Preparation method of construction waste-based recycled aggregate

Abstract

The invention provides a preparation method of a construction waste-based recycled aggregate, and belongs to the technical field of construction waste reutilization. The preparation method comprises the following steps: firstly, removing impurities through vibration screening and manual sorting; then carrying out double-shaft crushing and impact type crushing to form angular aggregate; controlling surface friction treatment by utilizing an aggregate surface feature intelligent identification model, and adding cement powder to improve the surface density at the same time; then, determining a sodium silicate concentration parameter by adopting a material ratio optimization neural network to carry out impregnation modification, and monitoring an impregnation effect by utilizing a pore filling uniformity evaluation model; carrying out surface coating treatment by adopting an organic silicon water repellent agent; finally, the aggregate performance is evaluated through a comprehensive performance balance point determination function, a parameter adjustment suggestion is generated, a closed-loop optimization iteration mechanism is formed, and the technical problems that the construction waste-based recycled aggregate is unstable in quality and high in water absorption rate, and the performance is difficult to accurately regulate and control are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of construction waste recycling, and in particular relates to a method for preparing construction waste-based recycled aggregate. Background Art

[0002] With the rapid development of urban renewal and infrastructure construction, the production of construction waste has increased dramatically, and its resource utilization has become a crucial component of the circular economy. Traditional technologies for preparing recycled aggregate from construction waste primarily rely on mechanical crushing and simple screening. Using single or multi-stage crushing equipment, construction waste is broken down into aggregate of a specific particle size for use in non-load-bearing concrete or road base materials. These technologies have been widely adopted in engineering projects both domestically and internationally, essentially achieving both construction waste reduction and resource utilization. However, traditional preparation processes have significant drawbacks. First, the surface of recycled aggregate often contains a large amount of cement mortar and fine particles, resulting in low apparent density, high water absorption, and insufficient strength. Second, the preparation process lacks precise control, leading to large fluctuations in aggregate properties and difficulty meeting the stability requirements of high-quality concrete. Third, existing surface treatment methods are mostly empirical, making it difficult to precisely adjust process parameters for the different sources and characteristics of construction waste, resulting in inconsistent treatment results. Currently, improving the quality of recycled aggregate relies primarily on manual judgment and trial-and-error adjustments, lacking intelligent parameter optimization mechanisms. This makes it difficult to address the technical issues of unstable recycled aggregate quality, high water absorption, and difficulty in precisely controlling properties. This seriously restricts the application and promotion of recycled aggregates in high-performance concrete, and there is an urgent need to develop intelligent and precise recycled aggregate quality improvement technology. Summary of the Invention

[0003] In view of this, the present invention provides a method for preparing construction waste-based recycled aggregate, which can solve the technical problems in the prior art such as unstable quality of recycled aggregate, high water absorption rate and difficulty in precise control of performance.

[0004] The present invention is achieved as follows: The present invention provides a preparation method of construction waste-based recycled aggregate, including construction waste sorting and pretreatment, primary crushing of construction waste, secondary crushing and molding of materials, intelligent surface friction treatment, intelligent sodium silicate impregnation modification, silicone hydrophobic coating treatment, and performance evaluation and parameter optimization iteration. A closed-loop parameter optimization iterative mechanism is formed through an intelligent recognition model of aggregate surface features, a material ratio optimization neural network, a pore filling uniformity evaluation model, and a comprehensive performance balance point determination function to achieve precise control of the water absorption rate, crushing value, and mud content of the recycled aggregate.

[0005] Among them, the steps of sorting and pre-treating construction waste are specifically to sort and pre-treat construction waste, using vibration screening equipment to divide construction waste into three particle sizes of more than 50 mm, 10 to 50 mm, and less than 10 mm, and remove wood, plastic, and metal impurities through manual sorting combined with magnetic separation equipment to ensure that the purity of the raw materials reaches more than 95%.

[0006] Among them, the step of primary crushing of the construction waste is specifically to transport the sorted 10 to 50 mm particle size construction waste to the double-shaft crusher for primary crushing, control the speed of the double-shaft crusher to 60 to 80 revolutions per minute, and control the output particle size to 10 to 25 mm. At the same time, a circulating screening system is set up to return the oversized materials to the double-shaft crusher for further crushing.

[0007] Among them, the step of secondary crushing and forming of the material is specifically to send the material after primary crushing into the impact crusher for secondary crushing, set the linear speed of the impact crusher rotor to 35 to 45 meters per second, and adjust the gap between the material and the material impact plate to 15 to 20 mm, so that the material particle size is reduced to 5 to 15 mm, forming angular recycled aggregate.

[0008] Among them, the step of intelligent surface friction treatment specifically adopts mechanical friction method to perform surface treatment on the recycled aggregate after secondary crushing, uses the intelligent recognition model of aggregate surface characteristics to perform real-time analysis on the surface condition of the recycled aggregate, controls the speed of the surface treatment machine to 120 to 150 revolutions per minute, controls the operation time of the surface treatment machine to 25 to 30 minutes, and adds cement powder with a mass ratio of 2% to 5% of the mass of the recycled aggregate.

[0009] Among them, the intelligent sodium silicate impregnation modification step specifically involves surface modification of the surface-treated recycled aggregate, using a material ratio optimization neural network to determine the optimal sodium silicate concentration parameters, immersing the recycled aggregate in a sodium silicate solution with a mass concentration of 5% to 8%, and impregnating it at 60 to 70°C for 2 to 3 hours, and monitoring the impregnation effect parameters of the impregnation treatment through a pore filling uniformity evaluation model.

[0010] Among them, the step of the organosilicon hydrophobic coating treatment is specifically to transfer the dried recycled aggregate to a rotary coater, spray an organosilicon hydrophobic agent with a mass concentration of 3% to 5%, control the speed of the rotary coater to 80 to 100 revolutions per minute, and control the coating time of the rotary coater to 10 to 15 minutes, so that the coverage rate of the organosilicon hydrophobic agent on the surface of the recycled aggregate reaches more than 95%.

[0011] Among them, the steps of performance evaluation and parameter optimization iteration are specifically to perform performance testing on the surface-modified recycled aggregate, obtain the apparent density parameters, water absorption parameters, crushing value parameters and mud content parameters of the recycled aggregate, and input the apparent density parameters, water absorption parameters, crushing value parameters, mud content parameters and impregnation effect parameters into the comprehensive performance balance point determination function.

[0012] Among them, in the performance evaluation and parameter optimization iteration steps, the comprehensive performance index parameters of the recycled aggregate and the recommended parameters for adjustment of the processing parameters are calculated through the comprehensive performance balance point determination function, and qualified recycled aggregates with a water absorption parameter lower than 3%, a crushing value parameter less than 18%, a mud content parameter less than 1% and comprehensive performance index parameters meeting the standard requirements are screened out.

[0013] Among them, the mechanical friction method specifically refers to the use of equipment to generate continuous friction between recycled aggregates and between recycled aggregates and the inner wall of the equipment, thereby removing loose particles on the surface of the recycled aggregates and smoothing sharp edges and corners. At the same time, the added cement powder is partially hydrated under the action of frictional heat and fills the micropores on the surface of the recycled aggregates, forming a denser and stronger surface structure.

[0014] The surface modification treatment mentioned above specifically refers to the process of changing the surface properties of recycled aggregate by physical or chemical methods. Sodium silicate solution is used to impregnate the surface of the recycled aggregate to form a sodium silicate gel network structure, fill the surface pores and enhance the interfacial bonding force, thereby reducing the water absorption and moisture migration capacity of the recycled aggregate.

[0015] The organosilicon hydrophobic agent mentioned herein specifically refers to methyltriethoxysilane or octyltriethoxysilane organosilicon compounds. The organosilicon hydrophobic agent can form a hydrophobic film on the surface of the recycled aggregate, making it difficult for water molecules to penetrate into the interior of the recycled aggregate, reducing the water absorption rate parameters of the recycled aggregate, and improving the stability of the recycled aggregate in concrete.

[0016] The above-mentioned intelligent recognition model of aggregate surface features specifically refers to an image recognition model based on a convolutional neural network structure. It uses a high-definition camera to collect images of the recycled aggregate surface in real time, analyzes the surface roughness parameters, pore distribution parameters and surface morphology characteristic parameters of the recycled aggregate, and outputs the optimal surface treatment machine speed parameters and the optimal surface treatment machine operation time parameters.

[0017] The material ratio optimization neural network mentioned above specifically refers to a recurrent neural network model used to predict the optimal sodium silicate concentration parameters and the optimal impregnation time parameters. It includes a hybrid recurrent neural network structure with one gated recurrent unit layer and two long-short-term memory network layers. The gated recurrent unit layer is responsible for processing the basic characteristic information of the recycled aggregate, and the long-short-term memory network layer is responsible for establishing parameter association relationships.

[0018] The pore filling uniformity evaluation model mentioned above specifically refers to an image segmentation neural network used to monitor the penetration of sodium silicate solution during the impregnation process of recycled aggregate in real time. The fully convolutional neural network based on the UNet architecture includes a four-layer downsampling convolution module and a corresponding four-layer upsampling convolution module, which is used to evaluate the impregnation effect.

[0019] The comprehensive performance balance point determination function is specifically a function used to calculate the comprehensive performance index parameters of the recycled aggregate and the recommended parameters for adjusting the processing parameters. The basic comprehensive index parameters are calculated by the weighted average method, and are corrected according to the impregnation effect parameters. The recommended parameters for adjusting the processing parameters are generated to form a closed-loop optimization iterative mechanism to ensure that the performance of the recycled aggregate meets the target requirements.

