the variables involved in the present application are explained in detail as shown in Table 4.

[0111] Table 4 Variable explanation table

[0112]

[0113] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for mechanical property optimization of nanomodified bio-based composites, characterized by, The application relates to a method for preparing a nano-modified bio-based composite material. The wood powder is pretreated, the surface of the wood powder is modified by using a silane coupling agent, and modified wood powder is obtained; A nanoparticle dispersion liquid is prepared, ultrasonic treatment is adopted, a titanate coupling agent is added for surface functionalization treatment, and a functionalized nanoparticle dispersion liquid is obtained; the uniformity of the nanoparticles in the functionalized nanoparticle dispersion liquid is evaluated through dispersion degree detection, a graph theory model of a nanoparticle dispersion network is established, each nanoparticle is taken as a node in the graph, the interaction force between the nanoparticles is taken as an edge weight, a minimum spanning tree algorithm is adopted to calculate a tree structure connecting all the nodes and having the minimum total weight, so that the optimal spatial distribution state of the nanoparticles is determined, the dispersion degree is calculated by counting the reciprocal of the ratio of the standard deviation of the distance between the nanoparticles in a unit volume to the average distance, and the dispersion degree value is required to be greater than 0.85, meanwhile, the aggregation degree is calculated by counting the ratio of the number of aggregates to the total number of particles, and the aggregation degree value is required to be less than 0.15, only when the dispersion degree and the aggregation degree meet the requirements at the same time, can it be ensured that the nanoparticles reach a good dispersion state; the modified wood powder is mixed with a polyethylene matrix material, melt blending is carried out in a double-screw extruder, and a wood powder polyethylene mixture is obtained; the functionalized nanoparticle dispersion liquid is synchronously added in the melt blending process, in-situ polymerization technology is adopted to make the nanoparticles form a three-dimensional network structure in the polyethylene matrix material, meanwhile, the spatial distribution of the nanoparticles is real-timely regulated by adopting an intelligent dispersion optimization model, and a nano-modified bio-based composite material is obtained. The mechanical property of the prepared nano-modified bio-based composite material is tested, a multi-objective optimization game model is established to optimize the process parameters, the nanoparticle addition amount, the mixing temperature and the mixing time parameters are adjusted according to the calculation results of the multi-objective optimization game model, and the foregoing steps are repeated until the tensile strength, the bending strength and the impact strength all reach preset target values.

2. The method for mechanical properties optimization of nanomodified bio-based composites according to claim 1, characterized in that, The pretreatment of the wood powder is specifically that the particle size of the wood powder is controlled in the range of 40 to 80 meshes, the wood powder is dried at a temperature of 80 to 100 DEG C for 4 to 6 hours, so that the water content of the wood powder is reduced to less than 2% in mass fraction, and then the surface of the wood powder is modified by using a silane coupling agent.

3. The method for mechanical properties optimization of nanomodified bio-based composites according to claim 2, characterized in that, The step of obtaining the functionalized nanoparticle dispersion liquid is specifically that the nanoparticles are dispersed in an ethanol solution at a mass fraction of 0.5% to 3%, ultrasonic treatment is adopted, the power is 300 to 500 W, and the nanoparticles are continuously dispersed for 30 to 60 minutes, and a titanate coupling agent is added for surface functionalization treatment.

4. The method for mechanical properties optimization of nanomodified bio-based composites according to claim 3, characterized in that, The step of mixing the modified wood powder with the polyethylene matrix material is specifically that the modified wood powder and the polyethylene matrix material are mixed at a mass ratio of 30 to 50:70 to 50, melt blending is carried out in a double-screw extruder at a temperature of 160 to 180 DEG C, and the screw rotation speed is controlled at 50 to 80 revolutions per minute.

5. The method for mechanical properties optimization of nanomodified bio-based composites according to claim 4, characterized in that, The step of forming a three-dimensional network structure of nanoparticles in a polyethylene matrix material after the step of using in-situ polymerization technology comprises the step of evaluating the integrity of the network structure by a meshing performance index, and the meshing performance index is required to be greater than 0.75, wherein the meshing performance index is a comprehensive index for evaluating the integrity of the three-dimensional network structure of nanoparticles in the polyethylene matrix material, and is calculated by observing the network connectivity by a scanning electron microscope and combining the conductivity test data, and the numerical range is 0 to 1, and the meshing performance index is required to be greater than 0.

75.

6. The method for mechanical properties optimization of nanomodified bio-based composites according to claim 5, characterized in that, The intelligent dispersion optimization model is specifically a nanoparticle spatial distribution prediction and optimization system based on a graph convolution network architecture, and the model comprises an input layer, a graph convolution layer, a message passing layer and an output layer, wherein the number of message passing steps is dynamically adjusted according to the number of typical aggregation zones of nanoparticle groups, and the model uses residual connection and batch normalization technology to improve training stability and convergence speed.

7. The method for mechanical properties optimization of nanomodified bio-based composites according to claim 6, characterized in that, In the step of testing the mechanical properties of the prepared nanomodified bio-based composite material, the step of calculating a performance gain index to evaluate the nanomodification effect is further included, wherein the performance gain index is specifically an evaluation parameter for quantifying the improvement effect of nanomodification on the mechanical properties of the nanomodified bio-based composite material, and is calculated by subtracting 1 from the ratio of the performance of the modified material to the performance of the original material, and a positive value indicates performance improvement.

8. The method for mechanical properties optimization of nanomodified bio-based composites according to claim 7, characterized in that, The multi-objective optimization game model is specifically a mathematical optimization model that comprehensively considers tensile strength, bending strength, impact strength, cost control and processing performance, and uses the Pareto optimal theory to establish a balance relationship between the objective functions, and solves the optimal parameter combination by a genetic algorithm.

Images