An ultra-weather-resistant industrial heavy-duty coating and a preparation method thereof

Abstract

The application provides an ultra-weather-resistant industrial heavy-duty paint and a preparation method thereof, and relates to the technical field of paint materials.The preparation method comprises the following steps: S1, raw material preparation;S2, pretreatment of pigments and fillers;S3, pretreatment of binders;S4, premixing;S5, grinding and dispersing;S6, paint mixing;S7, aging treatment;S8, filtering and packaging.The application solves the problems of poor synergy and blindness of process parameters of traditional paint binders, improves the comprehensive protection performance and production efficiency of the paint, and is suitable for high-demand industrial heavy-duty scenes.

Description

Technical Field

[0001] This invention relates to the field of coating materials technology, and in particular to an ultra-weather-resistant industrial heavy-duty anti-corrosion coating and its preparation method. Background Technology

[0002] Industrial heavy-duty anti-corrosion coatings must possess both excellent weather resistance and corrosion resistance to withstand the erosion of substrates by harsh environments such as ultraviolet radiation, salt spray, and chemical media. In existing technologies, coatings prepared from a single base material are difficult to balance weather resistance and adhesion. For example, pure fluorocarbon coatings have good weather resistance but insufficient adhesion to metal substrates, while pure epoxy resin coatings have strong adhesion but poor resistance to ultraviolet aging.

[0003] In terms of production process, the uniformity of pigment and filler dispersion directly affects coating performance. However, traditional grinding processes rely on manual experience to set rotation speed and time, and parameter adjustments lack scientific basis, resulting in large fluctuations in slurry fineness and a tendency for agglomeration. At the same time, traditional grinding processes suffer from low efficiency and high energy consumption, and frequent manual sampling and testing not only increase labor intensity but also affect production continuity.

[0004] In addition, the order of additive addition and curing conditions of existing coatings lack precise control, which can easily lead to insufficient synergistic effect of additives. During the storage of coatings, they are prone to layering and thickening. After application, the coating is prone to defects such as pinholes and orange peel, which further affect the protective effect and service life. Summary of the Invention

[0005] This invention provides an ultra-weather-resistant industrial heavy-duty anti-corrosion coating and its preparation method, focusing on three core dimensions: optimized raw material selection, refined process control, and intelligent algorithm-enabled grinding. First, a blend of fluorocarbon resin and epoxy resin is selected as the base material, combined with functional pigments and fillers such as nano-titanium dioxide and zinc powder, as well as specialized additives, clearly defining the specifications and proportions of each component. Second, pretreatment steps such as drying and dehydrating pigments and fillers, crushing and sieving, and adjusting the viscosity of the base material provide a high-quality raw material foundation for subsequent dispersion. Subsequently, the pre-mixing stage achieves preliminary homogenization of the pigments and fillers with the base material. The core innovation lies in the integration of a BP neural network prediction algorithm, a particle swarm optimization algorithm, and an adaptive parameter adjustment algorithm in the grinding and dispersion step, optimizing grinding parameters in real time to ensure uniform and stable slurry fineness. Finally, after paint preparation, curing, and precision filtration and packaging, the finished coating is formed.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A method for preparing an ultra-weather-resistant industrial heavy-duty anti-corrosion coating includes the following steps:

[0008] S1: Prepare raw materials, including base material, pigments and fillers, solvents and additives, and weigh each raw material according to the preset mass proportions;

[0009] S2: Pre-treat a symmetrical amount of pigments and fillers to obtain pre-treated pigments and fillers;

[0010] S3: Mix and adjust the viscosity of the symmetrical amount of base material to obtain a mixed base material, and divide it into a first part of the mixed base material and a second part of the mixed base material;

[0011] S4: The first part of the mixed base material is premixed with the pretreated pigments and fillers to obtain a premixed slurry, and the initial fineness of the premixed slurry is tested;

[0012] S5: The premixed slurry is fed into a sand mill, and the grinding process is controlled by a fusion of a BP neural network prediction algorithm, a particle swarm optimization algorithm, and an adaptive parameter adjustment algorithm. The BP neural network prediction algorithm predicts the fineness of the current slurry based on the initial fineness, the current speed of the sand mill, and the cumulative grinding time. The particle swarm optimization algorithm optimizes the grinding parameters within a preset speed and time range based on the predicted fineness. The adaptive parameter adjustment algorithm adjusts the speed and time in real time based on the optimal grinding parameters and the deviation between the detected actual fineness and the target fineness. It also feeds back the actual grinding parameters and the corresponding actual fineness to the BP neural network prediction algorithm to update the model until the slurry fineness meets the standard, thus obtaining the ground slurry.

[0013] S6: Add additives, solvents, and the second part of the mixed base material to the ground slurry, and adjust the viscosity to the preset range to obtain the paint after mixing.

[0014] S7: The paint after mixing is cured and its weather resistance, adhesion and salt spray resistance are tested. After meeting the standards, the cured paint is obtained.

[0015] S8: After the coating has matured, it is filtered and packaged to obtain the finished coating.

[0016] In this specification, in step S1, the base material includes fluorocarbon resin and epoxy resin, the pigments and fillers include nano titanium dioxide, zinc powder, mica powder and talc powder, the solvent includes xylene and butyl acetate, and the additives include dispersant, defoamer, leveling agent and ultraviolet absorber.

[0017] In this specification, step S2, the pretreatment process of pigments and fillers is as follows: the weighed pigments and fillers are vacuum dried until the moisture content is ≤0.5%, and then pulverized and sieved to a particle size of 50-100μm; the vacuum drying conditions are: temperature 80-100℃, vacuum degree -0.08--0.09MPa, drying for 2-3 hours. If the moisture content exceeds the standard, the drying time is extended by 30 minutes and then retested until the moisture content is ≤0.5%.

[0018] In this specification, in step S5, the particle swarm optimization algorithm uses the predicted fineness as a constraint condition and substitutes it into the fitness function to calculate the particle fitness. Within the range of rotation speed of 1500 to 2000 r / min and time of 30 to 180 min, the optimal rotation speed and optimal time are obtained through iterative optimization. The fitness function prioritizes ensuring that the predicted fineness is ≤50 μm and then minimizes the grinding energy consumption.

[0019] In this specification, in step S5, the adjustment amount of the adaptive parameter adjustment algorithm is determined in the following way: based on the deviation between the detected actual fineness and the target fineness of 50μm, and the difference between the optimal speed, optimal time and the current actual speed, and time, the speed adjustment amount and the time adjustment amount are calculated respectively, so that the actual speed and time approach the optimal speed and optimal time while compensating for the fineness deviation.

[0020] In this specification, in step S5, the BP neural network prediction algorithm, particle swarm optimization algorithm, and adaptive parameter adjustment algorithm form a closed-loop synergy: the BP neural network prediction algorithm provides prediction data for the particle swarm optimization algorithm, the particle swarm optimization algorithm provides target parameters for the adaptive parameter adjustment algorithm, and the adaptive parameter adjustment algorithm provides updated samples for the BP neural network prediction algorithm, until the closed loop ends when the actual detected slurry fineness is ≤50μm.

[0021] In this instruction manual, in step S3, the viscosity of the mixed base material is adjusted as follows: if the viscosity is lower than 500 mPa·s, add 1 to 2 parts of fluorocarbon resin; if the viscosity is higher than 800 mPa·s, add 0.5 to 1 part of xylene. Stir for 10 minutes and then retest until the viscosity meets the standard.

[0022] In this instruction manual, in step S4, the pretreated pigments and fillers are added at a rate of 5-8 kg / min. After the addition is completed, the mixture is stirred at 800-1000 r / min for 1-1.5 hours. If there are obvious visible particles, the stirring is extended for 20 minutes and then tested again until the uniformity meets the standard.

[0023] In this instruction manual, in step S7, the curing conditions are: temperature 25~30℃, stirring at 300r / min for 10 minutes every 6 hours; weather resistance requirement: gloss retention rate ≥80% after 1000h aging; adhesion requirement: grade ≤1; salt spray resistance requirement: no rust or blistering after 5000h.

[0024] An ultra-weather-resistant industrial heavy-duty anti-corrosion coating comprises raw material components formulated according to effective mass parts: base material, pigments and fillers, solvent and additives;

[0025] The base material is a compound system of fluorocarbon resin and epoxy resin. The fluorocarbon resin provides UV aging resistance through the high bond energy CF bond in the molecular structure, while the epoxy resin enhances the coating adhesion through the reaction of epoxy groups with hydroxyl groups on the surface of the metal substrate. The two work together to achieve a balance between the weather resistance and adhesion of the coating.

[0026] The pigments and fillers include nano-titanium dioxide, zinc powder, mica powder, and talc powder. Nano-titanium dioxide (anatase type, particle size 20-50nm) is used to scatter and absorb ultraviolet light and decompose surface contaminants. Zinc powder (purity ≥99.5%, particle size 5-10μm) acts as a sacrificial anode to achieve electrochemical corrosion protection through preferential oxidation. Mica powder (plate structure, particle size 10-20μm) and talc powder (layer structure, particle size 15-30μm) work together to form a "maze effect" to extend the penetration path of corrosive media.

[0027] The solvent is a compound system of xylene and butyl acetate, which is used to dissolve the base material and adjust the evaporation rate of the coating to avoid pinholes or cracks caused by excessive evaporation during coating application.

[0028] The additives include dispersants, defoamers, leveling agents, and ultraviolet absorbers. The dispersants prevent pigment and filler agglomeration through electrostatic repulsion, the defoamers reduce the surface tension of the system to eliminate bubbles, the leveling agents promote uniform spreading of the coating on the substrate surface, and the ultraviolet absorbers (benzotriazoles) specifically absorb 290-400nm ultraviolet light to protect the molecular chains of the base material.