[0020] The present invention uses an intelligent recognition model of aggregate surface characteristics to analyze the surface condition of recycled aggregate in real time and automatically adjust the surface treatment parameters; a material ratio optimization neural network is used to determine the optimal sodium silicate concentration, and the pore filling uniformity evaluation model is combined to monitor the impregnation effect; finally, a comprehensive performance balance point determination function is formed to achieve adaptive optimization of parameters. This method effectively solves the key problems in the traditional recycled aggregate preparation process. Intelligent recognition technology replaces manual experience judgment and realizes the precise control of recycled aggregate surface treatment parameters; the mechanical friction method combined with cement powder addition significantly improves the aggregate surface quality; sodium silicate impregnation and silicone hydrophobic treatment form multiple protective barriers, which greatly reduces the water absorption rate of recycled aggregate; the closed-loop parameter optimization iterative mechanism ensures the stability and consistency of product quality. It solves the technical problems of unstable quality of recycled aggregate, high water absorption rate and difficult to accurately control performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0023] like Figure 1 FIG. 1 is a flow chart of a method for preparing construction waste-based recycled aggregate provided by the present invention, and the method comprises the following steps:

[0024] S01. Sorting and pre-treatment of construction waste: Construction waste is sorted and pre-treated by using vibration screening equipment to separate the construction waste into three particle sizes: over 50 mm, 10 to 50 mm, and under 10 mm. Wood, plastic, and metal impurities are removed through manual sorting combined with magnetic separation devices to ensure that the purity of the raw materials reaches more than 95%;

[0025] S02. Primary crushing of construction waste: The sorted 10 to 50 mm particle size construction waste is transported to a twin-shaft crusher for primary crushing. The speed of the twin-shaft crusher is controlled at 60 to 80 revolutions per minute, and the output particle size is controlled at 10 to 25 mm. At the same time, a circulating screening system is set up to return oversized materials to the twin-shaft crusher for further crushing.

[0026] S03, secondary crushing and shaping of materials: The materials after primary crushing are fed into an impact crusher for secondary crushing. The rotor linear speed of the impact crusher is set to 35 to 45 meters per second, and the gap between the materials and the material impact plate is adjusted to 15 to 20 mm, so that the material particle size is reduced to 5 to 15 mm, forming angular recycled aggregates;

[0027] S04. Intelligent surface friction treatment: Surface treatment of the recycled aggregate after secondary crushing is performed using a mechanical friction method. The surface condition of the recycled aggregate is analyzed in real time using an intelligent recognition model for aggregate surface characteristics. The speed of the surface treatment machine is controlled to be 120 to 150 revolutions per minute based on the optimal speed parameter of the surface treatment machine output by the intelligent recognition model for aggregate surface characteristics. The operation time of the surface treatment machine is controlled to be 25 to 30 minutes based on the optimal operation time parameter of the surface treatment machine output by the intelligent recognition model for aggregate surface characteristics. Cement powder is added during the friction process at a mass ratio of 2% to 5% of the mass of the recycled aggregate to improve the surface density of the recycled aggregate.

[0028] S05. Intelligent sodium silicate impregnation modification: The surface-treated recycled aggregate is subjected to surface modification treatment. An optimal sodium silicate concentration parameter is determined using a material ratio optimization neural network. The recycled aggregate is immersed in a sodium silicate solution having a mass concentration of 5% to 8% determined by the optimal sodium silicate concentration parameter at 60 to 70° C. for 2 to 3 hours. The impregnation effect parameters of the impregnation treatment are monitored using a pore filling uniformity evaluation model. The sodium silicate solution is discharged and dried in an environment of 105 to 110° C. for 60 to 90 minutes.

[0029] S06, organic silicon hydrophobic coating treatment: transferring the dried recycled aggregate to a rotary coating machine, spraying an organic silicon hydrophobic agent with a mass concentration of 3% to 5%, controlling the rotation speed of the rotary coating machine to 80 to 100 revolutions per minute, and controlling the coating time of the rotary coating machine to 10 to 15 minutes, so that the coverage rate of the organic silicon hydrophobic agent on the surface of the recycled aggregate reaches more than 95%;

[0030] S07. Performance evaluation and parameter optimization iteration: The surface-modified recycled aggregate is subjected to performance testing to obtain apparent density parameters, water absorption parameters, crushing value parameters, and mud content parameters of the recycled aggregate. The apparent density parameters, water absorption parameters, crushing value parameters, mud content parameters, and impregnation effect parameters are input into a comprehensive performance balance point determination function to calculate comprehensive performance index parameters and processing parameter adjustment suggestion parameters of the recycled aggregate. Qualified recycled aggregates having a water absorption parameter lower than 3%, a crushing value parameter lower than 18%, and a mud content parameter lower than 1%, and whose comprehensive performance index parameters meet standard requirements, are screened based on the comprehensive performance index parameters. The processing parameter adjustment suggestion parameters are fed back to the aggregate surface feature intelligent recognition model, the material ratio optimization neural network, and the pore filling uniformity evaluation model for parameter optimization iteration.

[0031] The mechanical friction method specifically refers to using equipment to generate continuous friction between the recycled aggregates and between the recycled aggregates and the inner wall of the equipment, thereby removing loose particles on the surface of the recycled aggregates and smoothing out sharp edges and corners. At the same time, the added cement powder is partially hydrated under the action of frictional heat and fills the micropores on the surface of the recycled aggregates, forming a denser and stronger surface structure.

[0032] The surface modification treatment specifically refers to the process of changing the surface properties of the recycled aggregate by physical or chemical methods, using the sodium silicate solution to impregnate the surface of the recycled aggregate to form a sodium silicate gel network structure, fill the surface pores and enhance the interfacial bonding force, thereby reducing the water absorption and moisture migration capacity of the recycled aggregate;

[0033] The organosilicon hydrophobic agent specifically refers to a methyltriethoxysilane or octyltriethoxysilane organosilicon compound, which can form a hydrophobic film on the surface of the recycled aggregate, making it difficult for water molecules to penetrate into the interior of the recycled aggregate, thereby reducing the water absorption parameter of the recycled aggregate and improving the stability of the recycled aggregate in concrete.

[0034] The crushing value parameter specifically refers to the mass percentage of the recycled aggregate that is crushed after being subjected to a specified pressure under standard test conditions. It is an important indicator for evaluating the compressive strength of the recycled aggregate. The smaller the crushing value parameter, the higher the compressive strength of the recycled aggregate, and the more suitable it is for use in load-bearing structural concrete.

[0035] The mud content parameter specifically refers to the content of fine particles with a particle size of less than 0.075 mm in the recycled aggregate. Excessive mud content will affect the bonding performance and durability of concrete, so it needs to be strictly controlled within the specified range;

[0036] Among them, the aggregate surface feature intelligent recognition model specifically refers to an image recognition model based on a convolutional neural network structure, the aggregate surface feature intelligent recognition model collects the recycled aggregate surface image in real time through a high-definition camera, analyzes the recycled aggregate surface roughness parameter, the recycled aggregate surface pore distribution parameter and the recycled aggregate surface morphology characteristic parameter, and outputs the optimal surface treatment machine speed parameter and the optimal surface treatment machine operation time parameter; the specific structure of the aggregate surface feature intelligent recognition model is a deep neural network consisting of an input layer, five convolutional layers, three fully connected layers and an output layer, wherein the convolutional layer uses convolution kernels of different sizes to extract the multi-scale features of the recycled aggregate surface, the fully connected layer integrates the multi-scale features and maps them into the optimal surface treatment machine speed parameter and the optimal surface treatment machine operation time parameter, and the attention mechanism weight parameter in the aggregate surface feature intelligent recognition model is determined by the sum of the product of the surface treatment machine speed and 0.15 and the product of the cement powder addition mass ratio and 12; The step of establishing a training data set for the intelligent recognition model of aggregate surface features specifically includes collecting a total of 10,000 recycled aggregate surface image samples from different sources and different processing stages, photographing each recycled aggregate surface image from multiple angles and annotating the recycled aggregate surface roughness parameter, the recycled aggregate surface pore distribution parameter and the ideal processing parameter, expanding the data set to 50,000 images using random rotation, scaling and brightness adjustment data enhancement methods, and dividing the data set into a training set, a validation set and a test set in a ratio of 8:1:1; the step of training the intelligent recognition model of aggregate surface features specifically includes optimizing the parameters of the intelligent recognition model of aggregate surface features using a small batch stochastic gradient descent method with a batch size of 64, the learning rate is initially set to 0.001 and dynamically adjusted using a cosine annealing strategy, using an early stopping strategy to prevent overfitting during training, and stopping training when the validation set loss function value does not significantly decrease after 10 consecutive training cycles. Finally, the accuracy of the intelligent recognition model of aggregate surface features on the test set reaches more than 92%;

[0037] Among them, the material ratio optimization neural network specifically refers to a recurrent neural network model for predicting the optimal sodium silicate concentration parameter and the optimal impregnation time parameter; the specific structure of the material ratio optimization neural network is a hybrid recurrent neural network structure including a gated recurrent unit layer and two long short-term memory network layers, the gated recurrent unit layer is responsible for processing the basic characteristic information of the recycled aggregate, and the long short-term memory network layer is responsible for establishing the correlation between the optimal sodium silicate concentration parameter, the optimal impregnation time parameter and the impregnation effect parameter, and the memory unit number parameter in the material ratio optimization neural network is jointly determined by the recycled aggregate particle size distribution range parameter and the recycled aggregate initial porosity parameter, specifically, the memory unit number parameter is equal to the sum of the difference between the maximum value of the recycled aggregate particle size distribution range parameter and the minimum value of the recycled aggregate particle size distribution range parameter multiplied by 5 plus the recycled aggregate initial porosity parameter multiplied by 100; the step of establishing the training data set of the material ratio optimization neural network Specifically, the method comprises screening out 2,000 sets of treatment effect data under different recycled aggregate types, different sodium silicate concentrations and different impregnation time conditions from laboratory historical data, wherein the data include the recycled aggregate type, the recycled aggregate particle size distribution range parameter, the recycled aggregate initial porosity parameter, the sodium silicate concentration, the impregnation temperature, the impregnation time, the water absorption parameter after treatment and the strength improvement rate parameter, and using the Monte Carlo method to generate an additional 8,000 sets of simulation data to supplement the data set, which are merged to form a complete 10,000 sets of training data; the step of training the material ratio optimization neural network comprises using the mean square error as the loss function, using the adaptive moment estimation optimization algorithm for parameter optimization, setting the learning rate to 0.0005, setting the training cycle to 500, and performing a decay adjustment on the learning rate every 100 cycles; adding input noise during the training process to improve the generalization ability of the material ratio optimization neural network, and finally controlling the prediction error of the material ratio optimization neural network for the optimal sodium silicate concentration parameter to be within 0.3%;

