A method for rust removal and anti-corrosion construction on steel structure surface

Abstract

The invention provides a construction method for rust removal and corrosion prevention of steel structure surface, which belongs to the technical field of rust removal and corrosion prevention of steel structure, and comprises: using a drone to obtain multi-angle surface images of the steel structure, which are recorded as the first surface image. Then the steel structure is sprayed to obtain the second surface image after spraying. The two surface images are pre-processed and aligned, and the same standard is used for grid division, and the image features of each grid are calculated, which are recorded as the first segmentation feature and the second segmentation feature respectively. The rust removal measure parameter vector of each grid is calculated by using a pre-trained model, and a targeted rust removal operation is performed according to the vector. The image after rust removal is compared with the rust-free standard to determine whether the rust removal is completed. After the rust removal is completed, according to the preset spraying requirements, the primer, the intermediate paint and the topcoat are sprayed in sequence. Finally, the surface of the steel structure after painting is inspected, and the coating thickness is measured to ensure that the anti-corrosion effect meets the requirements. Targeted rust removal and corrosion prevention are achieved.

Description

Technical Field

[0001] The invention belongs to the technical field of rust removal and anti-corrosion of steel structures, and in particular relates to a construction method for rust removal and anti-corrosion of the surface of a steel structure. Background Art

[0002] Steel structures are widely used materials in modern construction and engineering fields, and their anti-corrosion performance is crucial to the safety and service life of the structures. However, steel is prone to surface corrosion during long-term use, which not only affects its appearance, but also seriously damages the mechanical properties of the steel structure. Therefore, how to effectively remove rust and prevent corrosion on the surface of steel structures has always been the focus of the industry.

[0003] Existing methods for rust removal and corrosion prevention of steel structures mainly include sandblasting, grinding and chemical rust removal. Sandblasting uses high-speed sand particles to impact the surface of steel to remove the rust layer; grinding rust removal uses electric sand grinders and other tools to grind the steel surface manually or mechanically; chemical rust removal uses chemical solvents such as acid washing and alkali washing to dissolve and peel off the rust layer on the steel surface. Although these methods can effectively remove rust on the surface of steel, there are still some problems in practical applications:

[0004] 1. The rust removal effect cannot be precisely controlled. The degree of rust removal by different methods varies, and it is difficult to dynamically adjust the best rust removal measures according to the specific rust conditions on the steel surface.

[0005] 2. The rust removal process lacks intelligent control. The specific operating parameters of rust removal (such as sandblasting pressure, grinding intensity, chemical ratio, etc.) are usually set empirically based on manual experience, which makes it difficult to optimize operating efficiency.

[0006] 3. Lack of targeted anti-corrosion coating design. The surface conditions of steel after rust removal are different. How to choose the most suitable anti-corrosion coating and its construction process to ensure the best anti-corrosion effect is also a big problem.

[0007] In other words, the existing technology has the technical problem that it is difficult to carry out targeted rust removal and corrosion prevention on different parts of the steel structure. Summary of the invention

[0008] In view of this, the present invention provides a method for rust removal and anti-corrosion construction on the surface of a steel structure, which can solve the technical problem in the prior art that it is difficult to carry out targeted rust removal and anti-corrosion on different parts of the steel structure.

[0009] The present invention is achieved in that:

[0010] The present invention provides a method for rust removal and anti-corrosion construction on the surface of a steel structure, which comprises the following steps:

[0011] S10, using a camera drone to obtain surface images of the steel structure to be subjected to rust removal and anti-corrosion construction at multiple angles, recorded as first surface images;

[0012] S20, using a spray device carried by the drone to spray the surface of the steel structure, and obtaining surface images of the steel structure at multiple angles after the spray treatment, which are recorded as second surface images;

[0013] S30, preprocessing the first surface image and the second surface image to obtain a first preprocessed image and a second preprocessed image, and aligning the first preprocessed image and the second preprocessed image at different angles;

[0014] S40, meshing the aligned first preprocessed image and the second preprocessed image using the same standard, and calculating the segmented image features of each grid, including the grid segmented image features of the first preprocessed image and the grid segmented image features of the second preprocessed image, which are respectively recorded as the first segmentation features and the second segmentation features;

[0015] S50, using the pre-trained rust removal measure determination model, calculating the segmented image features of each grid, and obtaining an array of rust removal measures for each grid, specifically, a parameter vector for each step of the entire rust removal process,

[0016] S60, performing rust removal operation manually or using rust removal equipment according to the rust removal measure vector of each grid, and determining whether the rust removal operation is completed by calculating the difference between the image after rust removal and the preset standard image without rust, that is, if the difference is greater than or equal to the preset difference threshold, then continue the rust removal operation, if the difference is less than the preset difference threshold, then complete the rust removal operation;

[0017] S70, after the rust removal operation is completed, spraying the steel structure with anti-corrosion paint, specifically applying primer, intermediate paint and topcoat in sequence according to preset spraying requirements; the spraying requirements are the thickness of each paint layer and the curing time;

[0018] S80. Inspect the surface of the steel structure after painting, measure the coating thickness, and ensure that the anti-corrosion effect meets the requirements.

[0019] The method for determining the thickness and curing time of each paint layer is specifically as follows:

[0020] S71, establishing a multi-objective game model, including a relational model and a correction model, wherein the relational model is used to describe the relationship between the spraying parameters of each layer of paint and the thickness and curing time of the paint layer, and the correction model is used to describe the influence of environmental conditions on the curing time. The decision variables of the multi-objective game model are the thickness and curing time of the primer, the intermediate paint and the topcoat as decision variables, and the objective function of the multi-objective game model is the anti-corrosion effect and the construction time;

[0021] S72, fitting the relationship model and using historical data;

[0022] S73, constructing the constraint conditions of the multi-objective game model, including the minimum and maximum thickness requirements of each layer of paint, the total thickness limit, the thickness ratio relationship between each layer, and the shortest and longest curing time;

[0023] S74, using a multi-objective optimization algorithm to solve the multi-objective game model, and obtaining a Pareto optimal solution set as the spraying requirement to be selected;

[0024] S75. Use the small-scale test verification method to test the selected spraying requirements, and select the spraying requirements corresponding to the best small-scale test results as the final spraying requirements.

[0025] Furthermore, the step S75 specifically includes:

[0026] S751. Select several small test areas in non-critical parts of the steel structure;

[0027] S752. Carry out spraying tests in various test areas according to different solutions in the Pareto optimal solution set;

[0028] S753, record the actual spraying parameters, environmental conditions, paint layer thickness and curing time of each test area;

[0029] S754, conduct corrosion resistance evaluation on the test area, including adhesion test and salt spray resistance test;

[0030] S755. Comprehensively consider the anti-corrosion effect, construction efficiency and cost factors, and select the optimal spraying requirements from the test results.