[0029] The core performance indicators of this ultra-weather-resistant industrial heavy-duty anti-corrosion coating meet the following requirements: gloss retention rate ≥80% after 1000h ultraviolet aging test, cross-cut adhesion test grade ≤1, and no rust or blistering after 5000h neutral salt spray test.

[0030] In summary, the present invention has at least the following beneficial effects:

[0031] 1. Significantly improved overall protective performance of the coating: Through the compound design of fluorocarbon resin and epoxy resin, combined with the synergistic effect of nano-level pigments and fillers, the coating has strong adhesion, excellent weather resistance and heavy corrosion resistance, which can effectively resist the erosion of the substrate by harsh environment and extend the service life of the substrate.

[0032] 2. Intelligent and precise control of the grinding process: Three algorithms are integrated to form a closed-loop optimization system, replacing traditional manual experience control, avoiding fluctuations in slurry fineness, ensuring uniform coating dispersion, and reducing coating defects from the source.

[0033] 3. Production efficiency and energy saving optimization: Dynamic adjustment of parameters in the grinding process reduces ineffective energy consumption, shortens the grinding cycle, and reduces the frequency of manual sampling and testing, thereby improving production continuity; refined control of each step reduces raw material loss and improves product qualification rate.

[0034] 4. More stable coating performance: By optimizing the order of additive addition, curing conditions and solvent ratio, the coating is less prone to separation and thickening during storage, has good leveling properties during construction, and produces a smooth and flat coating surface. It is suitable for various construction methods and has a wider range of applications. Attached Figure Description

[0035] Figure 1 This is a schematic flowchart of the preparation method of the ultra-weather-resistant industrial heavy-duty anti-corrosion coating involved in this invention.

[0036] Figure 2 This is a schematic diagram of the process for preparing the raw materials involved in this invention.

[0037] Figure 3 This is a schematic diagram of the pigment and filler pretreatment process involved in this invention.

[0038] Figure 4 This is a schematic diagram of the grinding and dispersion process involved in this invention. Detailed Implementation

[0039] like Figure 1 As shown in the figure, this embodiment provides a method for preparing an ultra-weather-resistant industrial heavy-duty anti-corrosion coating, including:

[0040] S1: Raw Material Preparation

[0041] To ensure the coating possesses superior weather resistance and heavy-duty corrosion protection, the types and specifications of raw materials must be strictly selected. All raw materials are sourced from the market and comply with relevant national standards. The raw material preparation process is as follows: Figure 2 As shown.

[0042] Base material: As the film-forming substance of coatings, it directly affects the adhesion and weather resistance of the coating.

[0043] 30-40 parts of fluorocarbon resin (50% solid content, hydroxyl value 80-100 mg KOH / g): Hydroxyl fluorocarbon resin is selected because its CF bond energy in its molecular structure is high (485 kJ / mol), and it has excellent resistance to ultraviolet aging. The hydroxyl groups can react with subsequent additives to increase the crosslinking density. 10-15 parts of epoxy resin (epoxy value 0.5-0.6 eq / 100g): Low molecular weight bisphenol A type epoxy resin is selected. When compounded with fluorocarbon resin, it can enhance the adhesion of the coating to the metal substrate (epoxy groups easily react with hydroxyl groups on the metal surface) and improve the chemical corrosion resistance.

[0044] Pigments and fillers: impart hiding power, corrosion resistance and mechanical strength to coatings. Particle size needs to be controlled to match subsequent grinding processes.

[0045] 5-8 parts of nano-titanium dioxide (particle size 20-50nm, anatase): Nano-sized particles enhance the scattering and absorption of ultraviolet light, improving weather resistance; anatase has higher photocatalytic activity than rutile, and can decompose surface contaminants; 15-20 parts of zinc powder (particle size 5-10μm, purity ≥99.5%): As a sacrificial anode, it protects the substrate from corrosion through electrochemical action; the particle size must be uniform to avoid agglomeration; 8-12 parts of mica powder (particle size 10-20μm, lamellar structure): The lamellar structure can form a labyrinth effect in the coating, extending the penetration path of corrosive media and improving water resistance; 5-10 parts of talc powder (particle size 15-30μm, whiteness ≥90%): The layered structure can improve the leveling of the coating and reduce the curing shrinkage rate.

[0046] Solvent: Used to adjust the viscosity of the coating. It must be compatible with the base material and have a moderate evaporation rate.

[0047] 8-12 parts xylene (industrial grade, purity ≥99%): excellent solubility in fluorocarbon and epoxy resins, with a moderate evaporation rate; 5-8 parts butyl acetate (industrial grade, purity ≥99%): when combined with xylene, the evaporation rate can be adjusted to prevent the coating from drying too quickly and causing pinholes.

[0048] Additives: Improve the application performance and storage stability of coatings.

[0049] Dispersant 1-2 parts (polycarboxylate): disperses pigments and fillers through electrostatic repulsion, preventing agglomeration; Defoamer 0.5-1 part (organosilicon): reduces the surface tension of the system and eliminates bubbles generated during grinding and stirring; Leveling agent 0.3-0.5 parts (acrylate): promotes uniform spreading of the coating on the substrate surface and reduces orange peel effect; Ultraviolet absorber 0.8-1.2 parts (benzotriazole): absorbs 290-400nm ultraviolet light and reduces base material degradation.

[0050] According to the above-mentioned mass proportions, use an electronic balance (accuracy 0.01g) to accurately weigh each raw material, put them into numbered sealed containers, and store them in a dry and ventilated raw material warehouse (temperature 20~25℃, relative humidity ≤60%) for later use.

[0051] S2: Pigment and filler pretreatment

[0052] The moisture content and particle size distribution of pigments and fillers directly affect the subsequent premixing and grinding effects (for example, excessive moisture content can cause foaming during coating curing, and uneven particle size increases grinding difficulty). Therefore, pretreatment is necessary. The pretreatment process for pigments and fillers is as follows: Figure 3 As shown.

[0053] Drying and dehydration: Pour all the pigments and fillers weighed in S1 (5-8 parts nano titanium dioxide, 15-20 parts zinc powder, 8-12 parts mica powder, and 5-10 parts talc powder) into the tray of the vacuum drying oven (spreading thickness ≤ 5cm to ensure uniform heating). Set the drying temperature to 80-100℃ (too high a temperature will cause zinc powder to oxidize, and too low a temperature will result in low drying efficiency) and the vacuum degree to -0.08--0.09MPa (the vacuum environment accelerates moisture evaporation). Dry for 2-3 hours.

[0054] After drying, the moisture content was determined by gravimetric method: 10g of sample was placed in a weighing bottle and dried in an oven at 105℃ until constant weight, and the moisture content was calculated. ( This refers to the quality before drying. (For the quality after drying), it needs to be controlled. If the moisture content exceeds the standard, extend the drying time by 30 minutes and retest until it meets the standard.

[0055] Crushing and sieving: Transfer the dried and qualified pigments and fillers into a high-speed pulverizer (model SF-130, equipped with a stainless steel crushing chamber), adjust the crushing speed to 3000-4000 r / min (too low a speed will not crush agglomerated particles, too high a speed will cause zinc powder to oxidize due to frictional heat), and crush for 15-20 minutes. After crushing, pass the material through a 200-mesh standard sieve (75μm aperture), and collect the undersize material (the oversize material is returned to the pulverizer for re-crushing).

[0056] The particle size distribution of pigments and fillers under sieves was detected using a laser particle size analyzer (model MS2000). The required D50 (median particle size) was 50-100 μm (this range ensures uniform dispersion in S4 premixing while reducing the grinding load in S5). The pretreated pigments and fillers were then placed in sealed bags for later use.

[0057] S3: Base Material Pretreatment

[0058] The mixing uniformity and viscosity of the base material directly affect the dispersion effect of pigments and fillers (too low a viscosity will cause pigments and fillers to settle, while too high a viscosity will make dispersion difficult), and pretreatment is required to adjust them to a suitable state.

[0059] Mixing and stirring: Add 30-40 parts of S1 weighed fluorocarbon resin and 10-15 parts of epoxy resin to a 500L stainless steel reactor (with jacket heating and anchor stirrer). Start the stirring device and set the stirring speed to 300-500 r / min (too low a speed will result in uneven mixing, and too high a speed will easily entangle air bubbles). At the same time, turn on the jacket heating and slowly raise the system temperature to 60-70℃ (this temperature can reduce the viscosity of the base material and promote molecular diffusion mixing, but it is below the glass transition temperature of fluorocarbon resin to avoid resin degradation).

[0060] Viscosity adjustment: Keep the temperature and speed constant, and stir for 1 to 2 hours. During this period, take 50 mL of sample every 20 minutes with a sampling spoon and use a rotational viscometer (model NDJ-5S, 25℃, rotor No. 3, 100 r / min) to test the viscosity of the mixed base. The target viscosity is 500 to 800 mPa·s (this range can ensure that pigments and fillers are uniformly coated in S4).

[0061] If the viscosity is below 500 mPa·s (due to fluctuations in resin solid content), add 1-2 parts of fluorocarbon resin (fluorocarbon resin has a higher viscosity than epoxy resin, which can quickly increase the viscosity of the system), stir for 10 minutes, and then retest.

[0062] If the viscosity is higher than 800 mPa·s (due to uneven distribution of resin molecular weight), add 0.5 to 1 part xylene (the solvent can reduce intermolecular forces), stir for 10 minutes and retest until the viscosity meets the standard.

[0063] Material separation and storage: Divide the qualified mixed base material into two parts. The first part accounts for 60% to 70% of the total mass (used for S4 premixing, as a dispersion medium for pigments and fillers), and the second part accounts for 30% to 40% (used for S6 paint mixing, to adjust the final paint viscosity). Transfer them to two sealed base material storage tanks with agitators (to prevent settling during standing) and store them at room temperature (20 to 25°C) for later use. The storage time should not exceed 24 hours (to avoid viscosity changes caused by solvent evaporation).

