Strengthening treatment and automatic sorting system for construction waste recycled aggregate

Abstract

The invention relates to a strengthening treatment and automatic separation system for construction waste recycled aggregate, and innovatively provides a technical route of'multistage crushing, magnetic separation, winnowing, flotation linkage and aggregate strengthening 'aiming at the core problems of more impurities and low strength of the recycled aggregate. The system comprises a pretreatment module (for removing large impurities), a multi-stage crushing module (for controlling grain shapes), a linkage sorting module (for removing plastics through winnowing, removing wood chips through flotation and carrying out photoelectric precise sorting), an aggregate strengthening module (for improving strength through microwave activation and nanometer coating) and an intelligent control center. By establishing a magnetic separation efficiency model, a grain shape optimization model and a strength prediction algorithm, the aggregate purity is greater than 98% (XRF detection), and the compressive strength is improved by 40% (ISO 679 verification); the energy consumption in the sorting link is reduced by 30%, and the reinforced aggregate can replace natural aggregate to prepare C40 concrete. The bottleneck of high-value utilization of the construction waste is broken through, and circular economy development is promoted.

Description

Technical Field

[0001] This invention relates to the field of intelligent strengthening and sorting technology of recycled aggregates, and in particular to a strengthening treatment and automatic sorting system for recycled aggregates from construction waste. Background Technology

[0002] As a core raw material for concrete, the large-scale application of recycled aggregates can reduce the mining of natural sand and gravel (saving 1.2 tons of natural resources per ton of recycled aggregate) and reduce carbon emissions (reducing CO2 emissions by 48 kg / ton compared to natural aggregates). However, currently, due to high impurity content (>5%) and large strength fluctuations (coefficient of variation ≥0.15), recycled aggregates can only be used in subgrade layers (concrete below C15), and their application rate in high-grade concrete (above C30) is less than 10%, which seriously restricts the process of resource utilization.

[0003] However, existing methods for processing and sorting recycled aggregates from construction waste have the following problems:

[0004] First, in the crushing process of construction waste, the single-stage coarse crushing (such as jaw crusher) and fine crushing (such as impact crusher) are not optimized in a coordinated manner, which leads to the overload of impact energy during secondary crushing, resulting in microcracks and a crushing value of >20% (the national standard requires Class I aggregate to be ≤12%). Moreover, there is no particle shape control mechanism, and the proportion of needle-shaped and flaky particles is >15%, which reduces the fluidity of concrete.

[0005] Secondly, in the aggregate sorting process, magnetic separation alone can only remove metals (efficiency 80-85%), while plastics and wood chips need to be sorted manually (missed detection rate > 8%). In addition, traditional flotation requires the addition of a large amount of chemicals (such as fatty acids), causing secondary pollution (COD emissions > 200 mg / L).

[0006] Finally, in the aggregate strengthening process, although the acid immersion method (5% hydrochloric acid concentration) improves surface activity, it also poses a risk of chloride ion corrosion (Cl- content > 0.06%). Mechanical grinding and strengthening have high energy consumption (> 30 kWh / ton), and the strength improvement is only 15-20%.

[0007] In summary, those skilled in the art need a construction waste recycled aggregate strengthening treatment and automatic sorting system that can simultaneously solve the problem of "purity-strength" synergistic optimization. Summary of the Invention

[0008] The purpose of this invention is to solve the above problems by designing an enhanced treatment and automatic sorting system for recycled aggregates from construction waste.

[0009] The technical solution of the present invention to achieve the above objectives is a system for enhanced treatment and automatic sorting of recycled aggregates from construction waste, the system comprising:

[0010] The pre-processing module performs preliminary crushing of construction waste and removes large-sized impurities from the construction waste based on a metal removal model, thereby obtaining pre-processed coarse material.

[0011] The multi-stage crushing module performs secondary crushing on the coarse material and screens the coarse material after secondary crushing based on the particle shape optimization model to obtain fine material;

[0012] The linkage sorting module sequentially performs air separation, flotation and photoelectric separation operations on the screened concentrate based on the air separation mathematical model, flotation mathematical model and photoelectric separation mathematical model to deeply separate non-metallic impurities in the concentrate to obtain aggregate.