[0031] Specifically, the step S10 includes: using a camera drone to obtain surface images of the steel structure to be derusted and anti-corroded at multiple angles, which are recorded as first surface images.

[0032] Specifically, the step S20 includes: using a spray device carried by the drone to uniformly spray the surface of the steel structure; obtaining surface images of the steel structure at multiple angles after the spray treatment, which are recorded as second surface images.

[0033] Specifically, step S30 includes: preprocessing the first surface image and the second surface image, including image denoising, brightness / contrast adjustment, geometric transformation, etc., to obtain a first preprocessed image and a second preprocessed image; and registering and aligning the first preprocessed image and the second preprocessed image at different angles to ensure that data in the same area can be accurately matched.

[0034] Specifically, step S40 includes: dividing the aligned image using a grid of a fixed size; for each grid area, calculating visual features such as texture, color, and edge to form a first segmentation feature and a second segmentation feature that describe the surface state of the grid area.

[0035] Specifically, step S50 includes: establishing a rust removal measure determination model based on deep learning, the training data of the model including a large number of steel structure surface images in different rust states and manual annotations of their optimal rust removal methods and parameters; taking the first segmentation feature and the second segmentation feature as input, the model outputs the rust removal method (such as sandblasting, grinding, chemical rust removal) and specific operating parameters (such as pressure, time, etc.) required for each grid area, forming a rust removal measure vector for each grid.

[0036] Specifically, step S60 includes: according to the rust removal measure vector of each grid, using corresponding manual operation or automated rust removal equipment to perform rust removal; collecting the surface image of the steel structure after rust removal, and calculating the pixel-level difference with the preset rust-free standard image, if the difference is greater than or equal to 0.1, continuing the rust removal operation until the difference is less than 0.1.

[0037] Specifically, the step S70 includes: selecting epoxy or polyurethane primers according to the selection criteria of primers (good adhesion, excellent anti-corrosion performance, compatibility with substrates and subsequent coatings); selecting epoxy intermediate paints according to the selection criteria of intermediate paints (enhancing overall coating thickness, improving anti-corrosion performance, and improving topcoat adhesion); selecting polyurethane or fluorocarbon topcoats according to the selection criteria of topcoats (good weather resistance, decorativeness, and UV resistance); and determining the spraying thickness and curing time of each layer of coating according to the optimization model described below.

[0038] Specifically, step S71 includes: establishing a relational model to describe the mathematical relationship between the spraying parameters of each layer of paint and the paint layer thickness and curing time, and establishing a correction model to describe the influence of environmental conditions on the curing time; taking the thickness and curing time of each layer of paint as decision variables, establishing two objective functions of anti-corrosion effect and construction time, and setting relevant constraints at the same time, forming a multi-objective game optimization problem.

[0039] Specifically, step S72 includes: using historical data to compare the parameters in the relationship model , , , , , etc. to perform regression fitting to improve the prediction accuracy of the model; at the same time, the parameters in the correction model are , , , , etc. for model training and optimization.

[0040] Specifically, step S73 includes: setting the minimum and maximum thickness of each layer of paint and , the total thickness cannot exceed ;Thickness ratio of primer to midcoat Cannot be less than ,Thickness ratio of intermediate paint to topcoat Cannot be less than ;Curing time for each coat of paint Need to be controlled arrive Within the range; set reasonable value range constraints for spraying parameters (pressure, speed, distance).

[0041] The multi-objective game model can be expressed as:

[0042] ;

[0043] in: ;

[0044] represents the decision variable vector, including the thickness of each layer of paint, curing time, spraying pressure, spraying speed and spraying distance. is the thickness of the i-th layer of paint (i = 1, 2, 3 represents primer, intermediate paint, and topcoat, respectively); is the curing time of the i-th layer of paint; is the spraying pressure of the i-th layer of paint; Spraying speed of the i-th layer of paint; is the spraying distance of the i-th layer of paint.

[0045] is the anti-corrosion effect objective function:

[0046] ;

[0047] In the formula, is the weight of the thickness of the i-th paint layer; is the maximum allowable thickness of the i-th layer of paint; is the total thickness influencing factor; is the upper limit of total thickness; is the optimal thickness deviation penalty factor; is the ideal optimal thickness of the i-th layer of paint.

[0048] is the construction time objective function:

[0049] ;

[0050] In the formula, : Actual curing time after taking into account environmental conditions; : Preparation time for the i-th coat of paint; : Overall setup time; : Ambient temperature; : Temperature change; actual curing time Calculated by the modified model.

[0051] The relational model is specifically expressed as follows:

[0052] ;

[0053] In the formula, is the ambient temperature, , , , , , All are fitting parameters, obtained through historical data regression;

[0054] The modified model is specifically expressed as:

[0055] ;

[0056] In the formula, : Temperature cycle change influence coefficient; : Frequency of temperature cycle change; : Phase of temperature cycle change; : Humidity influence coefficient; : Current humidity; : Base humidity; : Wind speed influence coefficient; : wind speed;

[0057] Theoretical curing time calculation:

[0058] ;

[0059] In the formula, : Curing time coefficient; : Humidity influence factor; : activation energy; Gas constant; , : Wind speed affects parameters.

[0060] The constraints include:

[0061] 1. Thickness limit constraints:

[0062] ;

[0063] In the formula, : Minimum thickness of the i-th layer of paint; : Maximum thickness of the i-th layer of paint; : Upper limit of total thickness.

[0064] 2. Thickness ratio constraint:

[0065] ;

[0066] In the formula, : Minimum thickness ratio of primer to intermediate paint; : Minimum thickness ratio of intermediate paint to topcoat.

[0067] 3. Curing time constraints:

[0068] ;

[0069] In the formula, : The shortest curing time of the i-th coat of paint; : Maximum curing time of the i-th coat of paint.

[0070] 4. Spraying parameter constraints:

[0071] ;

[0072] ;

[0073] ;

[0074] In the formula, , : Spraying pressure range; , : Spraying speed range; , : Spraying distance range.

[0075] This multi-objective game model comprehensively considers the two objectives of anti-corrosion effect and construction time, and also includes the influence of various process parameters and environmental conditions. The solution of the model will obtain a series of Pareto optimal solutions, representing different trade-offs between anti-corrosion effect and construction time.

[0076] In practical applications, we can select a suitable multi-objective optimization algorithm (such as NSGA-II or MOEA / D) to solve the model according to the specific situation, and select the final construction plan based on the results of the pilot test.

[0077] Specifically, step S74 includes: using NSGA-II or MOEA / D algorithm based on evolutionary algorithm, solving the two objective functions of anti-corrosion effect and construction time through iterative optimization, and obtaining a set of Pareto optimal solutions representing different degrees of trade-offs as the spraying requirements to be selected.