[0064] S4: Premix

[0065] The purpose of premixing is to initially disperse the pretreated pigments and fillers with part of the mixed base material to form a uniform slurry, laying the foundation for efficient grinding of S5 (if the premixing is uneven, too many pigment and filler agglomerates will lead to prolonged grinding time and increased energy consumption of S5).

[0066] Base material preparation: Inject the first part of the mixed base material obtained by S3 into a 1000L dispersion kettle (with a high-speed dispersion disc, 30cm in diameter), start the disperser, set the speed to 800-1000r / min (this speed can form a moderate vortex, which is convenient for the pigments and fillers to be evenly dispersed), and stir at a constant speed for 5 minutes to stabilize the base material system.

[0067] Adding pigments and fillers: Use a screw feeder (50 r / min) to slowly add all the pretreated pigments and fillers obtained from S2 into the dispersion vessel, controlling the feeding speed at 5-8 kg / min (too fast a speed will result in excessively high local pigment and filler concentrations, forming agglomerates that are difficult to disperse; too slow a speed will result in low efficiency). The feeding process should last 30-60 minutes (adjusted according to the total mass of pigments and fillers). During feeding, the dispersion disc should be submerged 5-10 cm below the surface of the base material to prevent pigments and fillers from flying away.

[0068] Mixing uniformity test: After feeding, maintain the disperser speed at 800-1000 r / min and continue stirring for 1-1.5 hours to allow the pigments and fillers to initially bond with the base material. After stirring, visually inspect the uniformity of the premixed slurry: Take a small amount of slurry and place it on a glass slide, spread it evenly with a scraper, and observe it under natural light. There should be no obviously visible particles (particle size ≥ 100 μm). If particles are present, extend the stirring time by 20 minutes and test again until the uniformity meets the standard.

[0069] Initial fineness record: The initial fineness of the premixed slurry after reaching the standard was measured using a scraper fineness gauge (range 0-150μm). (Take the average of 3 tests) and record the data (range 80-120μm). This value will serve as a key input parameter for the BP neural network prediction algorithm in S5, so the detection accuracy must be ensured (error ≤ 5μm). Transfer the premixed slurry to a transfer tank and then pump it to the feed inlet of the horizontal sand mill in S5 for later use.

[0070] S5: Grinding and Dispersion

[0071] In the process of grinding and dispersing premixed slurry in a horizontal sand mill, the original process controls the grinding effect by fixing the rotation speed and sampling at intervals to detect the fineness. However, this method suffers from low efficiency, high energy consumption, and poor fineness stability. To address these issues, this step introduces an intelligent control strategy that integrates Adaptive Parameter Adjustment (APA), Backpropagation (BP) neural network prediction algorithm, and Particle Swarm Optimization (PSO) algorithm, forming a closed-loop system of real-time prediction, parameter optimization, dynamic adjustment, and data feedback. The grinding and dispersion process is as follows: Figure 4 As shown.

[0072] Specifically, the BP algorithm predicts the current slurry fineness by collecting grinding parameters in real time, reducing the cost of frequent sampling; the PSO algorithm, based on the prediction results of BP, optimizes the grinding speed and time while meeting the fineness requirements, reducing energy consumption; the APA algorithm combines the optimal parameters output by PSO with the actual detected fineness deviation to correct operating parameters in real time, compensating for system fluctuations such as raw material batch differences and equipment wear; finally, the actual operating data generated by APA feeds back into the BP algorithm, continuously optimizing its prediction accuracy. The three work synergistically to achieve efficient, accurate, and low-consumption control of the grinding process.

[0073] Algorithm 1: BP Neural Network Prediction Algorithm (BP) – Real-time Prediction of Grinding Fineness

[0074] Core function: Based on key parameters in the grinding process (rotation speed, time, initial fineness), a nonlinear mapping model is established to predict the fineness of the slurry at the current moment in real time, providing basic data for parameter optimization of the PSO algorithm, while reducing the frequency of manual sampling and testing every 30 minutes.

[0075] 1.1 Model Construction

[0076] The backpropagation (BP) neural network is designed based on a three-layer architecture of input-hidden-output, and the parameter dimensions and number of neurons are optimized for the characteristics of the grinding process.

[0077] Input layer (3 neurons): The three parameters that have the most significant impact on grinding fineness are selected as inputs, specifically: the current operating speed of the horizontal sand mill. (Unit: r / min), value range 1500~2000 r / min (based on the rated speed range of the equipment); cumulative grinding time (Unit: min), value range 0–180 min (covering the original process grinding time of 30–180 min); initial fineness of premixed slurry (Unit: μm), which is the detection value of the premixed slurry in step S4, ranging from 80 to 120 μm.

[0078] Hidden layers; after multiple experiments (testing different combinations of 5-15 neurons), it was determined that two hidden layers should be used:

[0079] The first hidden layer contains 10 neurons and is used to extract low-order features of the input parameters (such as the linear relationship between rotation speed and time).

[0080] The second hidden layer contains 8 neurons, which are used to fuse low-order features and extract high-order features (such as the interaction of rotation speed, time, and initial fineness).

[0081] Both hidden layers use the Sigmoid activation function. This function can map input values ​​to the 0-1 interval, enhancing the model's nonlinear fitting ability.

[0082] Output layer (1 neuron): Output is the predicted slurry fineness at the current time step. (Unit: μm), the goal is to make The error between the actual fineness and the actual fineness is ≤5μm.

[0083] Network structure formula: Let the output of the j-th neuron in the first hidden layer be... The output of the k-th neuron in the second hidden layer is ,but:

[0084] (j=1,2,...,10);

[0085] (k=1,2,...,8);

[0086] ;

[0087] For the i-th parameter of the input layer ( , , );

[0088] The connection weights from the input layer to the first hidden layer (i=1,2,3 correspond to the input parameters, j=1,...,10 correspond to the neurons in the first layer);

[0089] This is the bias term for the j-th neuron in the first hidden layer, used to adjust the activation threshold of the neuron;

[0090] The connection weights from the first hidden layer to the second hidden layer (k=1,...,8 correspond to neurons in the second layer, j=1,...,10 correspond to neurons in the first layer);

[0091] This is the bias term for the k-th neuron in the second hidden layer;

[0092] The connection weights from the second hidden layer to the output layer (k=1,...,8 correspond to the neurons in the second layer);

[0093] This is the bias term for the output layer.

[0094] 1.2 Model Training

[0095] To ensure the prediction accuracy of the BP algorithm, the model needs to be trained using historical production data. The specific process is as follows:

[0096] Training data acquisition and preprocessing: Collect grinding production data from 500 batches over the past 6 months (each batch corresponds to a premixed slurry with one raw material ratio). Each batch of data includes:

[0097] Input parameters: (Real-time rotation speed is recorded every 10 minutes) (Cumulative time corresponding to the recorded moment) (Initial fineness of step S4 in this batch);

[0098] Output label: Actual level of detail at the corresponding moment (Manual inspection using a scraper fineness gauge).

[0099] Data preprocessing: outliers (such as sudden drops in speed due to equipment failure) are removed, and the input parameters are mapped to the [0,1] interval using Min-Max standardization to avoid affecting the training effect due to differences in units.

[0100] Loss function definition; the mean squared error (MSE) is used to measure the deviation between the predicted and actual values, and the formula is: ;

[0101] in, The total sample size is... Let m be the prediction granularity for the m-th sample. This represents the actual detection detail for the m-th sample.

[0102] Training process: Gradient descent is used to update network weights and biases. Specific parameter settings include: learning rate. (Control the weight update step size to avoid convergence oscillations); Number of iterations: 1000 (tested, the loss function tends to stabilize after 1000 iterations); Training termination condition: when When the model converges, training stops.

[0103] After training is complete, save the final weights and bias parameters. ), used for subsequent real-time prediction.

[0104] 1.3 Model Application

[0105] During the grinding process, the system collects the following data in real time and inputs it into the trained BP model:

[0106] Current speed of horizontal sand mill (Acquired in real time via device sensors); Cumulative grinding time (Timing starts from the start of grinding, automatically recorded by the system clock); Initial fineness of the premixed slurry (Fixed value, taken from the detection result of step S4).

[0107] Model output prediction detail This result will be directly used as the input parameter for the PSO algorithm to optimize the grinding parameters.

[0108] Algorithm 2: Particle Swarm Optimization (PSO) Algorithm – Global Optimization of Grinding Parameters

[0109] Core function: to predict the fineness of the backpropagation algorithm As a constraint, within the feasible range of grinding speed and time, we seek the optimal combination of parameters that minimizes grinding energy consumption, providing an adjustment target for the APA algorithm to achieve the dual objectives of achieving the fineness target and minimizing energy consumption.

[0110] 2.1 Model Construction

[0111] The PSO algorithm simulates the foraging behavior of bird flocks and searches for the optimal solution by moving particles in the solution space. The model parameters are designed as follows to address the optimization objective of this step:

[0112] Particle definition:

[0113] Each particle represents a set of grinding parameters to be optimized, i.e. ,in:

[0114] Grinding speed (unit: r / min), with a range matching the equipment: ;

[0115] Grinding time (unit: min), based on the time range of the original process: .

[0116] Fitness function design:

[0117] The fitness function is used to evaluate the quality of particles, and it needs to consider two objectives simultaneously: achieving the required fineness and minimizing energy consumption. The formula is: ;

[0118] in: (Target fineness, consistent with the original S5 process requirements);

[0119] Fineness penalty: If the prediction fineness does not meet the standard ( If the value is positive, the fitness value increases (the particle is penalized); if the target is met ( If the value is 0, then no penalty is imposed.