[0013] The aggregate strengthening module strengthens the aggregate according to the strength enhancement model (microwave irradiation chamber → nano-spraying system → hot air sintering furnace) to improve the mechanical properties and durability of the aggregate, thereby obtaining the strengthened aggregate.

[0014] The intelligent control center has built-in digital twin models of the preprocessing module, multi-stage crushing module, linkage sorting module, and aggregate strengthening module. The digital twin model can map the status of the entire system (the system composed of the preprocessing module, multi-stage crushing module, linkage sorting module, and aggregate strengthening module) in real time, and dynamically coordinate the action parameters of the preprocessing module, multi-stage crushing module, linkage sorting module, and aggregate strengthening module based on the reinforcement learning decision model to achieve multi-objective global optimization.

[0015] The pretreatment module includes: a heavy-duty chain feeder, an impact crusher, and an electromagnetic separator. Construction waste is fed into the impact crusher at a constant speed through the heavy-duty chain feeder, crushing the construction waste into particles with a size ≤100mm and conveying it to the electromagnetic separator. The electromagnetic separator removes large impurities according to the metal removal model and outputs the pretreated coarse material.

[0016] The mathematical expression for the metal removal model is:

[0017] η m =1-e -k·B·d

[0018] In the formula, η m denoted as metal removal rate (target ≥95%), k is the magnetic permeability of the material (0.8-1.2), B is the magnetic field strength (Tesla), and d is the length of the magnetic separation path (meters).

[0019] The multi-stage crushing module includes a jaw crusher, a vertical impact crusher, and a high-frequency vibrating screen. The jaw crusher can receive pre-treated coarse material and coarsely crush the particles in the coarse material to a size of 50mm. Then, the coarsely crushed coarse material is fed into the vertical impact crusher for fine crushing, and the particle size of the coarsely crushed coarse material is finely crushed to a size of 5-25mm. The finely crushed coarse material is fed into the high-frequency vibrating screen, and the high-frequency vibrating screen classifies and screens the finely crushed coarse material according to the particle shape optimization model to obtain refined material.

[0020] The mathematical expression for the particle shape optimization model is:

[0021]

[0022] In the formula, S q Let A be the aggregate sphericity, A be the aggregate projected area, and P be the aggregate projected perimeter.

[0023] The linkage sorting module includes a cyclone separator, a microbubble generator, and a hyperspectral camera. The cyclone separator receives the concentrate and adjusts the action parameters of the air separation process according to the air separation mathematical model to complete the air separation operation of the concentrate. The air-separated concentrate is then transported to the microbubble generator. The microbubble generator adjusts the flotation conditions according to the flotation mathematical model to perform flotation on the air-separated concentrate to obtain the flotated concentrate. The hyperspectral camera performs photoelectric separation on the flotated concentrate and adjusts the action parameters of the photoelectric separation process according to the photoelectric separation model to finally remove non-metallic impurities and obtain aggregate.

[0024] The expression for the wind-selection mathematical model is:

[0025]

[0026] In the formula, η w For air separation efficiency, the target value is >95%, ρ p The density of impurity particles ranges from 0.02 to 1.4 g / cm³. 3 , denoted as the equivalent diameter of the impurity particles, ranging from 2 to 10; v is the airflow velocity, ranging from 8 to 12; and μ is the aerodynamic viscosity, ranging from 1.8 × 10⁻⁶. -5 (Under normal temperature conditions), D is the characteristic diameter of the cyclone separator, with a value range of 1.2-2.5.

[0027] The expression for the flotation mathematical model is:

[0028]

[0029] In the formula, η f The flotation removal rate is defined as a target value >90%, K is the surface adsorption constant, and its value ranges from 0.3 to 0.6. C sThe actual concentration of the foaming agent is 0.05-0.1, C0 is the critical micelle concentration with a target value of 0.0083 mol / L (sodium dodecyl sulfate), and t is the effective flotation time with a value range of 60-120.