[0078] Specifically, step S75 includes: selecting several small test areas at non-critical parts of the steel structure, conducting spraying tests in each test area according to different schemes in the Pareto optimal solution set, and recording actual spraying parameters, environmental conditions, paint layer thickness and curing time; evaluating the anti-corrosion properties such as adhesion and salt spray resistance of the test area; comprehensively considering the anti-corrosion effect, construction efficiency and cost factors, and selecting the optimal spraying process parameters from the test results as the final spraying requirements.

[0079] Furthermore, the small test area is an area of ​​30 cm x 30 cm.

[0080] Wherein, the spraying parameters include spraying pressure, spraying distance and spraying speed.

[0081] Wherein, the environmental conditions include temperature, humidity and wind speed.

[0082] Among them, the parameter vector includes: rust removal method, operating parameters and operation time. The rust removal method includes sandblasting, grinding or chemical rust removal. Each rust removal method has different operating parameters. The operating parameters of sandblasting include sandblasting pressure, sandblasting distance, and sandblasting time; the operating parameters of grinding include grinding force and grinding time; the operating parameters of chemical rust removal include chemical rust remover ratio concentration and action time.

[0083] The difference threshold is 0.1.

[0084] Furthermore, the selection criteria for primers are: good adhesion, excellent corrosion protection, and compatibility with the substrate and subsequent coatings;

[0085] The selection criteria for the intermediate paint are: enhancing the overall coating thickness, improving the anti-corrosion performance, and improving the adhesion of the topcoat;

[0086] The selection criteria for topcoat are: good weather resistance, decorative properties and UV resistance.

[0087] Primers are usually epoxy or polyurethane primers; intermediate paints are generally epoxy intermediate paints; topcoats are usually polyurethane or fluorocarbon topcoats.

[0088] Wherein, the rust removal measures are determined by a model using a convolutional neural network structure.

[0089] The training steps of the rust removal measures determination model include:

[0090] Create a training data set, specifically:

[0091] 1. Collect image samples of various steel structure surface corrosion conditions, including different degrees of corrosion, corrosion in different environments, etc.

[0092] 2. Manually label each sample, including the degree of rust, surface condition, etc.

[0093] 3. Apply different rust removal methods to each sample and record the rust removal effect and required parameters.

[0094] 4. The image samples, rust status annotations, rust removal methods and effects are combined into training data pairs.

[0095] 5. Preprocess the data set, including image enhancement, data normalization, etc.

[0096] Train the neural network, specifically:

[0097] 1. Design the convolutional neural network (CNN) structure, including convolutional layers, pooling layers, fully connected layers, etc.

[0098] 2. Define the loss function and use a multi-task learning approach to simultaneously predict the rust removal method and operating parameters.

[0099] 3. Use optimization algorithms such as stochastic gradient descent (SGD) for model training.

[0100] 4. Use cross-validation method to evaluate model performance and perform hyperparameter tuning.

[0101] 5. Use the test set to verify the generalization ability of the model and ensure the applicability of the model under different corrosion conditions.

[0102] Compared with the prior art, the method for rust removal and anti-corrosion construction of a steel structure provided by the present invention has the following beneficial effects:

[0103] 1. It can accurately assess the actual corrosion situation on the surface of steel structures. By collecting multi-angle images through drones and analyzing image features with deep learning technology, the degree of corrosion in different areas can be accurately identified, providing a reliable basis for subsequent targeted rust removal. Compared with manual visual inspection, this method based on image analysis is more objective and reliable.

[0104] 2. Realize intelligent control of rust removal operations. According to the rust characteristics of different areas, the model is determined by using pre-trained rust removal measures to automatically recommend the best rust removal method (sandblasting, grinding or chemical rust removal) and its specific parameters, which greatly improves the efficiency and quality of rust removal operations. This intelligent control based on model optimization is far superior to the traditional method of setting parameters based on manual experience.

[0105] 3. An anti-corrosion coating process integrating multi-objective optimization is designed. After rust removal, the present invention establishes a multi-objective game model that comprehensively considers the anti-corrosion effect and construction time. Through mathematical optimization and small-scale experiments, the optimal spraying parameters of primer, intermediate paint and topcoat can be determined to ensure the dual optimization of coating effect and construction efficiency. This systematic anti-corrosion coating design is more scientific and reasonable than the traditional method that relies solely on experience.

[0106] In general, the intelligent steel structure surface rust removal and anti-corrosion construction method based on drones, image analysis and multi-objective optimization proposed in the present invention solves the technical problem that it is difficult to carry out targeted rust removal and anti-corrosion on different parts of the steel structure in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0107] Figure 1 A flow chart of the method provided by the present invention;

[0108] Figure 2 A flow chart of the method for determining the thickness and curing time of each paint layer;

[0109] Figure 3 This is a flowchart of step S75. DETAILED DESCRIPTION

[0110] In order to make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0111] like Figure 1 As shown, it is a flow chart of a method for rust removal and anti-corrosion construction of a steel structure surface provided by the present invention, and the method comprises the following steps:

[0112] S10, using a camera drone to obtain surface images of the steel structure to be subjected to rust removal and anti-corrosion construction at multiple angles, recorded as first surface images;

[0113] S20, using a spray device carried by the drone to spray the surface of the steel structure, and obtaining surface images of the steel structure at multiple angles after the spray treatment, which are recorded as second surface images;

[0114] S30, preprocessing the first surface image and the second surface image to obtain a first preprocessed image and a second preprocessed image, and aligning the first preprocessed image and the second preprocessed image at different angles;

[0115] S40, meshing the aligned first preprocessed image and the second preprocessed image using the same standard, and calculating the segmented image features of each grid, including the grid segmented image features of the first preprocessed image and the grid segmented image features of the second preprocessed image, which are respectively recorded as the first segmentation features and the second segmentation features;

[0116] S50, using the pre-trained rust removal measure determination model, calculating the segmented image features of each grid, and obtaining an array of rust removal measures for each grid, specifically, a parameter vector for each step of the entire rust removal process,

[0117] S60, performing rust removal operation manually or using rust removal equipment according to the rust removal measure vector of each grid, and determining whether the rust removal operation is completed by calculating the difference between the image after rust removal and the preset standard image without rust, that is, if the difference is greater than or equal to the preset difference threshold, then continue the rust removal operation, if the difference is less than the preset difference threshold, then complete the rust removal operation;

[0118] S70, after the rust removal operation is completed, the steel structure is sprayed with anti-corrosion paint, and primer, intermediate paint and topcoat are applied in sequence according to the preset spraying requirements; the spraying requirements are the thickness of each paint layer and the curing time;

[0119] S80. Inspect the surface of the steel structure after painting, measure the coating thickness, and ensure that the anti-corrosion effect meets the requirements.