[0038] Among them, the pore filling uniformity evaluation model specifically refers to an image segmentation neural network used to monitor the penetration of the sodium silicate solution during the impregnation of the recycled aggregate in real time; the specific structure of the pore filling uniformity evaluation model is a fully convolutional neural network based on the UNet architecture, which includes four layers of downsampling convolution modules and corresponding four layers of upsampling convolution modules, each of which is composed of two 3×3 convolution layers and a maximum pooling layer, and each of which is composed of a transposed convolution layer and two 3×3 convolution layers. The jump connection weight parameter in the pore filling uniformity evaluation model is determined by the sum of the product of the mass concentration of the sodium silicate solution and 0.25 and the product of the impregnation treatment temperature and 0.05; the step of establishing the training data set of the pore filling uniformity evaluation model specifically includes collecting 5000 images of different fillings. The recycled aggregate slice images in the filling stage are annotated by materials experts with the segmentation masks of the filled and unfilled areas in the recycled aggregate slice images. The recycled aggregate slice images are rotated, flipped and color-transformed, and data enhancement processing is performed to expand the number to 20,000. They are divided into training set, validation set and test set in a ratio of 7:2:1; the steps of training the pore filling uniformity assessment model specifically include using weighted cross entropy as the loss function, assigning different weights to filled areas and unfilled areas to solve the category imbalance problem, and using momentum gradient descent method for optimization. The learning rate is set to 0.001, the momentum coefficient is set to 0.9, and the training cycle is set to 200. Batch normalization is used during the training process to accelerate convergence and improve stability. Finally, the pixel-level accuracy of the pore filling uniformity assessment model on the test set reaches more than 95%;

[0039] The comprehensive performance balance point determination function is used to calculate the comprehensive performance index parameters of the recycled aggregate and the processing parameter adjustment suggestion parameters, and the input includes the water absorption parameter, the apparent density parameter, the crushing value parameter, the mud content parameter and the impregnation effect parameter, and the output is the comprehensive performance index parameters and the processing parameter adjustment suggestion parameters; the comprehensive performance balance point determination function first calculates the basic comprehensive index parameters by the weighted average method, wherein the water absorption parameter weight is 0.4, the apparent density parameter weight is 0.2, the crushing value parameter weight is 0.3, and the mud content parameter weight is 0.1, and then the basic comprehensive index parameters are corrected according to the impregnation effect parameter, and the correction coefficient is 0. The number is 1 plus the difference between the impregnation effect parameter and the standard impregnation effect parameter divided by the standard impregnation effect parameter. Finally, the corrected comprehensive performance index parameter is compared with the target threshold. If it is lower than the target threshold, the processing parameter adjustment suggestion parameter is generated. The processing parameter adjustment suggestion parameter includes the surface treatment machine speed increase or decrease value, the cement powder addition ratio adjustment value, the sodium silicate concentration adjustment value and the impregnation time adjustment value. The processing parameter adjustment suggestion parameter is used as a new input of the aggregate surface feature intelligent recognition model, the material ratio optimization neural network and the pore filling uniformity evaluation model to continue iterative optimization until the comprehensive performance index parameter reaches the target threshold or the number of iterations reaches the preset upper limit.

[0040] The specific implementation of the above steps is described in detail below.

[0041] The specific implementation of step S01 utilizes a vibrating screening system to precisely classify construction waste based on the principles of physical grading and mechanical sorting. First, the collected construction waste is transported to a vibrating screening system equipped with multiple layers of screens, with a top layer of 50 mm pores and a middle layer of 10 mm pores. The vibrating screening system uses an eccentric drive to generate three-dimensional vibrations, with a controlled frequency of 960 to 1080 vibrations per minute and an amplitude of 4 to 6 mm to ensure sufficient dispersion and efficient passage of the material through the screens. The vibrating screening process separates the construction waste into three particle sizes: over 50 mm, 10 to 50 mm, and under 10 mm. The classified material is then processed for impurity removal using a manual sorting table combined with an automated magnetic separator. The magnetic separator utilizes a permanent magnet with a strength of 0.5 to 0.8 Tesla, providing a magnetic field strong enough to capture ferrous debris from the construction waste. Non-magnetic impurities such as wood, plastic, and paper are removed using near-infrared recognition combined with manual sorting. After sorting, the purity of construction waste reaches over 95%, meeting the quality requirements of raw materials for subsequent processing. This step lays the foundation for the entire process, ensuring the basic quality of recycled aggregate through strict raw material sorting.

[0042] The specific implementation of step S02 utilizes mechanical crushing principles, using a twin-shaft crusher to perform controlled primary crushing of the sorted 10-50 mm particle size. The twin-shaft crusher is equipped with two counter-rotating gear shafts, each fitted with alloy steel crushing teeth with a hardness of 58-62 HRC. During the crushing process, the twin-shaft crusher's speed is controlled at 60-80 rpm, and the gear shaft spacing is adjusted to 25-30 mm to generate sufficient compression and shear forces. After entering the crushing chamber, the material is crushed by a combination of compression, shearing, and bending in the meshing area between the two shafts. The crushed material is discharged through a 15 mm wide discharge port and then enters a vibrating screening system for particle size classification. This screening system uses a 25 mm mesh to separate the crushed material into acceptable and oversized fractions. Oversized material is returned to the twin-shaft crusher via a closed conveyor system for further crushing, creating a closed-loop crushing process. The primary crushing process not only reduces the particle size of construction waste to 10 to 25 mm, but also removes some of the cement mortar adhering to the aggregate surface through extrusion, improving the quality of the subsequent recycled aggregate. The key to this step is to control the crusher speed and discharge particle size, and through appropriate equipment parameter settings, achieve initial crushing without over-crushing the material.

[0043] The specific implementation of step S03 is based on the principles of impact mechanics, using an impact crusher to perform secondary fine crushing on the material after primary crushing. The impact crusher is equipped with a high-speed rotating rotor, with multiple wear-resistant alloy impact hammers mounted on its surface. After entering the crushing chamber, the material is first struck by the high-speed rotating hammers, acquiring high-speed kinetic energy. It then strikes the material impact plate fixed to the machine wall, creating numerous cracks within the material during the impact process, achieving self-crushing. The impact crusher rotor linear speed is set to 35 to 45 meters per second, and the gap between the material and the impact plate is adjusted to 15 to 20 millimeters. During the crushing process, inter-material collisions are also a key factor. By adjusting the feed rate to 2 to 3 tons per hour, the material concentration in the crushing chamber is maintained at a moderate level, increasing the chance of inter-material collisions. The impact crusher's crushing mechanism is primarily based on the brittle nature of the material. The impact load causes the material to break along its internal microcracks and weak points, thereby preserving a relatively intact aggregate structure while removing most of the attached material. The recycled aggregates produced by this crushing method generally have an angular shape, a particle size range of 5 to 15 mm, a relatively rough surface, and good mechanical interlocking ability, making them suitable for use as concrete aggregate.

[0044] The specific implementation method of step S04 is to perform surface treatment on the recycled aggregate after secondary crushing based on the principle of tribology combined with artificial intelligence recognition technology. First, the aggregate is transported to the surface treatment machine, which adopts a rotating friction barrel design with the inner wall inlaid with wear-resistant rubber material. During the friction treatment process, the recycled aggregates collide with each other in the rotating barrel and rub against the barrel wall, removing loose particles on the surface and smoothing sharp edges. At the same time, cement powder with a mass ratio of 2% to 5% of the mass of the recycled aggregate is added to the treatment barrel. The particle size of the cement powder is less than 45 microns, and 42.5 grade ordinary Portland cement is used. The intelligent recognition model of aggregate surface features uses a high-definition camera installed on the processor to collect aggregate surface images in real time and analyze surface roughness parameters and pore distribution parameters. The recognition model consists of a deep neural network consisting of an input layer, five convolutional layers, three fully connected layers, and an output layer. The convolutional layers use kernels of varying sizes (3×3, 5×5, 3×3, 3×3, and 3×3, respectively) of 32, 64, 128, 256, and 512 to extract multi-scale features from the aggregate surface. The fully connected layers have 1024, 512, and 256 neurons, respectively. The ReLU activation function is used to integrate multi-scale features and map them to optimal surface treatment machine speed and operating time parameters. The attention mechanism weights in this model are determined by the sum of the product of the surface treatment machine speed and 0.15 and the product of the cement powder addition mass ratio and 12. The steps for establishing a training dataset for the intelligent aggregate surface feature recognition model include collecting 10,000 recycled aggregate surface image samples from different sources and at different processing stages. Each recycled aggregate surface image is photographed from multiple angles and annotated with the recycled aggregate surface roughness parameters, recycled aggregate surface pore distribution parameters, and ideal processing parameters. The dataset is expanded to 50,000 images using data augmentation methods such as random rotation, scaling, and brightness adjustment. The dataset is then divided into training, validation, and test sets in an 8:1:1 ratio. The training steps for the intelligent aggregate surface feature recognition model include optimizing the model parameters using mini-batch stochastic gradient descent with a batch size of 64. The learning rate is initially set to 0.001 and dynamically adjusted using a cosine annealing strategy. An early stopping strategy is used during training to prevent overfitting. Training is terminated when the validation set loss function value shows no significant decrease after 10 consecutive training cycles. The accuracy of the intelligent aggregate surface feature recognition model on the test set ultimately exceeds 92%. Based on the analysis results of the intelligent recognition model, the surface treatment machine speed is automatically adjusted to 120 to 150 rpm, and the treatment time is 25 to 30 minutes. The heat generated during the friction process causes some cement powder to hydrate, filling the micropores on the aggregate surface and forming a dense surface layer, thereby improving the surface strength and density of the aggregate.