[0030] The expression for the photoelectric sorting model is:

[0031] y = softmax(W·ReLU(Conv(X)))

[0032] In the formula, y is the impurity category probability vector with a dimension of [3×1], and the output is: aggregate / organic matter / metal. W is the weight matrix of the fully connected layer with a dimension of [256×3], used for training optimization. ReLU is the activation function, expressed as f(x)=max(0,x). The activation function is a non-linear transformation, mainly used to solve the gradient vanishing problem. Conv is the convolution kernel operation, using 3×3@16 channels, mainly used to extract hyperspectral features. X is the input hyperspectral data cube with a dimension of [3×1].

[0033] [256×256×128] is used to input the spatial resolution and 128 bands.

[0034] The aggregate strengthening module includes a nano-spraying system, a microwave irradiation chamber, and a hot air sintering furnace. The nano-spraying system is used to spray sol onto the heated aggregate surface to achieve coating reinforcement. Then, the aggregate coated with sol is introduced into the microwave irradiation chamber and heated to promote the healing of microcracks in the aggregate. Finally, the aggregate with healed microcracks is introduced into the hot air sintering path for sintering, thereby completing the solidification process.

[0035] The mathematical expression for the strength enhancement model is:

[0036]

[0037] In the formula, Δσ is the increase in compressive strength, α is the interfacial bonding coefficient, and its value ranges from 0.8 to 1.2. m V represents the elastic modulus of the nanofilm. f The coating volume fraction has a target value of ≥12%.

[0038] The nano-spraying system uses SiO2 sol with a particle size of 50±5nm and a film thickness of 10-15μm.

[0039] The mathematical expression for the reinforcement learning decision model is:

[0040] Q(s,a)=R(s,a)+γmax a′ Q(s′,a′)

[0041] In the formula, Q is the action value function, R is the reward function, s is the system state (temperature, particle size, etc.), a is the control action (crushing pressure, wind speed, etc.), s′ is the next system state, a′ is the next controllable action, and γ is the discount factor.

[0042] The mathematical expression for the reward function R is:

[0043] R = w1·(purity - 0.98) + w2·Δσ - w3·energy consumption

[0044] In the formula, w1, w2, and w3 are all weighting coefficients, with w1 = 10, w2 = 0.5, and w3 = 0.01 (purity priority).

[0045] The main inventive points of this invention are as follows:

[0046] First, the present invention adopts a sorting-enhancement synergistic mechanism, utilizing a series design of air classification and flotation, where air classification removes light impurities and flotation separates heavy organic matter. At the same time, microwave activation is carried out before nano-coating to repair internal defects and enhance the surface.

[0047] Secondly, this invention combines impurity identification with dynamic control, uses a hyperspectral camera to analyze aggregate composition in real time, and improves the identification accuracy to 98.7% based on a metal removal model and a particle shape optimization model.

[0048] Finally, this invention employs a combination of strength prediction and proactive intervention to establish a model relating aggregate strength to crushing energy consumption, namely... When the predicted strength deviation is greater than 5%, the vertical crusher speed is automatically adjusted.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] 1. This invention uses a combination of magnetic separation, air separation, flotation and photoelectric separation to achieve targeted removal of impurities, with a cumulative error rate of <2% and aggregate purity >98%;

[0051] 2. This invention employs a triple treatment method of microwave activation, nano-coating, and low-temperature sintering to achieve microcrack repair and surface densification of aggregates, thereby increasing the compressive strength of aggregates by 40%.

[0052] 3. This invention utilizes an intelligent control center to achieve dynamic adjustment of crushing power and reuse of sintering waste heat, reducing the energy consumption of the entire system by 30%. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of an enhanced treatment and automatic sorting system for recycled aggregates from construction waste, as described in this invention.

[0054] Figure 2This is a flowchart of the aggregate strengthening module described in this invention;

[0055] Figure 3 This is a data comparison table of Embodiment 1 of the present invention;

[0056] Figure 4 This is a data comparison table of Embodiment 2 of the present invention;

[0057] Figure 5 This is a data comparison table of Embodiment 3 of the present invention. Detailed Implementation

[0058] The present invention will now be described in detail with reference to the accompanying drawings;

[0059] An enhanced treatment and automatic sorting system for recycled aggregates from construction waste, such as... Figure 1-2 As shown, the system includes:

[0060] The pre-processing module performs preliminary crushing of construction waste and removes large-sized impurities from the construction waste based on a metal removal model, thereby obtaining pre-processed coarse material.