[0120] The spraying parameters include spraying pressure, spraying distance, and spraying speed; the environmental conditions include temperature, humidity, and wind speed;

[0121] The parameter vector includes: rust removal method, operating parameters and operation time. The rust removal method includes sandblasting, grinding or chemical rust removal. Each rust removal method has different operating parameters. The operating parameters of sandblasting include sandblasting pressure, sandblasting distance and sandblasting time; the operating parameters of grinding include grinding intensity and grinding time; the operating parameters of chemical rust removal include chemical rust remover ratio concentration and action time.

[0122] The difference threshold is 0.1.

[0123] Among them, the selection criteria of primers are: good adhesion, excellent anti-corrosion performance, and compatibility with the substrate and subsequent coatings; epoxy or polyurethane primers are usually selected;

[0124] The selection criteria for intermediate paint are: enhancing the overall coating thickness, improving anti-corrosion performance, and improving the adhesion of the topcoat; generally, epoxy intermediate paint is selected.

[0125] The selection criteria for topcoat are: good weather resistance, decorative properties and UV resistance; polyurethane or fluorocarbon topcoat are commonly used.

[0126] like Figure 2 As shown, the method for determining the thickness of each paint layer and the curing time is as follows:

[0127] S71, establishing a multi-objective game model, including a relational model and a correction model, wherein the relational model is used to describe the relationship between the spraying parameters of each layer of paint and the thickness and curing time of the paint layer, and the correction model is used to describe the influence of environmental conditions on the curing time. The decision variables of the multi-objective game model are the thickness and curing time of the primer, the intermediate paint and the topcoat as decision variables, and the objective function of the multi-objective game model is the anti-corrosion effect and the construction time;

[0128] S72, fitting the relationship model and using historical data;

[0129] S73, constructing the constraint conditions of the multi-objective game model, including the minimum and maximum thickness requirements of each layer of paint, the total thickness limit, the thickness ratio relationship between each layer, and the shortest and longest curing time;

[0130] S74, using a multi-objective optimization algorithm to solve the multi-objective game model, and obtaining a Pareto optimal solution set as the spraying requirement to be selected;

[0131] S75. Use the small-scale test verification method to test the selected spraying requirements, and select the spraying requirements corresponding to the best small-scale test results as the final spraying requirements.

[0132] like Figure 3 As shown, the step S75 specifically includes:

[0133] S751. Select several small test areas (e.g. 30cm x 30cm) at non-critical locations of the steel structure.

[0134] S752. According to different schemes in the Pareto optimal solution set, spraying tests are carried out in various test areas.

[0135] S753. Record the actual spraying parameters, environmental conditions, paint layer thickness and curing time for each test area.

[0136] S754. Evaluate the anti-corrosion performance of the test area, including adhesion test, salt spray resistance test, etc.

[0137] S755. Comprehensively consider the anti-corrosion effect, construction efficiency and cost factors, and select the optimal spraying requirements from the test results.

[0138] The specific implementation methods of the above steps are described in detail below:

[0139] The specific implementation of step S10 is: using a camera drone to obtain surface images of the steel structure to be derusted and anticorroded from multiple angles. The purpose of this step is to collect initial image data of the steel structure surface to provide a basis for subsequent image analysis and processing. The drone is equipped with a high-definition camera, and by taking all-round photos of the steel structure surface, multiple groups of surface images covering different angles and positions are obtained. These original images are recorded as the first surface images.

[0140] The specific implementation of step S20 is: using the spray device carried by the drone to spray the surface of the steel structure, and obtaining the surface images of the steel structure at multiple angles after the spray treatment. The purpose of this step is to collect the surface image data of the steel structure after the spray treatment, and provide a basis for subsequent image comparison and analysis. The spray device carried by the drone can evenly spray the surface of the steel structure, and the surface image of the steel structure after the spray treatment is recorded as the second surface image.

[0141] The specific implementation of step S30 is: preprocessing the first surface image and the second surface image to obtain the first preprocessed image and the second preprocessed image, and aligning the first preprocessed image and the second preprocessed image at different angles. The purpose of this step is to eliminate noise and distortion in the image acquisition process and ensure that the two sets of image data are completely corresponding in spatial position. The preprocessing operation includes image denoising, brightness / contrast adjustment, geometric transformation, etc., to ensure that subsequent image feature extraction and comparative analysis are more reliable. At the same time, through image registration technology, the first preprocessed image and the second preprocessed image at different angles are aligned to ensure that the data in the same area can be accurately matched.

[0142] The specific implementation of step S40 is: the aligned first preprocessed image and the second preprocessed image are meshed using the same standard, and the segmented image features of each mesh are calculated, including the mesh segmented image features of the first preprocessed image and the mesh segmented image features of the second preprocessed image. The purpose of this step is to extract image features that describe the surface state of the steel structure and provide a data basis for the subsequent determination of rust removal measures. First, the aligned image is divided using a grid of a fixed size to ensure that the image area in each grid is the same. Then, for each grid area, visual features such as texture, color, and edge are calculated to form a feature vector that describes the surface state of the grid area. The feature vector corresponding to the first preprocessed image is recorded as the first segmentation feature, and the feature vector corresponding to the second preprocessed image is recorded as the second segmentation feature.

[0143] The specific implementation method of step S50 is: using the pre-trained rust removal measure determination model, the segmented image features of each grid are calculated to obtain the rust removal measure array of each grid. The purpose of this step is to determine the best rust removal measures and their parameters for each area according to the actual situation of the steel structure surface. First, it is necessary to establish a rust removal measure determination model based on deep learning. The training data of the model includes a large number of steel structure surface images in different rust states, as well as the best rust removal methods and parameters for each image area manually annotated. During the training process, the model learns the mapping relationship from image features to rust removal measures. In practical applications, the first segmentation feature and the second segmentation feature are used as input, and the model outputs the rust removal method (such as sandblasting, grinding, chemical rust removal) and specific operating parameters (such as pressure, time, etc.) required for each grid area. This information constitutes the rust removal measure vector for each grid.

[0144] The specific implementation method of step S60 is: according to the rust removal measure vector of each grid, rust removal operation is performed manually or by using rust removal equipment, and the method of calculating the difference between the image after rust removal and the preset rust-free standard image is used to determine whether the rust removal operation is completed. The purpose of this step is to perform actual rust removal on the surface of the steel structure according to the rust removal measures determined above, and to evaluate the rust removal effect by comparing with the ideal state. First, according to the rust removal measure vector of each grid, rust removal is performed by using the corresponding manual operation or automatic rust removal equipment. Then, the surface image of the steel structure after rust removal is collected and compared and analyzed with the preset rust-free standard image. Specifically, the pixel-level difference between the two images is calculated to determine the degree of consistency between the current surface state and the ideal state. If the difference is greater than or equal to the preset difference threshold of 0.1, it indicates that the rust removal operation needs to be continued; otherwise, it means that the rust removal goal has been completed.