[0045] The specific implementation of step S05 is based on the principle of sodium silicate gel chemical reaction, combined with a neural network algorithm to deeply modify the recycled aggregate after surface treatment. First, the material ratio optimization neural network is used to analyze the characteristics of the recycled aggregate and determine the optimal sodium silicate concentration parameter. The neural network specifically adopts a hybrid recurrent neural network structure including a gated recurrent unit layer and two long short-term memory network layers. The gated recurrent unit layer is configured with 128 neurons, which is responsible for processing the basic characteristic information of the recycled aggregate. The two long short-term memory network layers are each configured with 256 neurons, which are responsible for establishing the correlation between the optimal sodium silicate concentration parameter, the optimal impregnation time parameter and the impregnation effect parameter. The number of memory units in the material ratio optimization neural network is jointly determined by the recycled aggregate particle size distribution range parameter and the recycled aggregate initial porosity parameter. Specifically, the number of memory units parameter is equal to the difference between the maximum value of the recycled aggregate particle size distribution range parameter and the minimum value of the recycled aggregate particle size distribution range parameter multiplied by 5 plus the sum of the recycled aggregate initial porosity parameter multiplied by 100. The training dataset for the material ratio optimization neural network was established by selecting 2,000 sets of treatment effect data from historical laboratory data, including parameters for recycled aggregate type, recycled aggregate particle size distribution range, recycled aggregate initial porosity, sodium silicate concentration, impregnation temperature, impregnation time, post-treatment water absorption, and strength gain. An additional 8,000 sets of simulated data were generated using the Monte Carlo method to supplement the dataset, which was then combined to form a complete training dataset of 10,000 sets. The training of the material ratio optimization neural network involved using mean squared error as the loss function and an adaptive moment estimation optimization algorithm for parameter optimization. The learning rate was set to 0.0005, the training cycle was set to 500, and the learning rate was decayed every 100 cycles. Input noise was added during training to improve the generalization ability of the material ratio optimization neural network. Ultimately, the prediction error of the material ratio optimization neural network for the optimal sodium silicate concentration parameter was kept within 0.3%. According to the output of the neural network, the recycled aggregate was immersed in a sodium silicate solution with a mass concentration of 5% to 8%. The solution was prepared using a modulus of 3.1 to 3.3 and a density of 1.38 to 1.42 g / cm 3The aggregate is diluted with industrial-grade water glass. The impregnation treatment is carried out at a constant temperature of 60 to 70°C for 2 to 3 hours. During the impregnation process, the sodium silicate solution penetrates the pores on the surface of the aggregate and reacts with the calcium ions in the aggregate to form calcium silicate gel. The gel gradually hardens and fills the pores on the surface of the aggregate. The entire impregnation process is monitored in real time by a pore filling uniformity assessment model. The model is based on a fully convolutional neural network with a UNet architecture, consisting of four layers of downsampling convolution modules and corresponding four layers of upsampling convolution modules. Each downsampling convolution module consists of two 3×3 convolution layers and a maximum pooling layer, and each upsampling convolution module consists of a transposed convolution layer and two 3×3 convolution layers. The jump connection weight parameter in the pore filling uniformity assessment model is determined by the sum of the product of the mass concentration of the sodium silicate solution and 0.25 and the product of the impregnation treatment temperature and 0.05. The training dataset for the pore filling uniformity assessment model was established by collecting 5,000 images of recycled aggregate slices at different filling stages. Materials experts then annotated the segmentation masks of filled and unfilled areas within the recycled aggregate slices. The images were then augmented with rotation, flipping, and color transformation to expand the dataset to 20,000 images. The dataset was then divided into training, validation, and test sets in a ratio of 7:2:1. The training of the pore filling uniformity assessment model involved using weighted cross entropy as the loss function, assigning different weights to filled and unfilled areas to address class imbalance. Optimization was performed using momentum gradient descent with a learning rate of 0.001, a momentum coefficient of 0.9, and a training cycle of 200. Batch normalization was used during training to accelerate convergence and improve stability. The pore filling uniformity assessment model achieved pixel-level accuracy exceeding 95% on the test set. After the impregnation is completed, the sodium silicate solution is drained and the recycled aggregate is transferred to an oven and dried at 105 to 110°C for 60 to 90 minutes to ensure that the gel network formed by the impregnation is completely dried and solidified.

[0046] The specific implementation of step S06 is to perform surface hydrophobic modification on the dried recycled aggregate based on the organosilicon hydrophobic principle. The recycled aggregate is transferred to a rotary coating machine, which adopts an inclined rotating drum structure with multiple sets of lifting plates installed on the inner wall, and the drum inclination angle is set to 15 to 20 degrees. The coating machine is started, the drum speed is controlled to 80 to 100 revolutions per minute, and at the same time, an organosilicon hydrophobic agent with a mass concentration of 3% to 5% is sprayed into the drum through a high-pressure atomizing nozzle. The organosilicon hydrophobic agent used is methyltriethoxysilane or octyltriethoxysilane. The concentration is controlled by solvent dilution, and the spraying rate is controlled at 15 to 25 ml per kilogram of aggregate. The coating process lasts for 10 to 15 minutes to ensure that the coverage rate of the organosilicon hydrophobic agent on the recycled aggregate surface reaches more than 95%. The coating principle is to use organosilicon compounds to form a molecular-level hydrophobic film on the aggregate surface. The silicon-oxygen bonds in the organosilicon molecules undergo condensation reaction with the hydroxyl groups on the aggregate surface to form a stable chemical bond, while the hydrophobic organic groups face outward to form a hydrophobic barrier. The silicone coating not only exhibits excellent hydrophobicity but also weather and chemical resistance, effectively preventing moisture and harmful substances from invading the aggregate. A high-precision humidity sensor monitors the surface humidity of the aggregate in real time. Coating is complete when the humidity stabilizes at 4-6%. The coated recycled aggregate is then left at room temperature for 24 hours to allow the silicone hydrophobic coating to fully cure. This step effectively reduces the recycled aggregate's water absorption by forming a durable hydrophobic layer on the aggregate surface, improving its long-term stability in concrete.

[0047] The specific implementation of step S07 is to conduct a comprehensive performance evaluation and parameter optimization iteration on the surface-modified recycled aggregate based on the multi-parameter optimization theory. First, the apparent density parameter, water absorption parameter, crushing value parameter and mud content parameter of the recycled aggregate are measured according to the standard test method. The apparent density is measured by the volumetric flask drainage method, and the typical value should be between 2400 and 2600 kg / m 3The water absorption rate is determined by the drying method and is required to be within 3%; the crushing value is determined by the standard crushing value test and is required to be less than 18%; and the mud content is determined by the water washing and screening method and is required to be less than 1%. The values ​​of each parameter obtained from the test are input into the comprehensive performance balance point determination function. This function first calculates the basic comprehensive index parameters using the weighted average method, with the following weights set: water absorption parameter weight 0.4, apparent density parameter weight 0.2, crushing value parameter weight 0.3, and mud content parameter weight 0.1. The basic comprehensive index parameters are then corrected based on the impregnation effect parameters. The correction factor is 1 plus the difference between the impregnation effect parameter and the standard impregnation effect parameter, divided by the standard impregnation effect parameter. The reference value of the standard impregnation effect parameter is 0.85, indicating that the pore filling rate during the impregnation process reaches 85%. The calculated comprehensive performance index parameters are compared with the target threshold of 0.75. If they are below the target threshold, recommended parameters for treatment parameter adjustments are generated. The recommended parameters for adjusting the processing parameters include the increase or decrease value of the surface treatment machine speed, the adjustment value of the cement powder addition ratio, the adjustment value of the sodium silicate concentration and the impregnation time. These recommended adjustment parameters serve as new inputs for the intelligent recognition model of aggregate surface characteristics, the material ratio optimization neural network and the pore filling uniformity evaluation model to start the next round of optimization iteration. The entire optimization iteration process adopts the Bayesian optimization algorithm to estimate the optimal parameter combination and improvement direction by constructing a probability model of the parameter space. The iterative process sets the maximum number of iterations to 8 times, and each iteration adjusts the processing parameters by 5-10%. After multiple rounds of iterative optimization, the recycled aggregate with comprehensive performance index parameters reaching the target threshold is finally obtained as the final qualified product. This step is based on the closed-loop feedback control principle to achieve continuous optimization of the recycled aggregate production process and improve product quality stability and yield.

[0048] The core technical idea of ​​the present invention is described in detail below.

[0049] 1. Intelligent surface feature recognition and adaptive parameter control: The aggregate surface feature intelligent recognition model adopted by the present invention collects images of the recycled aggregate surface in real time through a high-definition camera, and applies a deep convolutional neural network to analyze the surface roughness, pore distribution and morphological feature parameters, thereby achieving precise guidance on surface treatment parameters. This intelligent recognition technology based on machine vision replaces the parameter setting method that relies on manual experience in traditional processes, and realizes a fundamental transformation from subjective experience judgment to objective data analysis. In the traditional recycled aggregate preparation process, the adjustment of process parameters often depends on the experience of the operator, and it is difficult to accurately control construction waste of different sources and properties, resulting in large fluctuations in product quality. The intelligent recognition technology of the present invention can quickly analyze the surface features of the aggregate and automatically output the optimal processing parameters, so that the surface treatment process is always in an ideal state, ensuring the consistency and stability of the surface quality of the recycled aggregate, and effectively solving the technical bottleneck that the traditional process cannot cope with the diversity and complexity of raw materials.