[0061] The multi-stage crushing module performs secondary crushing on the coarse material and screens the coarse material after secondary crushing based on the particle shape optimization model to obtain fine material;

[0062] The linkage sorting module sequentially performs air separation, flotation and photoelectric separation operations on the screened concentrate based on the air separation mathematical model, flotation mathematical model and photoelectric separation mathematical model to deeply separate non-metallic impurities in the concentrate to obtain aggregate.

[0063] The aggregate strengthening module strengthens the aggregate according to the strength enhancement model (microwave irradiation chamber → nano-spraying system → hot air sintering furnace) to improve the mechanical properties and durability of the aggregate, thereby obtaining the strengthened aggregate.

[0064] The intelligent control center has built-in digital twin models of the preprocessing module, multi-stage crushing module, linkage sorting module, and aggregate strengthening module. The digital twin model can map the status of the entire system (the system composed of the preprocessing module, multi-stage crushing module, linkage sorting module, and aggregate strengthening module) in real time, and dynamically coordinate the action parameters of the preprocessing module, multi-stage crushing module, linkage sorting module, and aggregate strengthening module based on the reinforcement learning decision model to achieve multi-objective global optimization.

[0065] The pretreatment module includes: a heavy-duty chain feeder (uniform feeding), an impact crusher, and an electromagnetic separator (magnetic field strength ≥12,000 Gauss). Construction waste is fed into the impact crusher at a uniform speed through the heavy-duty chain feeder, crushing the construction waste into particles ≤100mm and conveying it to the electromagnetic separator. The electromagnetic separator removes large impurities according to the metal removal model and outputs pretreated coarse material.

[0066] The mathematical expression for the metal removal model is:

[0067] η m =1-e -k·B·d

[0068] In the formula, η m denoted as metal removal rate (target ≥95%), k is the magnetic permeability of the material (0.8-1.2), B is the magnetic field strength (Tesla), and d is the length of the magnetic separation path (meters).

[0069] The multi-stage crushing module includes a jaw crusher, a vertical impact crusher, and a high-frequency vibrating screen (grading and screening). The jaw crusher can receive pre-treated coarse material and coarsely crush the particles in the coarse material to a size of 50mm. Then, the coarsely crushed coarse material is fed into the vertical impact crusher for fine crushing, and the particle size of the coarsely crushed coarse material is finely crushed to a size of 5-25mm. The finely crushed coarse material is fed into the high-frequency vibrating screen, and the high-frequency vibrating screen performs grading and screening of the finely crushed coarse material according to the particle shape optimization model to obtain refined material.

[0070] The mathematical expression for the particle shape optimization model is:

[0071]

[0072] In the formula, S q A represents the aggregate sphericity (target ≥ 0.7), and A represents the aggregate projected area (mm²). 2 ), where P is the projected perimeter of the aggregate (mm).

[0073] The linkage sorting module includes a cyclone separator, a microbubble generator, and a hyperspectral camera. The cyclone separator receives the concentrate and adjusts the action parameters of the air separation process (including but not limited to the air velocity, for example, a preferred air velocity of 10 m / s) according to the air separation mathematical model to complete the air separation operation of the concentrate and then transports the air-separated concentrate to the microbubble generator. The microbubble generator adjusts the flotation conditions (including but not limited to the pH value, for example, pH = 9.5) according to the flotation mathematical model to perform flotation on the air-separated concentrate to obtain the flotated concentrate. The hyperspectral camera performs photoelectric separation on the flotated concentrate and adjusts the action parameters of the photoelectric separation process (including but not limited to the wavelength of the hyperspectral camera, for example, the wavelength of the hyperspectral camera is 900-1700 nm) according to the photoelectric separation model to finally remove non-metallic impurities and obtain aggregate.