[0145] The specific implementation method of step S70 is: after the rust removal operation is completed, the anti-corrosion coating is sprayed on the steel structure, and the primer, intermediate paint and topcoat are applied in sequence according to the preset spraying requirements. The purpose of this step is to apply anti-corrosion coating on the surface of the steel structure after rust removal to improve its corrosion resistance. First, according to the selection criteria of the primer (good adhesion, excellent anti-corrosion performance, compatibility with the substrate and subsequent coatings), epoxy or polyurethane primers are selected. Then, according to the selection criteria of the intermediate paint (enhancing the overall coating thickness, improving the anti-corrosion performance, and improving the adhesion of the topcoat), epoxy intermediate paint is selected. Finally, according to the selection criteria of the topcoat (good weather resistance, decorativeness and UV resistance), polyurethane or fluorocarbon topcoat is selected. The spraying thickness and curing time of each layer of coating need to be determined according to the optimization model described later.

[0146] The specific implementation method of step S71 is: establish a multi-objective game model, including a relational model and a correction model. The purpose of this step is to establish a multi-objective optimization model that comprehensively considers the anti-corrosion effect and construction efficiency, and provide a theoretical basis for determining the optimal spraying parameters. The relational model is used to describe the mathematical relationship between the spraying parameters (pressure, speed, distance) of each layer of paint and the thickness and curing time of the paint layer, and the correction model describes the influence of environmental conditions (temperature, humidity, wind speed) on the curing time. The thickness and curing time of each layer of paint are used as decision variables, and two objective functions of anti-corrosion effect and construction time are established. At the same time, relevant constraints (such as thickness range, curing time limit, etc.) are set to constitute a multi-objective game optimization problem.

[0147] The specific implementation of step S72 is: fitting the relationship model and using historical data. The purpose of this step is to estimate and calibrate the parameters in the relationship model based on the existing measured data using a suitable regression analysis method to improve the prediction accuracy of the model. , , , , , etc., all need to be determined through regression fitting of historical data. At the same time, the parameters in the model are corrected. , , , , Etc., it is also necessary to use relevant environmental condition data for model training and optimization.

[0148] The specific implementation of step S73 is: construct the constraints of the multi-objective game model, including the minimum and maximum thickness requirements of each layer of paint, the total thickness limit, the thickness ratio between each layer, and the shortest and longest curing time. The purpose of this step is to set various constraints that need to be met in the multi-objective optimization model according to actual production requirements and process specifications. For example, the minimum and maximum thickness of each layer of paint are and , the total thickness cannot exceed ;Thickness ratio of primer to midcoat Cannot be less than ,Thickness ratio of intermediate paint to topcoat Cannot be less than ;Curing time for each coat of paint Need to be controlled arrive At the same time, it is also necessary to set reasonable value range constraints for the spraying parameters (pressure, speed, distance).

[0149] The multi-objective game model can be expressed as:

[0150] ;

[0151] in: ;

[0152] represents the decision variable vector, including the thickness of each layer of paint, curing time, spraying pressure, spraying speed and spraying distance. is the thickness of the i-th layer of paint (i = 1, 2, 3 represents primer, intermediate paint, and topcoat, respectively); is the curing time of the i-th layer of paint; is the spraying pressure of the i-th layer of paint; is the spraying speed of the i-th layer of paint; is the spraying distance of the i-th layer of paint.

[0153] is the anti-corrosion effect objective function:

[0154] ;

[0155] In the formula, is the weight of the thickness of the i-th paint layer; is the maximum allowable thickness of the i-th layer of paint; is the total thickness influencing factor; is the upper limit of total thickness; is the optimal thickness deviation penalty factor; is the ideal optimal thickness of the i-th layer of paint.

[0156] is the construction time objective function:

[0157] ;

[0158] In the formula, : Actual curing time after taking into account environmental conditions; : Preparation time for the i-th coat of paint; : Overall setup time; : Ambient temperature; : Temperature change; actual curing time Calculated by the modified model.

[0159] The relational model is specifically expressed as follows:

[0160] ;

[0161] In the formula, is the ambient temperature, , , , , , All are fitting parameters, obtained through historical data regression;

[0162] The modified model is specifically expressed as:

[0163] ;

[0164] In the formula, : Temperature cycle change influence coefficient; : Frequency of temperature cycle change; : Phase of temperature cycle change; : Humidity influence coefficient; : Current humidity; : Base humidity; : Wind speed influence coefficient; : wind speed;

[0165] Theoretical curing time calculation:

[0166] ;

[0167] In the formula, : Curing time coefficient; : Humidity influence factor; : activation energy; : gas constant; , : Wind speed affects parameters.

[0168] The constraints include:

[0169] 1. Thickness limit constraints:

[0170] ;

[0171] In the formula, : Minimum thickness of the i-th layer of paint; : Maximum thickness of the i-th layer of paint; : Upper limit of total thickness.

[0172] 2. Thickness ratio constraint:

[0173] ;

[0174] In the formula, : Minimum thickness ratio of primer to intermediate paint; : Minimum thickness ratio of intermediate paint to topcoat.

[0175] 3. Curing time constraints:

[0176] ;

[0177] In the formula, : The shortest curing time of the i-th coat of paint; : Maximum curing time of the i-th coat of paint.

[0178] 4. Spraying parameter constraints:

[0179] ;

[0180] ;

[0181] ;

[0182] In the formula, , : Spraying pressure range; , : Spraying speed range; , : Spraying distance range.

[0183] The specific implementation method of step S74 is: using a multi-objective optimization algorithm to solve the multi-objective game model, and obtain a Pareto optimal solution set as the spraying requirement to be selected. The purpose of this step is to use an appropriate multi-objective optimization algorithm to solve the multi-objective game model established above, and obtain a set of Pareto optimal solutions that represent the trade-off between anti-corrosion effect and construction time. The NSGA-II or MOEA / D algorithm based on the evolutionary algorithm can be used to solve the multi-objective function through iterative optimization, and finally obtain a set of alternative solutions that weigh anti-corrosion and construction time to varying degrees. Each of these alternative solutions represents an uncontrollable optimal solution, which constitutes the Pareto frontier and provides a reference for subsequent small-scale test verification.