[0120] The grinding energy consumption (unit: kWh) is proportional to the cube of the rotational speed and the time, depending on the equipment characteristics. The formula is as follows: ,in (Obtained through equipment calibration experiments: Under idling conditions, the energy consumption at different speeds and times was tested, and the proportional coefficient was obtained by fitting.) );

[0121] (Fineness weight) (Energy Consumption Weight): Through orthogonal experiments, this weight combination can ensure that the fineness requirements are met first (avoiding the sacrifice of fineness to reduce energy consumption), while effectively optimizing energy consumption.

[0122] 2.2 Model Training

[0123] The PSO algorithm finds the optimal parameters through iterative updates of particles. The specific steps are as follows:

[0124] Initialize Particle Swarm: Set the particle swarm size =30 (Testing showed that 30 particles strike a balance between computational complexity and optimization accuracy), randomly generate the initial positions of 30 particles. and initial velocity :

[0125] Initial position: Randomly take values within [1500, 2000], Randomly take values within [30, 180];

[0126] Initial velocity: The velocity range is limited to (Avoid sudden changes in rotational speed), (Avoid excessive time adjustment).

[0127] Update of individual optimal and global optimal: Individual optimal : Record the position with the minimum fitness during the iterative process of the i-th particle (i.e., the minimum );

[0128] Global optimal : Record the position with the minimum fitness among all particles (i.e., the optimal parameters of the entire particle swarm).

[0129] Particle position and velocity update formula; At the (k + 1)-th iteration, the velocity and position of the particle are updated as follows:

[0130] ;

[0131] ;

[0132] ;

[0133] ;

[0134] k is the iteration number (initial k = 0); = 0.8 (inertia weight, controlling the tendency of the particle to maintain its original velocity); = 2 (learning factor, respectively controlling the intensity of the particle learning from the individual optimal and global optimal); (random number, increasing the randomness of the optimization search); 、 are the rotational speed and time of the individual optimal position of the i-th particle; 、 are the rotational speed and time of the global optimal position; is the constraint function, if x < a, then output a, if x > b, then output b, otherwise output x (ensuring that the parameters are within the allowable range of the device).

[0135] Termination condition: When the iteration number reaches 50, stop the optimization search and output the global optimal parameters .

[0136] 2.3 Model application

[0137] The PSO algorithm receives the prediction granularity output from the BP algorithm in real time. Substitute into the fitness function In the process, through the above 50 iterations of optimization, the result is obtained that... Minimum optimal parameters This optimal parameter will serve as the target adjustment value for the APA algorithm, guiding real-time parameter correction.

[0138] Algorithm 3: Adaptive Parameter Adjustment Algorithm (APA) – Real-time Dynamic Parameter Correction

[0139] Core Function: Due to factors such as batch variations in raw materials (e.g., fluctuations in pigment and filler particle size) and equipment wear (e.g., wear of grinding beads in a sand mill), the optimal parameters output by the PSO algorithm may deviate from actual operating conditions. The APA algorithm compares the deviation between the actual detected fineness and the target value, and combines this with the optimal parameters of PSO to adjust the grinding speed and time in real time, ensuring that the final fineness consistently meets the standards.

[0140] 1.1 Model Construction

[0141] The APA algorithm is based on the feedback control principle and achieves dynamic correction through three steps: deviation calculation, adjustment calculation, and parameter update.

[0142] Deviation calculation: every 60 minutes (or when BP forecast granularity) A manual sample is taken once every hour, and the actual fineness is measured using a scraper fineness gauge. (k is the number of samples, initially k=1), calculate the bias: ;in, .like This indicates that the actual fineness does not meet the standard and the grinding intensity needs to be increased; if This indicates that the standard has been met and the intensity can be maintained or reduced.

[0143] Adjustment calculation: based on deviation And the optimal parameters output by PSO Calculate the speed adjustment amount at the current moment. and time adjustment amount :

[0144] ;

[0145] ;

[0146] , This represents the actual operating speed and time at the previous moment (after the (k-1)th sampling).

[0147] (Speed ​​ratio coefficient): Based on regression analysis of historical data, it is found that for every 1μm of positive deviation, the speed needs to be increased by 2r / min to compensate.

[0148] (Time Proportioning Factor): Similarly, for every 1μm of positive deviation, the grinding time needs to be extended by 0.5min;

[0149] , (Optimal parameter tracking coefficient): Ensure that the adjustment direction is close to the optimal PSO parameter, and avoid the parameter deviating from the global optimum for a long time due to local deviation.

[0150] Actual parameters updated:

[0151] The rotational speed and time after the kth adjustment are: ; ;

[0152] At the same time, it is necessary to ensure that the parameters are within the allowable range of the device through constraint functions:

[0153] like ,but ;like ,but ;

[0154] like ,but ;like ,but .

[0155] 1.2 Model Training

[0156] The core of the APA algorithm is determining the adjustment coefficient. The training process is as follows:

[0157] Collect 300 sets of historical deviation data (including data from different raw material batches and equipment conditions). and the corresponding optimal adjustment amount).

[0158] Adjusted fineness deviation To achieve the target, the least squares method was used to fit the coefficients, and the above coefficient values ​​were finally determined (after fitting verification, this combination of coefficients can enable 95% of the adjustment cases to meet the target after one correction).

[0159] 1.3 Model Application

[0160] The APA algorithm runs in real time according to the following process: 1. Receive the optimal parameters output by the PSO algorithm. 2. Take samples every 60 minutes (or when the BP predicted fineness is close to the target) to detect the actual fineness. Calculate the deviation 3. Substitute into the adjustment formula to calculate. and Update actual operating parameters 4. The new samples are fed back into the BP algorithm to update its network weights (the BP model is retrained once for every 10 new samples accumulated).

[0161] The three algorithms form a closed-loop collaboration through real-time data transmission, with the specific interaction as follows:

[0162] 1. BP→PSO interaction:

[0163] The BP algorithm outputs the prediction detail every 5 minutes. The PSO algorithm substitutes this value into the fitness function. In this context, it serves as a fineness constraint. For example, if... (Not meeting the standard), then Medium-level penalty items are The particles need to be adjusted towards higher rotation speeds and longer durations to reduce the penalty; if If the target is met, the penalty is 0, and the particles will preferentially adjust to lower speed and shorter time to reduce energy consumption.

[0164] 2. PSO→APA Interaction:

[0165] The optimal parameters output by the PSO algorithm As the target anchor point of the APA algorithm, the adjustment amount formula is used... and This affects real-time adjustments. For example, if the current actual speed... ,and Then the contribution The adjustment amount pushes the actual speed closer to the optimal value.

[0166] 3. APA→BP interaction:

[0167] Actual running data generated by the APA algorithm New samples are added to the training set of the backpropagation (BP) network to update its weights. For example, when changes in raw material batches lead to variations in the initial fineness... As the level increases, the new samples correct the prediction bias of the BP model under high initial detail, thus... It is closer to the actual value.

[0168] For example, taking a batch of premixed slurry as an example, the specific application of algorithm fusion will be explained in detail:

[0169] Initial conditions: Initial fineness of the premixed slurry in step S4 Initial rotation speed at the start of grinding Initial time (First sampling time).

[0170] Step 1: BP prediction and PSO optimization (at 30 min)

[0171] 1. Input for the BP algorithm: , , ;

[0172] The output of the first hidden layer is calculated using the trained backpropagation (BP) model. (After processing with the Sigmoid function, the value is between 0 and 1); Output of the second hidden layer (Also between 0 and 1); final prediction detail (Not up to standard).

[0173] 2. The PSO algorithm uses Input, initiate optimization:

[0174] The position of a particle in the initial particle swarm is fitness ;

[0175] After 50 iterations, the globally optimal particle position is: At this point, BP predicts the fineness under this parameter. (Meets standards), adaptability (This is the minimum value).

[0176] Step 2: APA adjustment (sampling at 60 min)

[0177] 1. Manual sampling to test actual fineness (Due to the slightly larger particle size of the zinc powder in the raw material, the actual fineness is 3μm higher than the predicted value).

[0178] 2. Calculate the deviation: ;

[0179] 3. Parameters from the previous time step: , ;

[0180] 4. Calculate the adjustment amount:

[0181] ;

[0182] ;

[0183] 5. Update parameters:

[0184] (Valid within the range of 1500-2000);

[0185] (Valid within the range of 30 to 180);

[0186] Step 3: Feedback Updates and Secondary Optimization (at 120 minutes)

[0187] 1. Take the new sample By inputting the backpropagation (BP) algorithm, retraining the model, and updating the weights, the BP algorithm improves the prediction accuracy of the current parameters.

[0188] At 2.120 min, the BP prediction fineness (Approaching the target), PSO optimization yields new optimal parameters. ;

[0189] 3. Manual sampling to test actual fineness ,deviation ;

[0190] 4. APA Adjustment Amount:

[0191] ;

[0192] ;

[0193] 5. After updating the parameters, the rotational speed... ,time At this point, the actual fineness meets the standard. ).

[0194] Final results: Through the fusion of three algorithms, this batch of slurry achieved a fineness of ≤50μm in 120min, which shortened the time by 20% and reduced energy consumption by 18% compared with the original process (average 150min) (due to avoiding blindly increasing the rotation speed). Moreover, the fineness stability (standard deviation ≤3μm) was significantly better than the original process (standard deviation ≥8μm).

[0195] Core contribution: The three elements work together to achieve intelligent perception, global optimization, and dynamic correction in the grinding process, solving the problems of traditional processes that rely on experience, are inefficient, and have poor stability. This provides a slurry base with uniform fineness and stable performance for subsequent paint mixing steps, indirectly improving the weather resistance and corrosion resistance of the final coating.