[0074] The expression for the wind-selection mathematical model is:

[0075]

[0076] It is mainly used to quantitatively calculate the separation efficiency of light impurities (such as plastics and paper), guide the coordinated design of wind speed (v) and separation cylinder diameter (D), and ensure the separation of suspended impurities;

[0077] In the formula, η w For air separation efficiency, the target value is >95%, ρ p Impurity particle density (g / cm³) 3 The value range is 0.02-1.4 g / cm³. 3 , denoted as equivalent diameter of impurity particles (mm), ranging from 2 to 10; v is the airflow velocity (m / s), ranging from 8 to 12; μ is the aerodynamic viscosity (Pa·s), ranging from 1.8 × 10⁻⁶. -5 (Under normal temperature conditions), D is the characteristic diameter (m) of the cyclone separator, with a value range of 1.2-2.5;

[0078] For example: when ρ p Reduce (e.g., ρ of foam plastics) p =0.2), v needs to be increased to 12m / s to maintain η. w >95%;

[0079] The expression for the flotation mathematical model is:

[0080]

[0081] It is mainly used to describe the adsorption and removal rate of organic impurities (wood chips, rubber), optimize the foaming agent concentration (Cs) and flotation time (t), and reduce the amount of chemicals used;

[0082] In the formula, ηf The flotation removal rate is defined as a target value >90%, K is the surface adsorption constant, and its value ranges from 0.3 to 0.6 (experimentally calibrated). C s C0 is the actual concentration of the foaming agent (mol / L), with a value range of 0.05-0.1; C0 is the critical micelle concentration (CMC) (mol / L), with a target value of 0.0083 mol / L; and t is the effective flotation time (s), with a value range of 60-120.

[0083] For example: If C s / C0>5 (i.e., excessive addition), η can still be guaranteed even if t=60s. f >90%, avoiding waste;

[0084] The expression for the photoelectric sorting model is:

[0085] y = softmax(W·ReLU(Conv(X)))

[0086] Its main function is to achieve pixel-level identification of impurities based on convolutional neural networks (CNN), output the impurity category probability (y), and drive the air jet valve to precisely remove impurities.

[0087] In the formula, y is the impurity category probability vector with a dimension of [3×1], and the output is: aggregate / organic matter / metal. W is the weight matrix of the fully connected layer with a dimension of [256×3], used for training optimization. ReLU is the activation function, expressed as f(x)=max(0,x). The activation function is a non-linear transformation, mainly used to solve the gradient vanishing problem. Conv is the convolution kernel operation, using 3×3@16 channels, mainly used to extract hyperspectral features. X is the input hyperspectral data cube with a dimension of [3×1].

[0088] [256×256×128] is used to input the spatial resolution and 128 bands;

[0089] For example, if the input X to the photoelectric sorting model is a hyperspectral image of 900-1700nm, then the output y will be the result of triggering air jet when the probability of "organic matter" is >0.98.

[0090] The entire linkage sorting module operates as follows: If residual wood chips (d) remain after air separation p =3mm,ρ p =0.6), the wind-separation model calculates η w =87% → 13% remains and enters flotation. The flotation model uses Cs=0.08, t=90s to increase η. f Up to 93% → final residue <1%, photoelectric model captures residual impurities → air jet positioning accuracy ±1mm.

[0091] The aggregate strengthening module includes a microwave irradiation chamber (frequency 2.45 GHz, power 30 kW), a nano-spraying system (SiO2 sol, particle size 50 nm), and a hot air sintering furnace (250 ± 5 °C). The nano-spraying system is used to spray sol onto the heated aggregate surface to achieve a coating reinforcement. Then, the aggregate coated with sol is introduced into the microwave irradiation chamber and heated to promote the healing of microcracks within the aggregate. Finally, the aggregate with healed microcracks is introduced into the hot air sintering path for sintering, thereby completing the solidification process.