[0184] The specific implementation method of step S75 is: adopt a small test verification method to test the spraying requirements to be selected, and select the spraying requirements corresponding to the best small test results as the final spraying requirements. The purpose of this step is to verify the alternative solutions obtained by the above optimization in the actual production environment and select the best spraying process parameters. First, select several small test areas at non-critical parts of the steel structure, and conduct spraying tests in each test area according to different solutions in the Pareto optimal solution set, and record the actual spraying parameters, environmental conditions, paint layer thickness and curing time. Then, the test area is evaluated for anti-corrosion properties such as adhesion and salt spray resistance. Finally, considering the anti-corrosion effect, construction efficiency and cost factors, the optimal spraying process parameters are selected from the test results as the final spraying requirements.

[0185] The core idea of ​​the whole invention can be summarized as follows: using drones to collect multi-angle images of the steel structure surface, and through image preprocessing, feature extraction and other steps, a rust removal measure determination model based on deep learning is established. The model can determine the best rust removal method and parameters for each area according to the actual situation of the steel structure surface. Then, the steel structure is actually rust-removed according to the determined rust removal measures, and the rust removal effect is evaluated by comparing with the ideal state. Finally, based on the rust removal, a multi-objective game model is established, which comprehensively considers the anti-corrosion effect and construction efficiency, and determines the best anti-corrosion coating process parameters through optimization solution combined with small test verification. This method that integrates image analysis, deep learning, and multi-objective optimization can effectively improve the effect and efficiency of rust removal and anti-corrosion coating on the surface of steel structures.

[0186] Another specific implementation of the multi-objective game model and its constraints in steps S72-S73 is given below:

[0187] 1. Relationship model (describing the relationship between spraying parameters and paint layer thickness and curing time):

[0188] For each coat of paint (primer, intermediate paint, topcoat), paint layer thickness The calculation formula is:

[0189] ;

[0190] In the formula, For the Thickness of the paint layer (μm); is the spraying pressure (MPa); is the spraying speed (m / s); is the spraying distance (m); is the ambient temperature (K); is the fitting parameter, obtained through historical data regression;

[0191] Curing time The calculation formula is:

[0192] ;

[0193] In the formula, For the Curing time of one coat of paint (h); is the relative humidity (%); is the activation energy (J / mol); is the bulk constant (8.314 J / (mol·K)); is the wind speed (m / s); is the fitting parameter, obtained through historical data regression;

[0194] 2. Modified model (describes the effect of environmental conditions on curing time):

[0195] ;

[0196] In the formula, is the curing time after environmental correction (h); is the base relative humidity (%); To correct the parameters, they are obtained by fitting the experimental data;

[0197] 3. Constraints:

[0198] a) Thickness limit of each layer of paint:

[0199] ;

[0200] b)Total thickness limit:

[0201] ;

[0202] c) Thickness ratio between layers:

[0203] ;

[0204] d) Curing time limit:

[0205] ;

[0206] 4. Objective function:

[0207] a) Anticorrosion effect objective function:

[0208] ;

[0209] In the formula, is the weight coefficient of each layer of paint; The optimum thickness of each layer of paint; To adjust the parameters;

[0210] b) Construction time objective function:

[0211] ;

[0212] In the formula, Preparation time for each coat of paint; Time for equipment installation and commissioning; To predict temperature changes;

[0213] 5. The final multi-objective game model is expressed as:

[0214] ;

[0215] Including the above constraints

[0216] This multi-objective optimization problem can be solved using multi-objective optimization algorithms such as NSGA-II or MOEA / D to obtain the Pareto optimal solution set.

[0217] Specifically, the principle of the present invention is: to collect multi-angle image data of the steel surface through drones, combine deep learning and multi-objective optimization technology, and realize intelligent control of rust removal operations and anti-corrosion coating.

[0218] First, the high-definition camera carried by the drone is used to take all-round photos of the steel structure surface to obtain original image data from multiple angles (i.e., the first surface image). Then, on this basis, the drone sprays the steel surface again and obtains the surface image after spraying (i.e., the second surface image). By preprocessing these two sets of image data (such as denoising, brightness adjustment, geometric correction, etc.), they are ensured to be accurately aligned in space.

[0219] Next, the aligned images are divided into grids of fixed size, and visual features such as texture, color, and edge are extracted for each grid area to form a feature vector describing the surface state of the area. These feature vectors include the grid segmentation features of the first preprocessed image and the grid segmentation features of the second preprocessed image, providing a data basis for the subsequent determination of rust removal measures.

[0220] Based on these visual feature data, the present invention establishes a rust removal measure determination model based on deep learning. The training data of the model includes a large number of steel surface images in different rust states, as well as manually annotated optimal rust removal methods (sandblasting, grinding or chemical rust removal) and operating parameters. By learning the mapping relationship from image features to rust removal measures, the model can automatically recommend specific rust removal measures (methods and parameters) required for each grid area based on the input first and second segmentation features.

[0221] After determining the best rust removal measures for each area, the corresponding manual operation or automated rust removal equipment can be used for actual rust removal. After rust removal is completed, a new image of the steel surface is collected and the difference is calculated with the standard image of the ideal rust-free state. If the difference is greater than the preset threshold, rust removal continues until the ideal state is reached.

[0222] Next, the present invention designs a multi-objective game optimization model that comprehensively considers the anti-corrosion effect and construction efficiency for the steel surface after rust removal. The model includes a mathematical model that describes the relationship between spraying parameters (pressure, speed, distance) and paint layer thickness and curing time, while considering the influence of environmental conditions on curing time. The spraying parameters of each layer of paint (primer, intermediate paint, topcoat) are used as decision variables, and two objective functions of anti-corrosion effect and construction time are established. By solving this multi-objective optimization problem, a set of Pareto optimal solutions representing the trade-offs between different anti-corrosion effects and construction time can be obtained.

[0223] Finally, the optimal anti-corrosion coating process parameters were selected by conducting small tests on actual steel structures. Specifically, a small test area was set up in the non-critical parts of the steel structure, and spraying tests were carried out according to different schemes in the Pareto optimal solution set. The actual parameters were recorded and the anti-corrosion performance was evaluated. Finally, the scheme with the best comprehensive performance was selected from the test results as the final spraying requirement.

[0224] In order to better understand and implement the present invention, an embodiment of a specific application scenario of the present invention is provided below: During the long-term use of a large steel structure bridge, its surface gradually developed serious rust problems. As a maintenance and management unit, they urgently need to take effective rust removal and anti-corrosion coating measures to ensure the safe service life of the bridge.

[0225] First, the bridge management department sent a drone survey team to conduct a comprehensive scan and photography of the steel structure surface of the entire bridge. Through the high-definition camera carried by the drone, they obtained multiple sets of surface images (i.e., first surface images) at various angles and positions of the bridge. These image data recorded various corrosion conditions accumulated in the bridge steel structure during use.

[0226] Subsequently, the drone carried out another round of spray treatment on the bridge surface and obtained multi-angle surface images after the spray treatment (i.e., the second surface image). These images not only recorded the changes in the bridge surface after the spray treatment, but also provided important references for subsequent image comparison and analysis.