[0196] S6: Paint Mixing

[0197] Paint mixing is a crucial step in adjusting the viscosity and performance required for application by mixing the ground slurry with the remaining base material, additives, and solvents. Strict control of the order of addition and mixing conditions is essential (improper addition of additives can reduce synergistic effects; for example, adding defoamers and dispersants simultaneously will decrease the dispersion effect).

[0198] Base material compounding: Inject the second part of the mixed base material stored in S3 into the ground slurry in the intermediate storage tank (with stirring function), start the tank stirring device, set the speed to 600-800 r / min (this speed can avoid slurry splashing and ensure uniform mixing of base materials), stir for 10 minutes to make the system initially homogenized.

[0199] Additives: Add all the weighed additives in S1 in the following order: dispersant → defoamer → leveling agent → UV absorber (each additive needs to be diluted with a small amount of xylene beforehand to avoid excessively high local concentrations):

[0200] After adding the dispersant, stir for 15 minutes (to enhance the stability of pigments and fillers and prevent secondary agglomeration); after adding the defoamer, stir for 15 minutes (to eliminate air bubbles generated during base material compounding); after adding the leveling agent, stir for 20 minutes (to promote system leveling during subsequent solvent adjustment); after adding the UV absorber, stir for 20 minutes (to ensure uniform dispersion and improve overall weather resistance).

[0201] Solvent adjustment and viscosity control: Add 8-12 parts of xylene and 5-8 parts of butyl acetate (mix the two solvents at a mass ratio of 1.5:1 for better matching of evaporation rates), and continue stirring for 2-3 hours. During this period, control the system temperature to 25-35℃ by cooling with a jacket or heating (too high a temperature will cause the solvent to evaporate too quickly, while too low a temperature will result in incomplete dispersion of the additives).

[0202] After stirring, use a rotational viscometer (same as S3, 25℃, rotor No. 4, 50 r / min) to test the viscosity of the initial mixed coating. The target viscosity is 1000~1500 mPa·s (this range is suitable for brush or spray application):

[0203] If the viscosity is too high (e.g., 1600 mPa·s), add 1-2 parts of xylene (xylene evaporates relatively quickly, which can rapidly reduce the viscosity), stir for 10 minutes, and then test again.

[0204] If the viscosity is too low (e.g., 900 mPa·s), add 1 to 2 parts of the mixed base material prepared by S3, stir for 10 minutes and retest until the viscosity meets the standard, and obtain the paint after mixing.

[0205] S7: Curing process

[0206] The purpose of curing is to allow the components in the coating system to fully diffuse and adsorb, forming a stable colloidal structure (uncured coatings may separate or thicken during storage), and at the same time, to verify whether the coating meets the standards through performance testing.

[0207] Standing curing: After the paint meets the S6 standard, transfer the mixed paint into a 1000L curing tank (with jacket insulation and low speed stirrer), seal the tank, set the curing temperature to 25-30℃ (close to room temperature to avoid temperature fluctuations that may cause system instability), and let it stand for 12-24 hours (too short a time will result in incomplete curing, too long a time may cause solvent evaporation).

[0208] Intermediate stirring: During the curing process, start the stirring device of the curing tank every 6 hours and stir at 300r / min for 10 minutes (low-speed stirring can promote the uniform distribution of the additives in the system and avoid local enrichment). After stirring, let it stand.

[0209] Performance Testing and Feedback: After curing, samples are taken for testing of key performance characteristics (3 parallel samples are taken for each test, and the average value is used):

[0210] Weather resistance: Using a QUV aging test chamber (model QUV / se), according to ASTM G154 standard, aged for 1000 hours under UVB-313 lamps, light temperature of 60℃, and condensation temperature of 50℃, the gloss retention rate is required to be ≥80% (gloss meter model BYK-Gardner4520, 60° angle test).

[0211] Adhesion: According to the cross-cut test standard for paint and varnish film, use a cross-cut tester to draw a 1mm×1mm grid, observe after peeling off the tape, and the required grade should be ≤1;

[0212] Salt spray resistance: Using a salt spray test chamber (model YWX / Q-150), according to the test standards for the neutral salt spray resistance of paints and varnishes, 5% NaCl solution, temperature 35℃, continuous spraying for 5000h, the coating is required to be free of rust and blistering.

[0213] Once all performance indicators meet the standards, proceed to the next step; if any indicator fails to meet the standards (such as bubbling in salt spray resistance), return to S6 to readjust the paint (add 0.1 to 0.2 parts of UV absorber to enhance weather resistance, or add 0.1 parts of dispersant to improve pigment and filler dispersion), and re-curing and testing until the standards are met.

[0214] S8: Filter Packaging

[0215] Filtration can remove any tiny impurities that may remain in the system (such as incompletely ground pigment and filler particles), and packaging must ensure airtightness to prevent solvent evaporation. The specific procedures are as follows:

[0216] Precision filtration: After S7 performance meets the standards, the cured coating is filtered through a 100-mesh stainless steel filter screen (150μm pore size, finer than the S5 grinding fineness, ensuring the removal of impurities). A diaphragm pump provides power, and the filtration pressure is controlled at 0.3–0.5 MPa (too low a pressure will slow down the filtration speed, and too high a pressure will cause the screen to break). During the filtration process, the screen should be checked for blockage after every 500L of coating is filtered (judged by the pressure difference between the inlet and outlet; the screen should be replaced when the pressure difference is ≥0.1MPa).

[0217] Quantitative filling: The filtered finished coating is transferred to an automatic filling machine (accuracy ±0.1kg), and filled into iron drums with epoxy resin inner walls at a specification of 20kg per drum (to prevent the coating from reacting with the iron drum). During the filling process, the weight displayed on the electronic scale is monitored in real time. If the error exceeds ±0.1kg, the machine is immediately stopped and the flow valve of the filling machine is calibrated.

[0218] After sealing the packaging drum, store it in the finished product warehouse (temperature 15~30℃, well ventilated, away from fire sources), with a shelf life of 12 months.

[0219] Through meticulous control of each step and optimization of the S5 algorithm, the final coating exhibits superior weather resistance (gloss retention rate ≥85% after 1000h aging) and salt spray resistance (no rust after 5000h) compared to traditional processes (improving by 10% and 20% respectively). Furthermore, production efficiency is increased by 20%, and energy consumption is reduced by 18%, making it suitable for heavy corrosion protection scenarios such as marine engineering and chemical equipment.

[0220] In some embodiments, to further enhance the adaptability of the grinding process to nonlinear disturbances (such as grinding bead wear and raw material hardness fluctuations), a fuzzy PID algorithm (FPID) is introduced based on the original adaptive parameter adjustment algorithm (APA), BP neural network prediction algorithm (BP), and particle swarm optimization algorithm (PSO). These four algorithms are integrated to form a multi-layered closed loop of prediction-optimization-fuzzy decision-dynamic correction-feedback: BP prediction provides basic data for PSO and FPID; PSO outputs globally optimal parameters as a reference target for FPID; FPID dynamically adjusts PID parameters through fuzzy logic to correct grinding deviations; and APA implements the final execution of parameters based on the output of FPID, while simultaneously feeding actual data back to the BP update model.

[0221] Algorithm 4: Fuzzy PID Algorithm (FPID) – Nonlinear Deviation Correction

[0222] Core function: To address the fineness deviation caused by nonlinear factors such as equipment wear (e.g., reduced grinding efficiency due to decreased zirconium oxide bead size) and raw material hardness fluctuations (e.g., talc hardness deviation) during the grinding process, the fuzzy logic is used to adjust the proportional coefficient, integral coefficient, and derivative coefficient of the PID controller in real time, achieving more accurate deviation correction than traditional PID and providing better adjustment instructions for the APA algorithm.

[0223] 4.1 Model Construction

[0224] The FPID algorithm consists of two parts: a fuzzy controller and a PID controller. The fuzzy controller dynamically adjusts the PID parameters based on the deviation, and the PID controller outputs the final adjustment amount.

[0225] Input variables (fuzzy controller):

[0226] Select the two variables that have the most direct impact on grinding deviation: fineness deviation. (Unit: μm): The difference between the current actual fineness and the target fineness. ,in The actual fineness received by the FPID module (taken from the detection data of APA). (Target value); Rate of change of deviation (Unit: μm / min): The change in deviation per unit time. ,in The deviation of the k-th detection is... For the (k-1)th deviation, (Detection interval, consistent with APA sampling frequency).

[0227] The fuzzy sets of the input variables are: {Negative Large (NB), Negative Medium (NM), Negative Small (NS), Zero (ZO), Positive Small (PS), Positive Medium (PM), Positive Large (PB)}, with the following universe of discourse: (Based on historical deviation range settings).

[0228] Output variables (fuzzy controller):

[0229] The output is the adjustment amount of three parameters of the PID controller, used to correct the initial PID parameters: proportional coefficient adjustment amount. Integral coefficient adjustment Adjustment amount of differential coefficients .

[0230] The fuzzy sets of the output variables are all: {Negative Large (NB), Negative Medium (NM), Negative Small (NS), Zero (ZO), Positive Small (PS), Positive Medium (PM), Positive Large (PB)}, with the following universe of discourse: (Determined through orthogonal experiments to ensure that the parameter adjustment range is reasonable).

[0231] Fuzzy rule table: Based on historical data, 49 fuzzy rules are formulated (7×7 input combinations), for example:

[0232] like and (If the deviation is large and continues to increase), then , , (Increase the proportional effect, weaken the integral effect, and enhance the differential effect to suppress overshoot); if and (If the deviation is zero and stable), then , , (Keep the parameters unchanged).

[0233] PID controller formula: Let the PID parameters after fuzzy adjustment be:

[0234] ; ; ;in , , These are the initial PID parameters (tuned using the Ziegler-Nichols method).