[0092] It should be noted that the microwaves (2.45GHz) emitted by the microwave irradiation chamber are absorbed by polar molecules (such as free water and CSH gel) in the aggregate → the molecules in the aggregate oscillate at high frequency → the local surface temperature of the aggregate rises to 300-400℃ → activating the surface energy of the aggregate; the high temperature causes the unhydrated cement particles inside the aggregate to undergo secondary hydration, generating new CSH gel: 2CaO·SiO2+2H2O→Ca2SiO4·H2O(CSH), which fills microcracks (width <0.1mm) and increases the bulk strength by 15-20%;

[0093] The sol is a 50nm particle size SiO2 sol (20% solid content). After being sprayed through a nano-spraying system, it forms a three-dimensional network film on the aggregate surface, which penetrates into the micropores (pore size 0.5-5μm) on the aggregate surface, thereby reducing the porosity of the aggregate surface. SiO2 reacts with Ca(OH)2 on the aggregate surface to form calcium silicate gel: SiO2 + Ca(OH)2 → CaSiO3·H2O, which enhances the interfacial bonding force (bond energy is increased to 150kJ / mol, which is 10 times higher than physical adsorption).

[0094] The hot air sintering furnace uses 250℃ hot air to promote the condensation of Si-OH groups in silica sol: 2Si-OH→Si-O-Si+H2O↑ to form a dense Si-O-Si network, and the hardness of the nanofilm is increased to 6.5GPa (close to quartz). The sintering process of the hot air sintering furnace is as follows: gradient heating (50℃ / min) → holding (10min) → slow cooling (5℃ / min) → eliminating internal stress of the nanofilm and avoiding cracking.

[0095] The mathematical expression for the strength enhancement model is:

[0096]

[0097] In the formula, Δσ is the increase in compressive strength (MPa), α is the interfacial bonding coefficient, and its value ranges from 0.8 to 1.2. m For the elastic modulus of the nanofilm (GPa), V f The target value for the coating volume fraction is ≥12%.

[0098] It is important to note that the intensity enhancement model introduces an interface bonding coefficient α (calibrated via SEM microstructure) to correlate the microscopic interface state with the macroscopic intensity, thus avoiding "over-enhancement" (e.g., V). f >15% leads to aggregate agglomeration, saving ¥120 / ton in nanomaterial costs.

[0099] The nano-spraying system uses SiO2 sol with a particle size of 50±5nm and a film thickness of 10-15μm.

[0100] The mathematical expression for the reinforcement learning decision model is:

[0101] Q(s,a)=R(s,a)+γmax a′ Q(s′,a′)

[0102] In the formula, Q is the action value function, R is the reward function, s is the system state (temperature, particle size, etc.), a is the control action (crushing pressure, wind speed, etc.), s′ is the next system state, a′ is the next controllable action, and γ is the discount factor.

[0103] It is important to note that the reinforcement learning decision model uses a discount factor γ = 0.95, which makes the system focus more on long-term stability rather than short-term gains, reduces the frequency of human intervention (<1 time / shift), and decreases the failure rate by 40%.

[0104] The mathematical expression for the reward function R is:

[0105] R = w1·(purity - 0.98) + w2·Δσ - w3·energy consumption

[0106] In the formula, w1, w2, and w3 are all weighting coefficients, with w1 = 10, w2 = 0.5, and w3 = 0.01 (purity priority).

[0107] Example 1: Small and medium-sized construction waste recycling center (annual processing capacity of 200,000 tons)

[0108] Application environment:

[0109] Raw materials: Demolition waste (60% concrete blocks, 30% bricks and tiles, 10% impurities)

[0110] Site: Fixed production line with stable power supply

[0111] Objective: To produce recycled aggregate for C30 concrete.

[0112] The operation process is as follows:

[0113] Pre-processing module: Construction waste enters the heavy-duty impact crusher (power 132kW). Based on the metal removal model, the magnetic field strength is adjusted (B=1.2T) to remove metals such as reinforcing bars (efficiency 96.2%). Output coarse material: ≤100mm, metal residue <0.8%.

[0114] Multi-stage crushing module: coarse material passes through a jaw crusher (feed 50mm) + a vertical impact crusher (output 20mm). Hydraulic pressure is adjusted based on a particle shape optimization model to output fine material with a sphericity S. q =0.73 (National standard ≥ 0.65)

[0115] Linked sorting module:

[0116] Air separation: Set the air velocity to 10.5 m / s to remove plastic (efficiency 94%).