[0050] 2. Multi-stage surface modification synergistic mechanism: The present invention constructs a multi-stage surface modification system of mechanical friction treatment, sodium silicate impregnation modification and silicone hydrophobic coating, forming an all-round surface strengthening mechanism from physical structure optimization to chemical property improvement. The mechanical friction method combined with cement powder addition utilizes the friction heat effect to promote partial hydration of cement, fill the micropores on the aggregate surface and smooth the edges and corners; sodium silicate impregnation fills the surface pores by forming a sodium silicate gel network structure, thereby enhancing the interfacial bonding force; silicone hydrophobic coating ultimately forms a molecular-level hydrophobic film on the aggregate surface to block water penetration. This multi-stage synergistic treatment mechanism realizes a comprehensive modification of the recycled aggregate surface from the microscopic to the macroscopic level. Unlike the limitations of traditional single treatment methods that can only solve part of the problem, the multi-stage surface modification system of the present invention can simultaneously solve multiple defects of the recycled aggregate surface, such as looseness, high porosity, and strong water absorption, forming a layer-by-layer progressive and mutually synergistic surface strengthening effect, which greatly improves the overall performance and stability of the recycled aggregate.

[0051] 3. Closed-loop parameter optimization iterative system: The present invention establishes a closed-loop parameter optimization iterative system dominated by a comprehensive performance balance point determination function, and realizes an automated feedback mechanism from aggregate performance detection to process parameter adjustment. The system receives multi-dimensional performance data of recycled aggregates, such as apparent density parameters, water absorption parameters, crushing value parameters and mud content parameters, generates comprehensive performance index parameters through weighted calculation, and outputs parameter adjustment suggestions after comparison with the target threshold, forming feedback adjustments to each front-end intelligent model. The traditional recycled aggregate preparation process lacks a systematic performance evaluation and parameter optimization mechanism, and adjustments are often isolated and empirical, making it difficult to achieve multi-parameter collaborative optimization. The closed-loop optimization system of the present invention organically connects each process link to form a complete information flow and control flow, so that the entire preparation process can continuously self-adjust and optimize according to real-time feedback on aggregate performance, ensuring the precise controllability of the final product performance and stable quality.

[0052] The present invention forms an organically unified whole through the three core technical ideas of intelligent surface feature recognition and parameter adaptive control, multi-level surface modification synergistic efficiency mechanism and closed-loop parameter optimization iterative system, and jointly constructs a set of recycled aggregate performance precision control system. Intelligent recognition technology provides precise parameter guidance for multi-level surface modification, so that the surface treatment is always in the best state; multi-level surface modification comprehensively improves the surface quality and performance of recycled aggregate through the synergistic effect of physical and chemical means; the closed-loop parameter optimization system conducts a comprehensive evaluation of the treatment results and feedback adjustment, so that the entire preparation process continues to approach the target performance. The close cooperation of these three technical ideas breaks through the technical bottlenecks of traditional recycled aggregate preparation technology, such as strong dependence on experience, unstable treatment effect, and inability to accurately control, and realizes a technological leap from empirical and extensive to intelligent and precise, so that each process parameter and material property in the recycled aggregate preparation process always maintains the best matching state, ensuring the high level and consistency of the final product quality, and providing reliable technical support for the application of recycled aggregate in high-performance concrete.

[0053] Specifically, the principle of this invention is: Based on the theory of material surface engineering and intelligent parameter optimization, this invention constructs a system for precisely controlling the properties of recycled aggregates. Its core principle can be explained from the following aspects:

[0054] First, the intelligent aggregate surface feature recognition model employed in this invention is based on a convolutional neural network. Using a high-definition camera, it captures real-time images of recycled aggregate surfaces and analyzes their surface roughness, pore distribution, and topographical parameters. Trained on 50,000 sample images, the model achieves an accuracy rate exceeding 92%. It can accurately identify the differences in surface properties of recycled aggregates from different sources and output optimal treatment parameters. This overcomes the inaccurate parameter settings often associated with relying on manual judgment in traditional processes, enabling intelligent control of the surface treatment process.

[0055] Secondly, the mechanical friction method of the present invention, in synergistic with the cement powder, fully utilizes the frictional heat effect. During the friction process, the surface temperature of the recycled aggregate increases, promoting the partial hydration of the added cement powder, forming a calcium-silicon hydrated gel that fills the surface micropores. Simultaneously, the continuous mechanical friction removes loose surface particles, smoothes sharp edges, and improves the aggregate's particle morphology. This process not only increases the aggregate's surface density but also optimizes its geometry, laying a good foundation for subsequent chemical treatment.

[0056] Furthermore, the sodium silicate impregnation and organosilicon hydrophobic coating used in the present invention form a dual protection system. During the impregnation process, sodium silicate penetrates into the pores on the aggregate surface and reacts with calcium ions in the aggregate to form an insoluble calcium silicate gel, which fills the surface pores and enhances interfacial bonding. The organosilicon hydrophobic agent forms a hydrophobic film on the aggregate surface. Through the arrangement of hydrophobic groups at the molecular scale, it changes the wettability of the aggregate surface, making it difficult for water molecules to penetrate into the interior of the aggregate. The synergistic effect of the material ratio optimization neural network and the pore filling uniformity evaluation model ensures the precise control and effect evaluation of the chemical treatment process.

[0057] Crucially, the present invention establishes a closed-loop iterative parameter optimization mechanism. The comprehensive performance balance point determination function receives multiple performance parameters as input, generates a comprehensive performance index through weighted calculation, and outputs parameter adjustment recommendations after comparing them with the target threshold. These recommendations are fed back to the front-end intelligent models, forming a closed-loop optimization system that ensures that the performance of the recycled aggregate continuously approaches the target value. This mechanism effectively solves the difficulty of dynamically adjusting process parameters based on aggregate performance in traditional processes, enabling precise regulation of recycled aggregate performance.

[0058] In summary, the present invention constructs a system for precisely controlling the properties of recycled aggregates by organically combining intelligent recognition technology with multi-stage surface modification treatment, thereby solving the technical problems of unstable quality, high water absorption and difficulty in precisely controlling the properties of recycled aggregates.

[0059] A specific embodiment 1 of the present invention is provided below. The specific implementation of each step in this embodiment 1 is described in detail as follows.

[0060] The specific implementation of step S01 is based on the principles of physical classification and mechanical sorting, and uses vibration screening equipment to accurately classify and process construction waste. First, the collected construction waste is transported to the vibration screening system, which is equipped with multiple layers of screens. The top screen has an aperture of 50 mm and the middle screen has an aperture of 10 mm. The vibration screening equipment uses an eccentric wheel drive to generate three-dimensional vibration, and the vibration screening efficiency η s It can be expressed as:

[0061]

[0062] Where η s is the screening efficiency, in %; m p The mass of the correctly graded material is in kilograms; m tis the total feed mass in kilograms. The target value of screening efficiency should reach more than 95%, which is achieved by adjusting the vibration parameters. The vibration screening process divides the construction waste into three particle sizes: above 50 mm, 10 to 50 mm, and below 10 mm. The classified materials are then treated for impurities by a manual sorting table combined with an automated magnetic separation device. The magnetic separation device is designed with a permanent magnet with a strength of 0.5 to 0.8 Tesla, and the magnetic field strength is sufficient to capture iron debris in construction waste. For non-magnetic impurities such as wood, plastic, paper, etc., near-infrared recognition combined with manual sorting is used for removal. After sorting, the purity P of construction waste can be expressed as:

[0063]

[0064] Where, P is the purity of construction waste, unit is %; m c is the mass of materials such as masonry, concrete, etc. that can be used to prepare recycled aggregates, in kilograms; m t = Total material mass, expressed in kilograms. After treatment, construction waste should have a purity of at least 95%, meeting the quality requirements for subsequent processing. This step lays the foundation for the entire process, ensuring the essential quality of the recycled aggregate through rigorous raw material sorting.

[0065] The specific implementation method of step S02 is to adopt the principle of mechanical crushing and use a double-shaft crusher to carry out controlled primary crushing of the 10 to 50 mm particle size construction waste after sorting. The double-shaft crusher is equipped with two relatively rotating gear shafts, on which alloy steel crushing teeth with a hardness of 58 to 62HRC are installed. During the crushing process, the speed of the double-shaft crusher is controlled to 60 to 80 revolutions per minute, and the gear shaft spacing is adjusted to 25 to 30 mm to generate sufficient extrusion and shear force. The crushed material is discharged through a 15 mm wide discharge port and then enters the vibrating screening system for particle size classification. The screening system uses a 25 mm aperture screen to divide the crushed material into qualified products and oversized products. The oversized materials are returned to the double-shaft crusher through a closed conveying system for re-crushing, forming a closed-loop crushing process. The circulating load rate L in the closed-loop crushing c It can be expressed as:

[0066]

[0067] Where, L c is the cycle load rate, in %; m r is the mass of the returned crushed material, in kilograms; m fis the mass of fresh feed material in kilograms. The recycle load rate should ideally be controlled within the range of 30-50%. The primary crushing process not only reduces the particle size of construction waste to 10-25 mm, but also removes some cement mortar adhering to the aggregate surface through extrusion, improving the quality of the subsequent recycled aggregate. The key to this step lies in controlling the crusher speed and discharge particle size, ensuring initial crushing without over-crushing through appropriate equipment parameter settings.