[0235] Speed ​​correction amount output by PID controller and time correction amount for:

[0236] ;

[0237] ;

[0238] The formulas for speed and time corrections have the same structure, but due to different units, they need to be multiplied by dimensional coefficients in practical applications: speed correction multiplied by... Time correction amount multiplied by .

[0239] 4.2 Model Training

[0240] Training data: Collect 300 sets of grinding data under nonlinear perturbation (such as data after 50 hours of grinding bead use, and data on raw material hardness fluctuation of ±10%), each set containing And the corresponding optimal PID parameter adjustment amount (determined through trial and error, i.e., manually adjusting the parameters to make the deviation converge the fastest).

[0241] Membership function optimization: A triangular membership function is used, and the function parameters (such as vertex coordinates) are fitted by the least squares method to make the error between the output of the fuzzy controller and the optimal adjustment amount ≤5%.

[0242] Rule Validation: Validate the optimized rule table to ensure that in more than 90% of the test cases, the deviation convergence time is shortened by ≥30% compared with the initial rule; otherwise, return to adjust the rule table.

[0243] 4.3 Model Application

[0244] The FPID algorithm receives the following inputs in real time:

[0245] 1. Actual level of detail transmitted by the APA module Calculate the deviation and rate of change of deviation ;

[0246] 2. Optimal parameters output by the PSO module As a target reference for PID control, the adjustment direction is influenced by the weight terms in the fuzzy rules.

[0247] Output via fuzzy controller After updating the PID parameters, calculate the speed and time correction. This is then fed into the APA algorithm as part of the final adjustment.

[0248] Four-algorithm fusion logic and interaction process

[0249] 1. Interaction between BP→PSO and BP→FPID: Prediction fineness of BP algorithm output It serves as input to both PSO and FPID:

[0250] For PSO: Used to calculate the fitness function ;

[0251] For FPID: When Compared to actual detail When the deviation is ≥5μm (indicating a nonlinear error in the prediction), the fuzzy rule strength of FPID is increased (by adding weight to the rate of change of deviation), for example: This enhances the response to prediction bias.

[0252] 2. PSO→FPID interaction:

[0253] Optimal speed output by PSO and time As the target constraint of FPID, a weight term is added to the fuzzy rule:

[0254] If the current speed Then, the adjustment amount of the proportional coefficient for positive deviation in the fuzzy rule increases by 20% (i.e., This pushes the actual rotational speed closer to the optimal value;

[0255] Similarly, time adjustment ensures that the direction of FPID correction is consistent with the global optimum.

[0256] 3. FPID→APA Interaction: The correction amount of the FPID output. Combined with the original adjustment amount of APA, it becomes the final adjustment amount: ; ;in =0.5 (weighting coefficient, determined experimentally to balance linear adjustment and nonlinear correction).

[0257] The final parameters are updated as follows:

[0258] ;

[0259] ;

[0260] 4. APA→BP and APA→FPID feedback: Actual operating data of APA Simultaneously, feedback is sent to BP and FPID:

[0261] For BP: Update training samples and optimize the prediction model;

[0262] For FPID: Compare the actual deviation convergence curve with the ideal curve (overshoot ≤10%, settling time ≤60min). If the deviation is ≥20%, then fine-tune the vertex coordinates of the fuzzy membership function using the gradient descent method (e.g., adjust the vertex of "PB" from 30μm to 28μm) to improve rule adaptability.

[0263] For example

[0264] Taking the scenario where grinding efficiency decreases due to wear of the grinding beads as an example (grinding efficiency decreases by 15% after 50 hours of use), the initial conditions are the same as before: .

[0265] Step 1: BP prediction and PSO optimization (at 30 min)

[0266] 1. BP input: Because the wear of the grinding beads was not taken into account, the predicted fineness was... ;

[0267] 2. PSO optimization yielded .

[0268] Step 2: APA detection and FPID triggering (at 60 minutes)

[0269] 1. APA sampling and testing: Due to decreased grinding efficiency, the actual fineness... (7μm higher than normal), deviation ;

[0270] 2. Original adjustment amount for APA:

[0271] ;

[0272] ;

[0273] 3. FPID input:

[0274] deviation (Exceeding the domain [-30, 30], truncation is set to 30 μm);

[0275] Deviation change rate (Initial deviation is 0);

[0276] because Triggering a fix: .

[0277] 4. FPID fuzzy inference:

[0278] Input fuzzification: ;

[0279] Look up the rule table: Output (Take 2) (Take -0.05) (Take 0.8);

[0280] Update PID parameters: , , .

[0281] 5. Calculation of FPID correction (multiplied by dimensionless coefficient):

[0282] ;

[0283] ;

[0284] Step 3: Integration, Adjustment, and Parameter Update

[0285] 1. Total adjustment amount ( =0.5):

[0286] ;

[0287] ;

[0288] 2. Update parameters:

[0289] ;

[0290] ;

[0291] Step 4: Feedback Optimization (at 120 minutes)

[0292] 1. Actual fineness testing: , (Significantly improved compared to 70μm without FPID);

[0293] 2. Data feedback to BP and FPID: BP updates the model to adapt to the impact of grinding bead wear, and FPID verifies the deviation convergence curve (overshoot 8% ≤ 10%). No rule adjustment is required.

[0294] 3. Finally, the fineness was achieved at 150 minutes (48μm), which shortened the time by 30 minutes compared to using only three algorithms.

[0295] Key contributions: The introduction of FPID improves the system's response speed to nonlinear disturbances by 40%, reduces the overshoot of fineness adjustment from 25% to 8%, especially in scenarios such as grinding bead wear and raw material hardness fluctuations, the final coating fineness standard deviation is further reduced from 3μm to 1.5μm, providing a more stable slurry base for subsequent paint mixing and curing steps, and indirectly improving salt spray resistance by another 5% (5000h no-abnormality rate increased from 98% to 100%).

[0296] In one specific embodiment, S1: Raw material preparation (accurately weighing by mass parts)

[0297] A batch of raw materials was selected and weighed according to the following mass percentages (unit: kg, accuracy 0.01 kg). All raw materials conform to the corresponding national standards:

[0298] Base material: 35 parts fluorocarbon resin (50% solid content, hydroxyl value 90mgKOH / g), 12 parts epoxy resin (epoxy value 0.55eq / 100g);

[0299] Pigments and fillers: 6 parts nano titanium dioxide (30nm particle size, anatase type), 18 parts zinc powder (8μm particle size, 99.6% purity), 10 parts mica powder (15μm particle size, flakes), and 7 parts talc powder (20μm particle size, 92% whiteness).

[0300] Solvents: 10 parts xylene (industrial grade, 99.2% purity), 6 parts butyl acetate (industrial grade, 99.1% purity).

[0301] Additives: 1.5 parts dispersant (polycarboxylate), 0.8 parts defoamer (organosilicon), 0.4 parts leveling agent (acrylate), 1.0 part ultraviolet absorber (benzotriazole).

[0302] The above raw materials were respectively placed into sealed containers numbered "Base Material-1", "Pigment and Filler-1", "Solvent-1" and "Auxiliary Agent-1" and stored in the raw material warehouse (temperature 23℃, humidity 55%) for later use.

[0303] S2: Pigment and filler pretreatment

[0304] Drying and dehydration: Pour the pigments and fillers (6kg nano titanium dioxide + 18kg zinc powder + 10kg mica powder + 7kg talc powder = 41kg) into the tray of the vacuum drying oven (spreading thickness 4cm), set the drying temperature to 90℃ and the vacuum degree to -0.085MPa, and dry for 2.5 hours.

[0305] Sampling and testing after drying: Take 10g of sample and measure its mass before drying. =10.03g, mass after drying =10.00g, moisture content W=(10.03-10.00) / 10.03×100%=0.3% (≤0.5%, meets the standard).

[0306] Crushing and sieving: Transfer the dried pigments and fillers into a high-speed pulverizer, set the speed to 3500 r / min, and crush for 18 minutes. After crushing, sieve through a 200-mesh sieve (75 μm aperture), collect 39 kg of the undersize material (2 kg of the oversize material is returned for regrinding).

[0307] Laser particle size analyzer test: D50 (median particle size) of the sieve material = 80μm (within the range of 50~100μm, which meets the standard), and the pretreated pigments and fillers are obtained and put into sealed bags for later use.

[0308] S3: Base Material Pretreatment

[0309] Mixing and stirring: Add 35kg of fluorocarbon resin and 12kg of epoxy resin (total mass 47kg) to a 500L reactor, start stirring (400r / min), and heat the jacket to 65℃.

[0310] Viscosity adjustment: Keep warm and stir for 1.5 hours, and take samples to test the viscosity every 20 minutes: 20 minutes: viscosity 620 mPa·s (25℃, rotor No. 3, 100 r / min); 40 minutes: viscosity 700 mPa·s (meets the standard, 500~800 mPa·s), stop adjusting.

[0311] Material separation and storage: Divide the 47kg mixed base material into two parts: Part 1: 47×65%=30.55kg (for S4 premixing); Part 2: 47×35%=16.45kg (for S6 paint mixing), and store them separately in two sealed storage tanks at room temperature.

[0312] S4: Premix

[0313] Base material preparation: Inject 30.55 kg of the first part of the mixed base material into a 1000L dispersion tank, start the disperser (900 r / min), and stir for 5 minutes.

[0314] Pigment and filler addition: Use a screw feeder to add 41 kg of pretreated pigments and fillers to the dispersion vessel, controlling the feeding rate at 6 kg / min. The feeding time is approximately 41 ÷ 6 ≈ 6.83 hours (410 minutes). During feeding, the dispersion disc should be submerged 8 cm below the liquid surface to prevent scattering.

[0315] Mixing uniformity test: After the material is added, stir at 900 r / min for 1.2 hours. Take a slurry and spread it evenly on a glass slide. There should be no visible particles (meets the standard).