[0117] Flotation: Injection of foaming agent (satisfying C) s =0.08mol / L), separating sawdust (efficiency 91%)

[0118] Photoelectric selection: CNN model identifies residual impurities, air jet positioning removes them (accuracy ±1mm), output aggregate: purity 98.7%.

[0119] Aggregate reinforcement module: Microwave irradiation (45 seconds) repairs microcracks → Nano-spraying (V f =11%) → Hot air sintering (250℃), strength enhancement model calculation Δσ=14.8MPa (theoretical value 15.2MPa)

[0120] Intelligent Control Center: Digital twin displays aggregate porosity in real time (18% → 12%), reinforcement learning decision-making reduces flotation time by 20 seconds (saving 15% of reagents);

[0121] The data comparison table for Example 1 is as follows: Figure 3 As shown.

[0122] Example 2: Mobile Emergency Treatment Station (Post-Disaster Construction Waste Disposal)

[0123] Application environment:

[0124] Raw materials: Post-earthquake rubble (containing more than 15% impurities such as glass, wood, and plastic)

[0125] Site: Vehicle-mounted equipment, powered by diesel generator

[0126] Objective: To rapidly produce roadbed fill aggregate

[0127] The operation process is as follows:

[0128] The intelligent control center activates emergency mode, and the reinforcement learning decision model Q(s,a) automatically simplifies the process (turns off flotation and strengthens air separation);

[0129] The pretreatment module dynamically increases the magnetic field strength to 1.5T (k=1.1, d=1.8m) using the magnetic separation model, achieving a metal removal rate of 99.1%.

[0130] Multi-stage crushing module, relaxed requirements for particle shape optimization model (S) q ≥0.6), increase the processing speed to 50t / h;

[0131] Linked sorting module:

[0132] Air separation, with the air velocity adjusted to 12m / s, achieves a plastic removal rate of 97%.

[0133] Photoelectric selection, CNN model focusing on glass recognition (wavelength 1550nm), with a clearance rate of 99%;

[0134] Aggregate strengthening module: Strength increased by 12% when microwave irradiation is used only (30 seconds);

[0135] The data comparison table for Example 2 is as follows: Figure 4 As shown.

[0136] Example 3: Zero-carbon building materials industrial park (annual production of 1 million tons of recycled aggregate)

[0137] Application environment:

[0138] Raw materials: Waste from urban renewal projects (complex composition with large fluctuations in impurities)

[0139] Site: Photovoltaic power supply + industrial waste heat utilization

[0140] Objective: To produce C50 high-strength concrete aggregate.

[0141] The operation process is as follows:

[0142] Intelligent Control Center: Digital twin model predicts fluctuations in raw material impurities (a sudden increase in sawdust content to 12%), and reinforcement learning is used to improve flotation C. s Up to 0.12 mol / L, increase nano-spraying V f =13% (compensation intensity)

[0143] Linked sorting module:

[0144] Flotation, according to η f The model automatically adds chemicals, achieving a sawdust removal rate of 93.5%.

[0145] Photoelectric selection, CNN model identifies tiny rubber particles (0.5mm);

[0146] Aggregate strengthening module: It adopts a complete process of microwave-nano-sintering, and Δσ is calibrated in real time through a strength enhancement model. The hot air sintering furnace can utilize waste heat (energy saving 35%).

[0147] Multi-objective optimization: The reward function R = 10·({purity} - 0.98) + 0.5·Δσ - 0.01·{energy consumption} achieves dynamic equilibrium.

[0148] The data comparison table for Example 3 is as follows: Figure 5 As shown.

[0149] The above technical solutions only embody the preferred technical solutions of the present invention.

[0150] The changes that technical personnel in the field may make to certain parts all reflect the present invention.

[0151] The principles of the invention are within the scope of protection of this invention.