[0227] To ensure that the two sets of image data completely correspond in spatial position, the bridge management department preprocessed these images, including image denoising, brightness adjustment, geometric correction and other steps, and finally obtained the first preprocessed image and the second preprocessed image.

[0228] Next, the management department divided the corrected image into grids of fixed size, and calculated the visual features such as texture, color, and edge for each grid area to form a feature vector describing the surface state of the area. Among them, the grid segmentation feature of the first preprocessed image is recorded as , the grid segmentation feature of the second pre-processed image is recorded as .

[0229] With the help of a large number of steel structure surface corrosion images and corresponding manually annotated data, the bridge management department established a deep learning-based rust removal measure determination model. The input of the model is and , the output is the rust removal method (sandblasting, grinding or chemical rust removal) required for each grid area and its specific parameters. The training process of the model is as follows:

[0230] 1. 100 image samples of steel structure surfaces with different degrees of corrosion were collected, covering slight corrosion, moderate corrosion and severe corrosion.

[0231] 2. Manually annotate these image samples, including the degree of rust (mild, moderate, severe) and the optimal rust removal method (sandblasting, grinding or chemical rust removal) and operating parameters required for each grid area.

[0232] 3. Combine image samples, manual annotation information, and rust removal effect data into training data pairs to establish a deep learning model. The specific data is shown in Table 1:

[0233] Table 1 Training data set samples

[0234]

[0235] 4. Use a convolutional neural network structure to simultaneously predict the rust removal method and operating parameters, define the loss function and use the stochastic gradient descent method for model training.

[0236] 5. Evaluate model performance through cross-validation, and optimize the network structure and hyperparameters.

[0237] 6. Use the test set to verify the generalization ability of the model under different degrees of corrosion to ensure its applicability in practical applications.

[0238] Through the above training process, the bridge management department finally established a high-precision rust removal measure determination model. The model was applied to the bridge surface image data, and the optimal rust removal method and operation parameters were given for each grid area as shown in Table 2:

[0239] Table 2 Recommended rust removal measures for different areas

[0240]

[0241] As can be seen from Table 2, for slightly corroded areas such as bridge decks and piers, the model recommends sandblasting for rust removal. The specific parameters are: sandblasting pressure 150 bar, sandblasting distance 150 mm, and sandblasting time 30 s.

[0242] For areas with moderate corrosion on bridge towers and some piers, the model recommends grinding to remove rust. The specific parameters are: grinding force 3 N, grinding time 120 s.

[0243] For serious corrosion in some dark and humid local areas of the bridge, the model recommends chemical rust removal with the following specific parameters: using 10% hydrochloric acid solution and acting for 240 seconds.

[0244] According to the model's recommendation, the bridge management department organized construction workers to carry out rust removal operations in different areas. During the construction process, they constantly checked the rust removal effect and determined whether the rust removal goal was achieved by calculating the difference from the ideal rust-free state image. For areas where the difference was still greater than 0.1, the rust removal operation continued until the difference was less than 0.1, and the rust removal of the area was completed.

[0245] After the rust removal work was completed, the management department immediately designed the optimal anti-corrosion coating solution for the bridge surface. First, they established a multi-objective game model, including the following key parts:

[0246] Relational Model:

[0247] Describe the spraying parameters (pressure) of each layer of paint (primer, intermediate paint, topcoat) ,speed ,distance ) and paint layer thickness and curing time The mathematical relationship between:

[0248] ;

[0249] in, Parameters to be determined.

[0250] Modified model:

[0251] For ambient temperature ,humidity and wind speed Curing time Impact:

[0252] ;

[0253] in, is the correction factor.

[0254] Objective function:

[0255] Anticorrosion effect target :

[0256] ;

[0257] Construction time target :

[0258] ;

[0259] Constraints:

[0260] Thickness calculation formula

[0261] Thickness limit: ;

[0262] Thickness ratio: ;

[0263] Curing time: ;

[0264] Spraying parameters: ;

[0265] By using the NSGA-II multi-objective optimization algorithm to solve this model, the management department obtained a set of Pareto optimal solutions representing different trade-offs between anti-corrosion effects and construction time.

[0266] Next, they selected several small test areas of 30 cm × 30 cm in non-critical parts of the bridge and conducted spray tests in these test areas according to different schemes in the Pareto optimal solution set. The actual spray parameters, environmental conditions, paint layer thickness, curing time and other data of each scheme were recorded in detail, and the anti-corrosion performance of these test areas, such as adhesion and salt spray resistance, was evaluated.

[0267] Finally, by comprehensively considering the anti-corrosion effect, construction efficiency and cost factors, the management department selected the following optimal anti-corrosion coating process parameters from the test results as shown in Table 3:

[0268] Table 3 Anticorrosion coating process parameters

[0269]

[0270] Primer: Epoxy, thickness 100 μm, curing time 6 h

[0271] Intermediate coating: Epoxy, thickness 150 μm, curing time 8 h

[0272] Topcoat: Polyurethane, thickness 80 μm, curing time 4 h

[0273] The spraying parameters are:

[0274] Primer - pressure 220 bar, spray speed 0.5 m / s, spray distance 300 mm

[0275] Intermediate coating - pressure 250 bar, spraying speed 0.4 m / s, spraying distance 280 mm

[0276] Topcoat - pressure 180 bar, spray speed 0.6 m / s, spray distance 320 mm

[0277] According to the above optimal process plan, the management department organized construction personnel to carry out comprehensive anti-corrosion coating on the entire bridge. After the coating was completed, they tested the coating thickness to ensure that the actual thickness of each layer of paint was consistent with the design requirements.

[0278] Through the above intelligent rust removal and anti-corrosion coating measures, the overall anti-corrosion performance of the bridge has been greatly improved. Compared with the previous method of rust removal and coating that relied solely on manual experience, the solution of the present invention not only significantly improved the construction quality and efficiency, but also reduced labor costs. The management department highly praised this new technical solution and plans to further promote its application in future maintenance management.

[0279] The chart shows the thickness of each layer of coating, the curing time and the spraying parameters. It can be seen that the scheme of the present invention achieves a good balance between anti-corrosion effect and construction efficiency through multi-objective optimization.