[0316] Initial fineness record: The scraper fineness gauge was used to measure the fineness three times, with results of 88μm, 92μm, and 90μm respectively. The average value was... After recording the data, it is transferred to a transfer tank and then transported to the S5 sand mill.

[0317] S5: Grinding and Dispersion

[0318] Equipment preparation: Load zirconia beads (particle size 1.5mm) into a horizontal sand mill (WS-30 type), filling amount = 30L × 65% = 19.5L, and turn on the cooling water (inlet temperature 22℃).

[0319] BP prediction: Input parameters BP model output prediction granularity .

[0320] PSO optimization: with Substituting the parameters into the fitness function, the optimal parameters are obtained after 50 iterations: , .

[0321] FPID and APA Integration Adjustment:

[0322] Samples were taken at 60 minutes; actual fineness. ,deviation ;

[0323] Original adjustment amount for APA: , ;

[0324] FPID Correction: Output after fuzzy inference , ;

[0325] Total adjustment ( =0.5): , ;

[0326] Updated parameters: n=1700+97.5=1797.5r / min, t=30+21.25=51.25min.

[0327] Standard judgment: If the fineness is 48μm (≤50μm) after 120min, stop grinding and obtain the ground slurry, which is then transferred to the intermediate storage tank.

[0328] S6: Paint Mixing

[0329] Base material compounding: Add 16.45 kg of the second part of the mixed base material to the intermediate storage tank (containing about 80 kg of ground slurry) and stir (700 r / min) for 10 minutes.

[0330] Additives: Add 1.5 kg of dispersant and stir for 15 minutes; add 0.8 kg of defoamer and stir for 15 minutes; add 0.4 kg of leveling agent and stir for 20 minutes; add 1.0 kg of ultraviolet absorber and stir for 20 minutes.

[0331] Solvent adjustment: Add 10 kg xylene + 6 kg butyl acetate (mass ratio 1.67:1), stir for 2.5 hours, and control the temperature at 30℃.

[0332] Viscosity control: The viscosity was tested to be 1300 mPa·s (25℃, rotor No. 4, 50 r / min, within the range of 1000~1500 mPa·s, which meets the standard), and the paint after mixing was obtained.

[0333] S7: Curing process

[0334] Standing curing: Transfer the mixed paint into a curing kettle, seal it, set the temperature to 28℃, and let it stand for 18 hours to cure.

[0335] Intermittent stirring: Stir at 300 rpm for 10 minutes every 6 hours (3 times in total).

[0336] Performance testing: Weather resistance: After 1000 hours of QUV aging, gloss retention rate = 88% (≥80%); Adhesion: Cross-cut test grade = 0 (≤1); Salt spray resistance: No rust or blistering after 5000 hours (meets standards).

[0337] S8: Filter Packaging

[0338] Precision filtration: Filtered through a 100-mesh stainless steel screen at a controlled pressure of 0.4 MPa, the weight of the filtered coating is approximately 110 kg.

[0339] Quantitative filling: The automatic filling machine fills 5 barrels at 20kg / barrel (20.02kg, 19.98kg, 20.00kg, 19.95kg, 20.05kg), with an error of ≤±0.1kg.

[0340] After the packaging drum is sealed, it is stored in a warehouse (temperature 25℃, good ventilation), with a shelf life of 12 months.

[0341] Example results: The batch of coatings was tested and all performance characteristics met the requirements for ultra-weather-resistant and heavy-duty anti-corrosion coatings. It is suitable for coating of steel structures on marine platforms and is expected to have an outdoor service life of ≥15 years.

[0342] In some embodiments, an ultra-weather-resistant industrial heavy-duty anti-corrosion coating includes raw material components formulated according to effective mass parts: base material, pigments and fillers, solvents and additives;

[0343] The base material is a compound system of fluorocarbon resin and epoxy resin. The fluorocarbon resin provides UV aging resistance through the high bond energy CF bond in the molecular structure, while the epoxy resin enhances the coating adhesion through the reaction of epoxy groups with hydroxyl groups on the surface of the metal substrate. The two work together to achieve a balance between the weather resistance and adhesion of the coating.

[0344] The pigments and fillers include nano-titanium dioxide, zinc powder, mica powder, and talc powder. Nano-titanium dioxide (anatase type, particle size 20-50nm) is used to scatter and absorb ultraviolet light and decompose surface contaminants. Zinc powder (purity ≥99.5%, particle size 5-10μm) acts as a sacrificial anode to achieve electrochemical corrosion protection through preferential oxidation. Mica powder (plate structure, particle size 10-20μm) and talc powder (layer structure, particle size 15-30μm) work together to form a "maze effect" to extend the penetration path of corrosive media.

[0345] The solvent is a compound system of xylene and butyl acetate, which is used to dissolve the base material and adjust the evaporation rate of the coating to avoid pinholes or cracks caused by excessive evaporation during coating application.

[0346] The additives include dispersants, defoamers, leveling agents, and ultraviolet absorbers. The dispersants prevent pigment and filler agglomeration through electrostatic repulsion, the defoamers reduce the surface tension of the system to eliminate bubbles, the leveling agents promote uniform spreading of the coating on the substrate surface, and the ultraviolet absorbers (benzotriazoles) specifically absorb 290-400nm ultraviolet light to protect the molecular chains of the base material.

[0347] The core performance indicators of this ultra-weather-resistant industrial heavy-duty anti-corrosion coating meet the following requirements: gloss retention rate ≥80% after 1000h ultraviolet aging test, cross-cut adhesion test grade ≤1, and no rust or blistering after 5000h neutral salt spray test.

[0348] In summary, this invention revolves around three core dimensions: optimized raw material selection, refined process control, and intelligent algorithm-enabled grinding. Through the scientific formulation of the raw material system, precise control of process steps, and optimization of the grinding process by intelligent algorithms, it achieves a dual improvement in the coating's performance of ultra-weather resistance and heavy-duty corrosion protection. Simultaneously, it solves problems such as poor base material synergy and blind process parameters in traditional industrial heavy-duty anti-corrosion coatings. The specific technical logic is as follows:

[0349] 1. Raw material optimization and screening: Abandoning the limitations of single base materials, we select fluorocarbon resin and epoxy resin as the base material, and combine it with functional pigments and fillers such as nano titanium dioxide and zinc powder, as well as special additives. We clarify the specifications, proportions and mechanisms of action of each component, laying the raw material foundation for the weather resistance and corrosion resistance of the coating.

[0350] 2. Refined process control: Pre-treatment steps such as drying and dehydrating pigments and fillers, crushing and sieving, and adjusting the viscosity of the base material are used to eliminate the impact of raw material defects on coating performance. Then, through pre-mixing, paint preparation, curing, and precision filtration and packaging, the parameters (such as temperature, speed, and time) of each step are quantitatively controlled to ensure the stability of the coating system.

[0351] 3. Intelligent Algorithm-Enabled Grinding: In the core grinding and dispersion step, the BP neural network prediction algorithm, particle swarm optimization algorithm, and adaptive parameter adjustment algorithm (some embodiments introduce fuzzy PID algorithm) are integrated to build a closed-loop control system of "real-time prediction - parameter optimization - dynamic adjustment - data feedback", which replaces the traditional manual experience control, ensures uniform and stable slurry fineness, and reduces coating defects from the source.

[0352] How to achieve ultra-weather resistance

[0353] The superior weather resistance primarily targets the ability to resist erosion from natural environments such as ultraviolet aging, temperature fluctuations, and atmospheric oxidation. This invention achieves this function through a two-dimensional approach: raw material compatibility design and refined process control.

[0354] 1. Enhanced weather resistance at the base material level: The base material system is a blend of fluorocarbon resin and epoxy resin. The CF bond energy in the fluorocarbon resin molecule is as high as 485kJ / mol, which has excellent resistance to ultraviolet aging and can effectively resist the degradation of the coating by ultraviolet rays in sunlight. The epoxy resin makes up for the insufficient adhesion of pure fluorocarbon coatings, and at the same time forms a cross-linked network with the fluorocarbon resin, improving the structural stability of the coating and reducing the cracking of the coating caused by temperature changes.

[0355] 2. Synergistic weather-resistant design of pigments, fillers, and additives:

[0356] Adding nano-titanium dioxide (anatase type, particle size 20-50nm) can enhance the scattering and absorption of ultraviolet rays due to its nano-sized particles. At the same time, the photocatalytic activity of anatase can decompose pollutants on the coating surface and reduce the impact of scale buildup on weather resistance.

[0357] Adding benzotriazole UV absorbers can specifically absorb ultraviolet rays in the range of 290–400 nm, further blocking the damage of ultraviolet rays to the molecular chains of the base material.

[0358] Acrylic leveling agents are selected to improve the leveling properties of the coating after application, making the coating surface smooth and flat, and reducing the number of weak points in weather resistance caused by surface defects.

[0359] 3. Process control ensures weather resistance stability:

[0360] The pigment and filler pretreatment process involves vacuum drying (moisture content ≤0.5%) and crushing and sieving (particle size 50~100μm) to avoid pinholes and bubbles after coating curing due to pigment and filler agglomeration or excessive moisture content, and to reduce the risk of ultraviolet rays penetrating from defects.

[0361] The curing treatment is carried out at 25-30℃, with low-speed stirring every 6 hours to allow the components of the coating to fully diffuse and adsorb, forming a stable colloidal structure and ensuring that the coating is not prone to chalking or loss of gloss during long-term outdoor use.

[0362] The intelligent algorithm control of grinding and dispersion ensures that the slurry fineness is ≤50μm and the distribution is uniform. After the coating film is formed, the density is improved, the penetration path of ultraviolet rays and oxygen is reduced, and finally the weather resistance index of gloss retention rate ≥80% is achieved after 1000h aging.