Claims

1. A system for enhanced treatment and automatic sorting of recycled aggregates from construction waste, characterized in that, The system includes: The pre-processing module performs preliminary crushing of construction waste and removes large-sized impurities from the construction waste based on a metal removal model to obtain coarse material; The multi-stage crushing module performs secondary crushing of coarse materials and screening based on a particle shape optimization model to obtain fine materials; The linkage separation module sequentially performs air separation, flotation and photoelectric separation operations on the concentrate based on the air separation mathematical model, flotation mathematical model and photoelectric separation mathematical model to separate non-metallic impurities in the concentrate to obtain aggregate. The aggregate strengthening module strengthens the aggregate according to the strength enhancement model to obtain strengthened aggregate; The intelligent control center incorporates digital twin models of preprocessing, multi-stage crushing, linkage sorting, and aggregate strengthening modules. Based on a reinforcement learning decision model, it dynamically coordinates the action parameters of these modules to achieve multi-objective global optimization.

2. The enhanced treatment and automatic sorting system for recycled aggregate from construction waste according to claim 1, characterized in that, The pretreatment module includes: a heavy-duty chain feeder, an impact crusher, and an electromagnetic separator. The mathematical expression for the metal removal model is: or m =1-e -k·B·d In the formula, η m denoted as metal removal rate, k as material permeability coefficient, B as magnetic field strength, and d as magnetic separation path length.

3. The enhanced treatment and automatic sorting system for recycled aggregate from construction waste according to claim 1, characterized in that, The multi-stage crushing module includes: a jaw crusher, a vertical impact crusher, and a high-frequency vibrating screen; The mathematical expression for the particle shape optimization model is: In the formula, S q Let A be the aggregate sphericity, A be the aggregate projected area, and P be the aggregate projected perimeter.

4. The enhanced treatment and automatic sorting system for recycled aggregate from construction waste according to claim 1, characterized in that, The linkage sorting module includes: a cyclone separator, a microbubble generator, and a hyperspectral camera.

5. The enhanced treatment and automatic sorting system for recycled aggregate from construction waste according to claim 4, characterized in that, The expression for the wind-separation mathematical model is: In the formula, η w For air separation efficiency, ρ p The density of impurity particles, Where is the equivalent diameter of the impurity particles, v is the airflow velocity, μ is the aerodynamic viscosity, and D is the characteristic diameter of the cyclone separator. The expression for the flotation mathematical model is: In the formula, η f Let K be the flotation removal rate, K be the surface adsorption constant, and C be the flotation removal rate. s C0 is the actual concentration of the foaming agent, C0 is the critical micelle concentration, and t is the effective flotation time. The expression for the photoelectric sorting model is: y = softmax(W·ReLU(Conv(X))) In the formula, y is the impurity category probability vector, W is the weight matrix of the fully connected layer, ReLU is the activation function, Conv is the convolution kernel operation, and X is the input hyperspectral data cube.

6. The enhanced treatment and automatic sorting system for recycled aggregate from construction waste according to claim 1, characterized in that, The aggregate strengthening module includes: a nano-spraying system, a microwave irradiation chamber, and a hot air sintering furnace.

7. The enhanced treatment and automatic sorting system for recycled aggregate from construction waste according to claim 6, characterized in that, The mathematical expression for the strength enhancement model is: In the formula, Δσ is the increase in compressive strength, α is the interfacial bonding coefficient, and E m V represents the elastic modulus of the nanofilm. f This represents the volume fraction of the coating.

8. The enhanced treatment and automatic sorting system for recycled aggregate from construction waste according to claim 7, characterized in that, The nano-spraying system uses SiO2 sol with a particle size of 50±5nm, and the thickness of the coating is 10-15μm.

9. The enhanced treatment and automatic sorting system for recycled aggregate from construction waste according to claim 1, characterized in that, The mathematical expression for the reinforcement learning decision model is: Q(s,a)=R(s,a)+γmax a′ Q(s′,a′) In the formula, Q is the action value function, R is the reward function, s is the system state, a is the control action, s′ is the next system state, a′ is the next controllable action, and γ is the discount factor.

10. The enhanced treatment and automatic sorting system for recycled aggregate from construction waste according to claim 9, characterized in that, The mathematical expression for the reward function R is: R = w1·(purity - 0.98) + w2·Δσ - w3·energy consumption In the formula, w1, w2, and w3 are all weighting coefficients.

Images