[0280] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for rust removal and anti-corrosion construction on the surface of a steel structure, characterized in that: The following steps are involved: S10, using a camera drone to obtain surface images of the steel structure to be subjected to rust removal and anti-corrosion construction at multiple angles, recorded as first surface images; S20, using a spray device carried by the drone to spray the surface of the steel structure, and obtaining surface images of the steel structure at multiple angles after the spray treatment, which are recorded as second surface images; S30, preprocessing the first surface image and the second surface image to obtain a first preprocessed image and a second preprocessed image, and aligning the first preprocessed image and the second preprocessed image at different angles; S40, meshing the aligned first preprocessed image and the second preprocessed image using the same standard, and calculating the segmented image features of each grid, including the grid segmented image features of the first preprocessed image and the grid segmented image features of the second preprocessed image, which are respectively recorded as the first segmentation features and the second segmentation features; S50, using a pre-trained rust removal measure determination model, calculate the segmented image features of each grid to obtain a rust removal measure array for each grid, specifically a parameter vector for each step of the entire rust removal process; the rust removal measure determination model adopts a convolutional neural network structure, and the training data of the rust removal measure determination model includes a large number of steel structure surface images in different rust states, as well as the optimal rust removal method and parameters for each image area manually annotated. During the training process, the rust removal measure determination model learns the mapping relationship from image features to rust removal methods and parameters; the parameter vector includes: rust removal method, operating parameters and operating time. The rust removal method includes three types: sandblasting, grinding or chemical rust removal. Each rust removal method has different operating parameters. The operating parameters of the sandblasting include sandblasting pressure, sandblasting distance, and sandblasting time; the operating parameters of the grinding include grinding intensity and grinding time; the operating parameters of the chemical rust removal include chemical rust remover ratio concentration and action time; S60, performing rust removal operation manually or using rust removal equipment according to the rust removal measure vector of each grid, and determining whether the rust removal operation is completed by calculating the difference between the image after rust removal and the preset standard image without rust, that is, if the difference is greater than or equal to the preset difference threshold, then continue the rust removal operation, if the difference is less than the preset difference threshold, then complete the rust removal operation; S70, after the rust removal operation is completed, spraying the steel structure with anti-corrosion paint, specifically applying primer, intermediate paint and topcoat in sequence according to preset spraying requirements; the spraying requirements are the thickness of each paint layer and the curing time; S80. Check the surface of the steel structure after painting, measure the coating thickness, and ensure that the anti-corrosion effect meets the requirements; The method for determining the thickness and curing time of each paint layer is specifically as follows: S71, establishing a multi-objective game model, including a relational model and a correction model, wherein the relational model is used to describe the relationship between the spraying parameters of each layer of paint and the thickness and curing time of the paint layer, and the correction model is used to describe the influence of environmental conditions on the curing time. The decision variables of the multi-objective game model are the thickness and curing time of the primer, the intermediate paint and the topcoat as decision variables, and the objective function of the multi-objective game model is the anti-corrosion effect and the construction time; S72, fitting the relationship model and using historical data; S73, constructing the constraint conditions of the multi-objective game model, including the minimum and maximum thickness requirements of each layer of paint, the total thickness limit, the thickness ratio relationship between each layer, and the shortest and longest curing time; S74, using a multi-objective optimization algorithm to solve the multi-objective game model, and obtaining a Pareto optimal solution set as the spraying requirement to be selected; S75. Use the small test verification method to test the spraying requirements to be selected, and select the spraying requirements corresponding to the best small test results as the final spraying requirements; Wherein, the step S75 specifically includes: S751. Select several small test areas in non-critical parts of the steel structure; S752. Conduct spraying tests in various test areas according to different solutions in the Pareto optimal solution set; S753, record the actual spraying parameters, environmental conditions, paint layer thickness and curing time of each test area; S754, conduct corrosion resistance evaluation on the test area, including adhesion test and salt spray resistance test; S755. Comprehensively consider the anti-corrosion effect, construction efficiency and cost factors, and select the optimal spraying requirements from the test results; The multi-objective game model is expressed as: ； in: , represents the decision variable vector, including the thickness of each layer of paint, curing time, spraying pressure, spraying speed and spraying distance; is the thickness of the i-th layer of paint, i = 1, 2, 3 represent primer, intermediate paint, and topcoat respectively; is the curing time of the i-th layer of paint; is the spraying pressure of the i-th layer of paint; is the spraying speed of the i-th layer of paint; is the spraying distance of the i-th layer of paint; is the anti-corrosion effect objective function: ； In the formula, is the weight of the thickness of the i-th paint layer; is the maximum allowable thickness of the i-th layer of paint; is the total thickness influencing factor; is the upper limit of total thickness; is the optimal thickness deviation penalty factor; is the ideal optimal thickness of the i-th layer of paint; is the construction time objective function: ； In the formula, : Actual curing time after taking into account environmental conditions; : Preparation time for the i-th coat of paint; : Overall setup time; : Ambient temperature; : Temperature change; actual curing time Calculated by the modified model; in, ; In the formula, is the ambient temperature, , , , , , All are fitting parameters, obtained through historical data regression; The modified model is specifically expressed as: ; In the formula, : Temperature cycle change influence coefficient; : Frequency of temperature cycle change; Phase of temperature cycle change; : Humidity influence coefficient; : Current humidity; : Base humidity; : Wind speed influence coefficient; : wind speed; Theoretical curing time calculation: ； In the formula, : Curing time coefficient; : Humidity influence factor; : activation energy; : gas constant; , : Wind speed influence parameters; The constraints include: Thickness limit constraint: ； ； In the formula, : Minimum thickness of the i-th layer of paint; : Maximum thickness of the i-th layer of paint; : Upper limit of total thickness; Thickness ratio constraint: ； ； In the formula, : Minimum thickness ratio of primer to intermediate paint; : Minimum thickness ratio of intermediate paint to topcoat; Curing time constraints: ； In the formula, : The shortest curing time of the i-th coat of paint; : Maximum curing time of the i-th coat of paint; Spraying parameter constraints: ； ； ； In the formula, , : Spraying pressure range; , : Spraying speed range; , : Spraying distance range.

2. A method for rust removal and anti-corrosion construction of a steel structure surface according to claim 1, characterized in that: The small test area is an area of ​​30 cm x 30 cm.

3. A method for rust removal and anti-corrosion construction of steel structure surface according to claim 1, characterized in that: The spraying parameters include spraying pressure, spraying distance and spraying speed.

4. A method for rust removal and anti-corrosion construction of steel structure surface according to claim 1, characterized in that: The environmental conditions include temperature, humidity and wind speed.

5. A method for rust removal and anti-corrosion construction of steel structure surface according to claim 1, characterized in that: The difference threshold is 0.

1.

6. A method for rust removal and anti-corrosion construction of steel structure surface according to claim 1, characterized in that: The selection criteria for primers are: good adhesion, excellent anti-corrosion properties, and compatibility with the substrate and subsequent coatings; The selection criteria for the intermediate paint are: enhancing the overall coating thickness, improving the anti-corrosion performance, and improving the adhesion of the topcoat; The selection criteria for topcoat are: good weather resistance, decorative properties and UV resistance.

Images