[0068] The specific implementation of step S03 is based on the principles of impact mechanics, using an impact crusher to perform secondary fine crushing on the material after primary crushing. The impact crusher is equipped with a high-speed rotating rotor, and multiple wear-resistant alloy impact hammers are mounted on the rotor surface. After entering the crushing chamber, the material is first struck by the high-speed rotating hammers, acquiring high-speed kinetic energy. It then strikes the material impact plate fixed to the machine wall, creating numerous cracks within the material during the impact process, achieving self-crushing. The impact crusher rotor linear speed is set to 35 to 45 meters per second, and the gap between the material and the impact plate is adjusted to 15 to 20 mm.

[0069] By adjusting the feed rate to 2-3 tons per hour, the material concentration in the crushing chamber is ensured to be moderate, which increases the chance of collision between materials. r The calculation formula is:

[0070]

[0071] Where R r is the particle size reduction ratio, dimensionless; d in is the average particle size of the feed, in millimeters; d out is the average particle size of the output material, expressed in millimeters. After secondary crushing, the particle size reduction ratio is controlled within a range of 1.5 to 2.5, and the output particle size range is controlled within a range of 5 to 15 mm. The crushing mechanism of the impact crusher is primarily based on the brittle nature of the material. The impact load causes the material to fracture along its internal microcracks and weak surfaces, thereby preserving a relatively intact aggregate structure while removing most attachments. The recycled aggregate produced by this crushing method generally has an angular shape and a relatively rough surface, exhibiting good mechanical interlocking properties, making it suitable for use as a concrete aggregate.

[0072] Step S04 utilizes tribological principles combined with artificial intelligence (AI) recognition technology to surface treat the recycled aggregate after secondary crushing. The aggregate is first conveyed to a surface treatment machine, which utilizes a rotating friction drum with a wear-resistant rubber inner wall. During the friction treatment process, the recycled aggregate collides with each other and rubs against the drum wall, removing loose surface particles and smoothing out sharp edges.

[0073] At the same time, cement powder with a mass ratio of 2% to 5% of the mass of the recycled aggregate is added into the processing barrel. The particle size of the cement powder is less than 45 microns, and 42.5 grade ordinary Portland cement is used.

[0074] Frictional heat causes local temperature rise, promoting hydration reaction on the surface of cement particles and forming a stronger adhesion layer. The intelligent recognition model of aggregate surface features uses a high-definition camera installed on the processor to collect aggregate surface images in real time and analyze surface roughness parameters and pore distribution parameters. The attention mechanism weight parameter W in this model is a Determined by the following formula:

[0075] W a =0.15·n t +12·R c ;

[0076] Where W a is the attention mechanism weight parameter, dimensionless; n t is the surface treatment machine speed, in revolutions per minute; R c The mass ratio of cement powder added is expressed in %. Based on the analysis results of the intelligent recognition model, the system automatically adjusts the surface treatment machine speed to 120 to 150 rpm and the treatment time to 25 to 30 minutes. The heat generated during the friction process causes some cement powder to hydrate, filling the micropores on the aggregate surface and forming a dense surface layer, thereby improving the aggregate's surface strength and density.

[0077] The specific implementation of step S05 is based on the chemical reaction principle of sodium silicate gel, combined with the neural network algorithm to deeply modify the recycled aggregate after surface treatment. First, the material ratio optimization neural network is used to analyze the characteristics of the recycled aggregate and determine the optimal sodium silicate concentration parameter. The number of memory units in the neural network is parameter N. m Determined by the following formula:

[0078] N m =5·(d max -d min )+100·P0;

[0079] Where N m is the memory unit quantity parameter, the unit is piece; d max is the maximum value of the recycled aggregate particle size distribution range, in millimeters; d min is the minimum value of the recycled aggregate particle size distribution range, in millimeters; P0 is the initial porosity parameter of the recycled aggregate, in 1. According to the output of the neural network, the recycled aggregate was immersed in a sodium silicate solution with a mass concentration of 5% to 8%. The penetration depth of sodium silicate on the aggregate surface and in the pores during the immersion process, D p It can be expressed as:

[0080]

[0081] Where D p is the penetration depth in millimeters; k p is the permeability coefficient, ranging from 0.8 to 1.2, which is related to the pore structure of the aggregate; t is the immersion time, in hours; γ is the surface tension of the liquid, in Newtons per meter; θ is the contact angle between the liquid and the aggregate, in degrees; r is the average pore radius, in micrometers; η is the viscosity of the liquid, in Pa·s. The reaction rate R of the sodium silicate solution with the calcium ions in the aggregate r It can be expressed as:

[0082]

[0083] Where R r is the reaction rate in moles per liter per hour; k r is the reaction rate constant, which is 10 4 ~10 5 liters per mole per hour; is the sodium silicate concentration in moles per liter; is the calcium ion concentration in moles per liter; E a is the activation energy, ranging from 20 to 30 kilojoules per mole; R is the gas constant, 8.314 joules per mole per kelvin; T is the reaction temperature, in kelvins, ranging from 333 to 343 kelvins (i.e., 60 to 70°C). The chemical reaction formula for the reaction of sodium silicate with calcium ions in aggregate to form calcium silicate gel is:

[0084] Na2SiO3+Ca(OH)2+nH2o→CaO·SiO2·nH2O+2NaOH;

[0085] The calcium silicate gel formed during the reaction gradually fills the pores on the aggregate surface and increases the surface density. p It can be expressed as:

[0086]

[0087] Where K p is the solution permeability, in Darcy; k is the aggregate pore permeability coefficient, with a value range of 10 -12 ~10 -14 square meters; ρ is the solution density, in kilograms per cubic meter; g is the acceleration of gravity, which is 9.8 meters per square second; μ is the solution dynamic viscosity, in Pa·s. The jump connection weight parameter W in the pore filling uniformity evaluation model s Determined by the following formula:

[0088]

[0089] Where W s is the jump connection weight parameter, dimensionless; is the mass concentration of the sodium silicate solution, expressed in %, and T is the impregnation temperature, expressed in degrees Celsius. After impregnation, the sodium silicate solution is drained and the recycled aggregate is transferred to an oven and dried at 105 to 110°C for 60 to 90 minutes to ensure that the gel network formed by the impregnation is completely dry and solidified.

[0090] The specific implementation of step S06 is to perform surface hydrophobic modification on the dried recycled aggregate based on the principle of organic silicon hydrophobicity. The recycled aggregate is transferred to a rotary coating machine, which adopts an inclined rotary drum structure with multiple sets of lifting plates installed on the inner wall. The drum inclination angle is set to 15 to 20 degrees. Start the coating machine, control the drum speed to 80 to 100 revolutions per minute, and at the same time, spray an organic silicon hydrophobic agent with a mass concentration of 3% to 5% into the drum through a high-pressure atomizing nozzle. The coating uniformity U c It can be expressed as:

[0091]

[0092] Where U c is the coating uniformity, in %; σ c is the standard deviation of the coating thickness on the aggregate surface, in microns; is the average coating thickness in microns. The target uniformity is 90% or higher. The organosilicon hydrophobic agent used is methyltriethoxysilane or octyltriethoxysilane. The concentration is controlled by solvent dilution, and the spray rate is controlled at 15-25 ml per kilogram of aggregate. The hydrolysis-condensation reaction equation of methyltriethoxysilane is:

[0093] (CH3)Si(OC2H5)3+3H2o→(CH3)Si(OH)3+3C2H5OH;

[0094] (CH3)Si(OH)3+3(≡Si-OH)→(≡Si-O-)3Si(CH3)+3H2O;

[0095] In the formula, ≡Si-OH represents the silanol group on the aggregate surface. The coating process lasts for 10 to 15 minutes to ensure that the coverage rate of the organosilicon hydrophobic agent on the recycled aggregate surface reaches more than 95%. The coating principle is to use organosilicon compounds to form a molecular-level hydrophobic film on the aggregate surface. The silicon-oxygen bonds in the organosilicon molecules condense with the hydroxyl groups on the aggregate surface to form a stable chemical bond, while the hydrophobic organic groups face outward to form a hydrophobic barrier. The thickness of the organosilicon coating is h. c It can be expressed as:

[0096]

[0097] Where h c is the coating thickness in microns; V s is the volume of the sprayed silicone solution, in milliliters; ρ s is the density of the solution in grams per milliliter; η d is the deposition efficiency, ranging from 0.7 to 0.85; S a is the total surface area of ​​aggregate, in square centimeters; ρ c The coating density is expressed in grams per cubic centimeter. The silicone coating not only exhibits excellent hydrophobicity but also weather and chemical resistance, effectively preventing moisture and harmful substances from invading the aggregate. A high-precision humidity sensor monitors the humidity changes on the aggregate surface in real time. When the humidity stabilizes at 4-6%, the coating is complete. The contact angle of the aggregate after coating is θ. c It can be expressed as:

[0098]

[0099] Where θ c is the water contact angle of the coated aggregate surface, in degrees; θ0 is the contact angle of the uncoated aggregate, usually 30 to 50 degrees; Δθ is the maximum contact angle increment, ranging from 60 to 80 degrees; k t is the time constant in hours -1 , the value range is 0.1~0.2; t c = is the curing time, measured in hours. The coated recycled aggregate is left at room temperature for 24 hours to allow the silicone hydrophobic coating to fully cure. This step effectively reduces the recycled aggregate's water absorption by forming a durable hydrophobic layer on the aggregate surface, improving its long-term stability in concrete.