[0363] How to achieve heavy-duty anti-corrosion function

[0364] The heavy-duty corrosion resistance mainly targets the protection against harsh industrial environments such as salt spray, chemical media, and electrochemical corrosion. This invention achieves this through a triple mechanism of electrochemical corrosion prevention, physical barrier protection, and coating structure optimization.

[0365] 1. Electrochemical corrosion protection mechanism: High-purity zinc powder (particle size 5-10μm, purity ≥99.5%) is added as a sacrificial anode. When the coating comes into contact with an electrolyte (such as salt spray), the zinc powder will preferentially undergo an oxidation reaction, protecting the metal substrate from electrochemical corrosion. At the same time, the zinc salt generated by the oxidation of zinc powder can fill the pores of the coating, further enhancing the corrosion protection effect.

[0366] 2. Physical barrier protection design:

[0367] Mica powder (flaky structure, particle size 10-20μm) and talc powder (layered structure, particle size 15-30μm) are selected. The two form an interlaced "maze effect" in the coating, which greatly prolongs the penetration path of corrosive factors such as salt spray and chemical media into the substrate.

[0368] A 100-mesh stainless steel screen (pressure 0.3-0.5MPa) is used for filtration to remove minute impurities from the coating, ensuring that the coating film is free of particle defects and reducing the intrusion channels of corrosive media.

[0369] 3. Coating structure and adhesion optimization:

[0370] The epoxy groups of epoxy resin can react with the hydroxyl groups on the surface of the metal substrate, thereby improving the adhesion of the coating to the metal (adhesion grade ≤ 1) and preventing the coating from losing its anti-corrosion function due to peeling.

[0371] In the base material pretreatment stage, the viscosity of the mixed base material is adjusted to 500-800 mPa·s to ensure that the pigments and fillers are fully combined with the base material. After the paint is mixed, the viscosity of the paint is controlled at 1000-1500 mPa·s to adapt to spraying, brushing and other construction methods and ensure uniform coating thickness.

[0372] The closed-loop algorithm control of grinding and dispersion ensures stable slurry fineness and improves the density of the coating film after it is formed. It can withstand 5000 hours of salt spray erosion without rust or blistering, meeting the stringent requirements of heavy-duty industrial corrosion protection.

[0373] The table below shows the correlation between raw material ratios and performance of ultra-weather-resistant industrial heavy-duty anti-corrosion coatings:

[0374]

[0375] Synergistic effect of raw materials: The base material (fluorocarbon resin + epoxy resin) compound achieves a balance of "weather resistance + adhesion", the pigments and fillers (zinc powder + mica powder) synergistically achieve dual protection of "electrochemical corrosion prevention + physical barrier", and the additive combination (dispersant + leveling agent + ultraviolet absorber) ensures coating uniformity and weather resistance stability.

[0376] Proportioning Standard: The mass fractions in the table represent the preset range for this embodiment. Actual production requires fine-tuning based on the target scenario (e.g., increasing the zinc powder / mica powder ratio for marine engineering, and increasing the fluorocarbon resin / UV absorber ratio for outdoor steel structures).

Claims

1. A method for preparing an ultra-weather-resistant industrial heavy-duty anti-corrosion coating, characterized in that, Includes the following steps: S1: Prepare raw materials, including base material, pigments and fillers, solvents and additives, and weigh each raw material according to the preset mass parts; the base material includes fluorocarbon resin and epoxy resin, the pigments and fillers include nano titanium dioxide, zinc powder, mica powder and talc powder, the solvent includes xylene and butyl acetate, and the additives include dispersant, defoamer, leveling agent and ultraviolet absorber. S2: Pre-treat a symmetrical amount of pigments and fillers to obtain pre-treated pigments and fillers; S3: Mix and adjust the viscosity of the symmetrical amount of base material to obtain a mixed base material, and divide it into a first part of the mixed base material and a second part of the mixed base material; S4: The first part of the mixed base material is premixed with the pretreated pigments and fillers to obtain a premixed slurry, and the initial fineness of the premixed slurry is tested; S5: The premixed slurry is fed into a sand mill, and the grinding process is controlled by a fusion of a BP neural network prediction algorithm, a particle swarm optimization algorithm, and an adaptive parameter adjustment algorithm. The BP neural network prediction algorithm predicts the fineness of the current slurry based on the initial fineness, the current speed of the sand mill, and the cumulative grinding time. The particle swarm optimization algorithm uses the predicted fineness as a constraint and substitutes it into the fitness function to calculate the particle fitness. Within the range of speed 1500–2000 r / min and time 30–180 min, the optimal speed and optimal time are obtained through iterative optimization. The fitness function prioritizes ensuring that the predicted fineness is ≤50 μm and then minimizes the grinding energy consumption. The adjustment amount of the adaptive parameter adjustment algorithm is determined by the following method: based on the deviation between the detected actual fineness and the target fineness of 50 μm, as well as the optimal speed and optimal time. The difference between the actual rotational speed and time is used to calculate the speed adjustment and time adjustment, respectively, so that the actual rotational speed and time approach the optimal speed and time while compensating for the fineness deviation. On this basis, a fuzzy PID algorithm is introduced, and the four are integrated to form a multi-layer closed loop of prediction-optimization-fuzzy decision-dynamic correction-feedback: the BP neural network prediction algorithm provides prediction data for the particle swarm optimization algorithm, the particle swarm optimization algorithm provides target parameters for the adaptive parameter adjustment algorithm and the fuzzy PID algorithm, the fuzzy PID algorithm dynamically adjusts the PID parameters through fuzzy logic to correct the grinding deviation, the adaptive parameter adjustment algorithm realizes the final execution of the parameters based on the output of the fuzzy PID algorithm, and feeds back the actual grinding parameters and the corresponding actual fineness to the BP neural network prediction algorithm to update the model until the fineness of the slurry meets the standard, and the ground slurry is obtained. S6: Add additives, solvents, and the second part of the mixed base material to the ground slurry, and adjust the viscosity to the preset range to obtain the paint after mixing. S7: The paint after mixing is cured and its weather resistance, adhesion and salt spray resistance are tested. After meeting the standards, the cured paint is obtained. S8: After the coating has matured, it is filtered and packaged to obtain the finished coating.

2. The preparation method of the ultra-weather-resistant industrial heavy-duty anti-corrosion coating according to claim 1, characterized in that, In step S2, the pigment and filler pretreatment process is as follows: the weighed pigment and filler are vacuum dried to a moisture content of ≤0.5%, and then pulverized and sieved to a particle size of 50-100μm. The vacuum drying conditions are: temperature 80-100℃, vacuum degree -0.08--0.09MPa, drying for 2-3 hours. If the moisture content exceeds the standard, the drying time is extended by 30 minutes and the test is repeated until the moisture content is ≤0.5%.

3. The preparation method of the ultra-weather-resistant industrial heavy-duty anti-corrosion coating according to claim 1, characterized in that, In step S3, the viscosity of the mixed base material is adjusted as follows: if the viscosity is lower than 500 mPa·s, add 1 to 2 parts of fluorocarbon resin. If the viscosity is higher than 800 mPa·s, add 0.5 to 1 part xylene, stir for 10 minutes and retest until the viscosity meets the standard.

4. The preparation method of the ultra-weather-resistant industrial heavy-duty anti-corrosion coating according to claim 1, characterized in that, In step S4, the pretreated pigments and fillers are added at a rate of 5-8 kg / min. After the addition is completed, the mixture is stirred at 800-1000 r / min for 1-1.5 hours. If there are obvious visible particles, the stirring is extended for 20 minutes and then tested again until the uniformity meets the standard.

5. The preparation method of the ultra-weather-resistant industrial heavy-duty anti-corrosion coating according to claim 1, characterized in that, In step S7, the curing conditions are: temperature 25~30℃, stirring at 300r / min for 10 minutes every 6 hours; weather resistance requirement: gloss retention rate ≥80% after 1000h aging, adhesion requirement: grade ≤1, salt spray resistance requirement: no rust or blistering after 5000h.

6. A super weather-resistant industrial heavy-duty anti-corrosion coating, characterized in that, The super weather-resistant industrial heavy-duty anti-corrosion coating is prepared by the preparation method of any one of claims 1 to 5, wherein the super weather-resistant industrial heavy-duty anti-corrosion coating comprises raw material components formulated according to effective mass parts: base material, pigments and fillers, solvent and additives; The base material is a compound system of fluorocarbon resin and epoxy resin. The fluorocarbon resin provides UV aging resistance through the high bond energy CF bond in the molecular structure, while the epoxy resin enhances the coating adhesion through the reaction of epoxy groups with hydroxyl groups on the surface of the metal substrate. The two work together to achieve a balance between the weather resistance and adhesion of the coating. The pigments and fillers include nano-titanium dioxide, zinc powder, mica powder and talc powder. Nano-titanium dioxide is used to scatter and absorb ultraviolet rays and decompose surface pollutants. Zinc powder acts as a sacrificial anode to achieve electrochemical corrosion protection through preferential oxidation. Mica powder and talc powder work together to form a labyrinth effect to extend the penetration path of corrosive media. The solvent is a compound system of xylene and butyl acetate, which is used to dissolve the base material and adjust the evaporation rate of the coating to avoid pinholes or cracks caused by excessive evaporation during coating application. The additives include dispersants, defoamers, leveling agents, and ultraviolet absorbers. The dispersants prevent pigment and filler agglomeration through electrostatic repulsion, the defoamers reduce the surface tension of the system to eliminate bubbles, the leveling agents promote uniform spreading of the coating on the substrate surface, and the ultraviolet absorbers specifically absorb 290-400nm ultraviolet light to protect the molecular chains of the base material.

Images