[0100] The specific implementation of step S07 is to conduct a comprehensive performance evaluation and parameter optimization iteration on the surface-modified recycled aggregate based on the multi-parameter optimization theory. First, the apparent density parameter, water absorption parameter, crushing value parameter and mud content parameter of the recycled aggregate are measured according to the standard test method. The apparent density is measured by the volumetric flask drainage method, and the typical value should be between 2400 and 2600 kg / m 3 The water absorption rate is determined by the drying method and is required to be controlled below 3%. The crushing value is determined by the standard crushing value test and is required to be less than 18%. The mud content is determined by the water washing and screening method and is required to be less than 1%. The various parameter values ​​obtained from the test are input into the comprehensive performance balance point determination function, which first calculates the basic comprehensive index parameter I by the weighted average method. b :

[0101]

[0102] Where, I bis the basic comprehensive index parameter, dimensionless; WA is the water absorption parameter, unit is %; WA ref is the reference value of water absorption, which is taken as 5%; ρ is the apparent density parameter, the unit is kg / m 3 ρ ref The reference value of apparent density is 2650kg / m 3 ;CV is the crushing value parameter, unit is %;CV ref is the reference value of crushing value, which is taken as 30%; MC is the mud content parameter, the unit is %; MC ref is the reference value of mud content, which is 2%; w1, w2, w3, and W4 are the weights of each parameter, which are 0.4, 0.2, 0.3, and 0.1 respectively. Then, the basic comprehensive index parameters are modified according to the impregnation effect parameters to calculate the comprehensive performance index parameters I c :

[0103]

[0104] Where, I c is a comprehensive performance index parameter, dimensionless; E p is the impregnation effect parameter, indicating the pore filling rate, in %; E p，ref The calculated comprehensive performance index parameter is compared with the target threshold of 0.75. If it is lower than the target threshold, the processing parameter adjustment recommendation parameter is generated. The processing parameter adjustment recommendation parameter includes the surface treatment machine speed increase or decrease value Δn t , Cement powder addition ratio adjustment value ΔR c , sodium silicate concentration adjustment value ΔC s And the immersion time adjustment value Δt s , the calculation formulas are:

[0105] Δn t =k n (0.75-I c )·n t ;

[0106] ΔR c =k r (0.75-I c )·R c ;

[0107] ΔC s =k c (0.75-I c )·C s ;

[0108] Δt s =k t (0.75-I c )·ts ;

[0109] Where Δn t k is the increase or decrease value of the surface treatment machine speed, in revolutions per minute; n is the speed adjustment coefficient, the value range is 1.5~2.5; I c is a comprehensive performance index parameter, dimensionless; n t is the current surface treatment machine speed, in revolutions per minute; ΔR c The adjustment value of cement powder addition ratio, unit is %; k r R is the cement addition ratio adjustment coefficient, ranging from 2.0 to 3.0; c The current cement powder addition ratio, in %; ΔC s is the sodium silicate concentration adjustment value, in %; k c is the concentration adjustment coefficient, ranging from 1.0 to 2.0; C s is the current sodium silicate concentration, in %; Δt s is the immersion time adjustment value, in hours; k t is the time adjustment coefficient, ranging from 0.5 to 1.5; t s The current soaking time in hours.

[0110] These recommended adjustment parameters serve as new inputs to the aggregate surface feature intelligent recognition model, the material ratio optimization neural network, and the pore filling uniformity evaluation model, initiating the next round of optimization iterations. The entire optimization iteration process uses the Bayesian optimization algorithm to estimate the optimal parameter combination and improvement direction by constructing a probability model of the parameter space. The Bayesian optimization algorithm is based on Gaussian process regression. Its core idea is to use known parameter points to construct the probability distribution of the objective function, and then select the next parameter point to be evaluated based on the acquisition function. The objective function modeling uses the Gaussian process:

[0111]

[0112] Where f(x) is the objective function, i.e. the comprehensive performance index parameter I c ; x is a parameter vector, including the surface treatment machine speed n t , Cement powder addition ratio R c , sodium silicate concentration C s and immersion time t s ; m(x) is the mean function, which is usually set to a constant; k(x, x′) is the kernel function, which uses the radial basis function (RBF):

[0113]

[0114] Where, is the signal variance, which controls the variation of the function value and has a value of 0.1 to 0.5; l i is the characteristic length scale, which controls the smoothness of the i-th parameter and has a value range of 0.05 to 0.2; d is the parameter dimension, which is 4 here; x i and x i ′ are the i-th components of the parameter vectors x and x′ respectively.

[0115] The acquisition function uses the expected improvement (EI), which is expressed as:

[0116]

[0117] Where EI(x) is the expected improvement function value; f(x) is the predicted mean of the objective function; f(x + ) is the current optimal objective function value; σ(x) is the predicted standard deviation; Φ(·) and φ(·) are the cumulative distribution function and probability density function of the standard normal distribution, respectively.

[0118] The iterative process of Bayesian optimization is:

[0119] 1. Initialization: Randomly select n0 points from the parameter space (n0 is 4 to 6) and calculate the corresponding objective function value I c ;

[0120] 2. Fitting Gaussian process: Fitting the Gaussian process model based on all current observation points;

[0121] 3. Find a new point: Calculate the expected improvement EI(x) of all possible points in the parameter space and select the point with the largest EI(x) as the next point to be evaluated x new ;

[0122] 4. Evaluate the new point: calculate the new point x new The corresponding objective function value I c ;

[0123] 5. Update the data set: add the new points and their function values ​​to the observation data set;

[0124] 6. Determine the termination condition: if the maximum number of iterations (set to 8) or the objective function value I is reached c If the target threshold value of 0.75 is reached, the iteration is terminated; otherwise, return to step 2 to continue the iteration.

[0125] The adjustment range of processing parameters in each iteration is controlled within the range of 5-10% to avoid overshoot or instability caused by excessive parameter changes. The parameter constraints in the optimization process include:

[0126] 120≤n t ≤150;

[0127] 2%≤R c ≤5%;

[0128] 5%≤C s ≤8%;

[0129] 2h≤t s ≤3h;

[0130] Through multiple rounds of iterative optimization, the final product is a recycled aggregate with comprehensive performance indicators that meet the target threshold. This step, based on the principle of closed-loop feedback control, enables continuous optimization of the recycled aggregate production process, improving product quality stability and yield rate.

[0131] Through the synergistic effect of the above seven steps, the entire process of efficiently converting construction waste into high-quality recycled aggregate is achieved, resulting in qualified recycled aggregate with a water absorption rate of less than 3%, a crushing value of less than 18%, a mud content of less than 1%, and comprehensive performance indicators that meet standard requirements. The technical advantages of this embodiment are described below.

[0132] First, intelligent surface feature recognition and adaptive parameter control technology achieves a shift from subjective judgment to objective data analysis. By using a deep convolutional neural network to analyze aggregate surface characteristics in real time, the system automatically outputs optimal processing parameters, ensuring the ideal surface treatment process. This overcomes the inaccurate parameter settings associated with traditional processes that rely on manual experience, effectively addresses the complexity and diversity of construction waste from different sources, and improves product quality consistency.

[0133] Secondly, a multi-stage surface modification synergistic mechanism establishes a comprehensive surface enhancement system, from optimizing physical structure to improving chemical properties. Mechanical friction and cement powder work synergistically to improve the physical structure, sodium silicate impregnation forms a gel network to fill pores, and a hydrophobic silicone coating provides molecular-level waterproofing. This progressive layer-by-layer treatment approach overcomes the limitations of traditional single-treatment technologies while simultaneously addressing the multiple defects of recycled aggregates, such as loose surface, high porosity, and strong water absorption, resulting in a synergistic surface enhancement effect.

[0134] Third, a closed-loop parameter optimization and iteration system integrates aggregate performance testing and process parameter adjustments into an automated feedback mechanism. This system integrates multi-dimensional performance data through a comprehensive performance balance point determination function, generates parameter adjustment recommendations, and feeds them back to the intelligent models, achieving multi-parameter collaborative optimization. This closed-loop optimization mechanism overcomes the isolated and empirical nature of traditional processes, enabling continuous self-adjustment and optimization throughout the entire preparation process.

[0135] These three technical approaches support and collaborate closely with each other, breaking through the bottleneck of traditional recycled aggregate preparation technology and achieving a technological leap from empirical, extensive production to intelligent, precise production. They ensure that all process parameters and material properties are optimally matched during the recycled aggregate preparation process, significantly improving the quality and stability of the final product and laying a solid foundation for the large-scale application of recycled aggregate in high-performance concrete.

[0136] In order to better understand and implement the present invention, the following provides Example 2 of a specific application scenario of the present invention: In a construction waste resource research project, researchers prepared recycled aggregate from concrete waste generated by the demolition of a 30-year-old office building. First, the collected construction waste was sorted and pre-processed using a vibration screening device model ZS-1200, with a vibration frequency of 17.5 Hz and an amplitude of A x =5.2mm, A y =5.0mm, A z =4.8mm, with phase differences of φ1 = π / 2 and φ2 = π / 4. After screening, three particle sizes of construction waste were obtained: over 50mm, 10-50mm, and under 10mm, accounting for 15%, 65%, and 20% of the total, respectively. Subsequently, a 0.65 Tesla magnetic separation device combined with manual sorting removed impurities such as wood, plastic, and metal, bringing the construction waste purity to 97.3%.

[0137] Construction waste with a particle size of 10-50mm was selected as the raw material for recycled aggregate preparation and conveyed to the SL-800 twin-shaft crusher for primary crushing. The speed was set at 75 rpm, and the gear shaft spacing was adjusted to 28mm. The crushed material was discharged through a 15mm discharge port and screened through a vibrating screening system with a 25mm screen. Approximately 75% of the material was qualified, while approximately 25% was oversized, forming a closed-loop crushing system with a circulating load rate of 33%. The particle size of the primary crushed product mainly ranged from 12 to 22mm, with an average particle size of 17mm.

[0138] After primary crushing, the material enters the impact crusher for secondary crushing. The rotor linear speed is set at 42 m / s, the gap between the material and the impact plate is 18 mm, and the feed rate is controlled at 2.5 tons / hour. The particle size of the recycled aggregate after secondary crushing mainly ranges from 6 to 14 mm, with an average particle size of approximately 10 mm. It has distinct angular characteristics and a particle size reduction ratio of 1.7.

[0139] After secondary crushing, the recycled aggregate is conveyed to a surface treatment machine for surface friction treatment. An intelligent recognition model for aggregate surface characteristics, through real-time image acquisition and analysis, determines the optimal surface treatment machine speed to be 138 rpm and the optimal operating time to be 27 minutes. During the treatment process, 42.5-grade ordinary Portland cement powder is added at a mass ratio of 3.5% of the recycled aggregate mass. During the friction treatment, the relative sliding speed of the material is approximately 6.9 m / s. The frictional heat generated raises the temperature inside the treatment drum to approximately 42°C, promoting partial hydration of the cement powder and achieving an adhesion rate of 81%. After surface treatment, the surface angularity of the recycled aggregate is significantly reduced, and the surface roughness is reduced by approximately 35%.

[0140] The recycled aggregate after surface treatment is modified by sodium silicate impregnation. The material ratio optimization neural network sets the number of memory units to 5×(14-6)+100×0.086=48.6≈49 based on the initial porosity parameter of the recycled aggregate of 0.086 and the particle size distribution range of 6 to 14 mm, and calculates and outputs the optimal sodium silicate concentration of 6.8%. The recycled aggregate is immersed in a 6.8% concentration sodium silicate solution and immersed at 65°C for 2.5 hours. During the impregnation process, the pore filling uniformity evaluation model monitors the impregnation effect parameters and the real-time jump connection weight parameter W s =0.25×6.8+0.05×65=4.95. After the impregnation treatment was completed, the impregnation effect parameter (pore filling rate) reached 88.5%. The sodium silicate solution was then discharged and the recycled aggregate was dried in an environment of 108°C for 75 minutes.

[0141] The dried recycled aggregate was transferred to a rotary coating machine at an 18-degree inclination angle for a silicone hydrophobic coating. The rotation speed was set at 90 rpm. A 4.2% methyltriethoxysilane solution was sprayed into the drum via a high-pressure atomizing nozzle at a rate of 20 ml / kg of aggregate for 12 minutes. The coating uniformity reached 93.5%, and the silicone hydrophobic coating coverage on the recycled aggregate surface reached 96.8%. The water contact angle of the coated recycled aggregate surface increased from 45 degrees to 112 degrees, demonstrating excellent hydrophobic properties.

[0142] The performance test of the treated recycled aggregate is shown in Table 1:

[0143] Table 1 Comparison of properties of recycled aggregate before and after modification

[0144] Performance indicators unit Original recycled aggregate Modified recycled aggregate Reference values ​​for natural aggregates Apparent density <![CDATA[kg / m 3 ]]> 2385 2482 2650 Water absorption ％ 7.26 2.15 1.50 Crushing value ％ 23.8 15.6 14.5 Mud content ％ 1.85 0.62 0.50 Porosity ％ 8.6 3.2 2.8

[0145] The test results were input into the comprehensive performance balance point determination function, and the comprehensive performance index parameter reached 0.862, exceeding the target threshold of 0.75, indicating that the modified recycled aggregate has excellent performance and meets the requirements of engineering applications.

[0146] The 28-day compressive strength of C30 concrete prepared with modified recycled aggregate reaches 36.5 MPa, meeting the design requirements. The elastic modulus is 3.12×10 4 MPa, the shrinkage rate is 4.82×10 -4 , the antifreeze performance reaches F100 level, and the resistance to chloride ion penetration is excellent.

[0147] This example shows that the method for preparing construction waste-based recycled aggregate of the present invention effectively improves the physical and mechanical properties and durability of the recycled aggregate. Compared with traditional recycled aggregate processing methods, the method of the present invention has obvious advantages. Traditional recycled aggregate preparation mainly relies on mechanical crushing and simple screening, lacking targeted surface treatment and modification processes, resulting in high water absorption rate of recycled aggregate (usually between 6% and 9%), low strength (crush value is usually between 22% and 28%), loose surface, high porosity and other problems, which seriously limit the application of recycled aggregate in high-performance concrete. However, the present invention significantly reduces the water absorption rate of recycled aggregate (by 10.4%), improves strength (crush value reduced by 14.5%), and improves surface quality (mud content reduced by 16.5%) through a systematic combination of intelligent surface friction treatment, sodium silicate impregnation modification and organosilicon hydrophobic coating. In particular, the artificial intelligence-assisted parameter optimization method adopted by the present invention achieves precise control and dynamic adjustment of processing parameters, significantly improving the quality stability and yield of recycled aggregate, and providing technical support for the large-scale application of recycled aggregate in high-grade concrete.

[0148] It should be noted that the variables involved in the present invention are explained in detail as shown in Table 2 below.

[0149] Table 2 Variable Explanation Table (Part 1)

[0150]

[0151] Table 3 Variable Explanation Table (Part 2)

[0152]

[0153]

[0154] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be covered by the scope of protection of the present invention.

Claims

1. A method for preparing construction waste-based recycled aggregate, characterized in that: It includes construction waste sorting and pretreatment, primary crushing of construction waste, secondary crushing and molding of materials, intelligent surface friction treatment, intelligent sodium silicate impregnation modification, silicone hydrophobic coating treatment, and performance evaluation and parameter optimization iteration. Through the intelligent recognition model of aggregate surface characteristics, the material ratio optimization neural network, the pore filling uniformity evaluation model and the comprehensive performance balance point determination function, a closed-loop parameter optimization iterative mechanism is formed to achieve precise control of the water absorption rate, crushing value and mud content of recycled aggregate.

2. The method for preparing construction waste-based recycled aggregate according to claim 1, characterized in that: The steps of sorting and pre-treating construction waste are specifically to sort and pre-treat construction waste, using vibration screening equipment to separate construction waste into three particle sizes of more than 50 mm, 10 to 50 mm, and less than 10 mm, and remove wood, plastic, and metal impurities through manual sorting combined with magnetic separation equipment to ensure that the purity of the raw materials reaches more than 95%.

3. The method for preparing construction waste-based recycled aggregate according to claim 2, characterized in that: The step of primary crushing of construction waste is specifically to transport the sorted 10 to 50 mm particle size construction waste to a double-shaft crusher for primary crushing, control the speed of the double-shaft crusher to 60 to 80 revolutions per minute, control the output particle size to 10 to 25 mm, and at the same time set up a circulating screening system to return oversized materials to the double-shaft crusher for further crushing.

4. The method for preparing construction waste-based recycled aggregate according to claim 3, characterized in that: The step of secondary crushing and forming the material is to feed the primary crushed material into an impact crusher for secondary crushing, set the impact crusher rotor linear speed to 35 to 45 meters per second, adjust the gap between the material and the material impact plate to 15 to 20 mm, reduce the material particle size to 5 to 15 mm, and form angular recycled aggregate.

5. The method for preparing construction waste-based recycled aggregate according to claim 4, characterized in that: The intelligent surface friction treatment step specifically comprises the following steps: using a mechanical friction method to perform surface treatment on the recycled aggregate after secondary crushing, using an intelligent recognition model of aggregate surface features to perform real-time analysis on the surface condition of the recycled aggregate, controlling the speed of the surface treatment machine to be 120 to 150 revolutions per minute, controlling the operating time of the surface treatment machine to be 25 to 30 minutes, and adding cement powder with a mass ratio of 2% to 5% of the mass of the recycled aggregate.

6. The method for preparing construction waste-based recycled aggregate according to claim 5, characterized in that: The intelligent sodium silicate impregnation modification step specifically comprises performing surface modification on the surface-treated recycled aggregate, using a material ratio optimization neural network to determine the optimal sodium silicate concentration parameter, immersing the recycled aggregate in a sodium silicate solution with a mass concentration of 5% to 8%, and impregnating the solution at 60 to 70° C. for 2 to 3 hours. The impregnation effect parameters of the impregnation treatment are monitored using a pore filling uniformity evaluation model.

7. The method for preparing construction waste-based recycled aggregate according to claim 6, characterized in that: The organosilicon hydrophobic coating treatment step specifically comprises transferring the dried recycled aggregate to a rotary coater, spraying an organosilicon hydrophobic agent with a mass concentration of 3% to 5%, controlling the speed of the rotary coater to 80 to 100 revolutions per minute, and controlling the coating time of the rotary coater to 10 to 15 minutes, so that the coverage rate of the organosilicon hydrophobic agent on the surface of the recycled aggregate reaches more than 95%.

8. The method for preparing construction waste-based recycled aggregate according to claim 7, characterized in that: The steps of performance evaluation and parameter optimization iteration are specifically to perform performance testing on the surface-modified recycled aggregate, obtain the apparent density parameters, water absorption parameters, crushing value parameters and mud content parameters of the recycled aggregate, and input the apparent density parameters, water absorption parameters, crushing value parameters, mud content parameters and impregnation effect parameters into the comprehensive performance balance point determination function.

9. The method for preparing construction waste-based recycled aggregate according to claim 8, characterized in that: In the performance evaluation and parameter optimization iteration steps, the comprehensive performance index parameters of the recycled aggregate and the recommended parameters for adjustment of the processing parameters are calculated through the comprehensive performance balance point determination function, and qualified recycled aggregates with a water absorption parameter lower than 3%, a crushing value parameter less than 18%, a mud content parameter less than 1% and comprehensive performance index parameters meeting the standard requirements are screened out.

10. The method for preparing construction waste-based recycled aggregate according to claim 9, characterized in that: The mechanical friction method specifically refers to the use of equipment to generate continuous friction between recycled aggregates and between recycled aggregates and the inner wall of the equipment, thereby removing loose particles on the surface of the recycled aggregates and smoothing sharp edges and corners. At the same time, the added cement powder is partially hydrated under the action of frictional heat and fills the micropores on the surface of the recycled aggregates, forming a denser and stronger surface structure.

Images