A zinc layer peeling control method based on galvanizing liquid alloy element proportioning

Abstract

The application discloses a zinc layer peeling control method based on galvanizing liquid alloy element proportioning, which firstly calculates diffusion and reaction parameters of alloy elements through a multi-component diffusion reaction coupling model based on CALPHAD, and constructs a diffusion reaction phase diagram; simulates atomic motion through a molecular dynamics-evolutionary neural network analysis algorithm, extracts key characteristic parameters, and establishes a mapping relationship between the key characteristic parameters and the bonding performance of a zinc layer; then collects real-time process parameters through a SIMATIC PCS 7 system of Siemens, integrates multidimensional data through a MES for Metals platform, analyzes parameter correlation and identifies a main influence range, and generates an alloy element proportioning adjustment scheme; finally, the scheme is converted into a control instruction through the SIMATIC PCS 7 system of Siemens, and dynamic regulation and control are realized. The method solves the problem of data disconnection, improves regulation and control accuracy and stability, reduces the probability of zinc layer peeling, and improves product quality consistency.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of zinc layer peeling control, and in particular to a zinc layer peeling control method based on alloy element proportioning of a zinc plating solution. BACKGROUND

[0002] In the production process of zinc-plated products, zinc layer peeling is a key problem affecting product quality and performance, which may cause accelerated corrosion of the base material, shorten the service life of the product, and bring many hidden troubles to subsequent processing and application. With the continuous improvement of the quality requirements of the industry for zinc-plated products, how to accurately control the zinc plating process parameters to reduce the phenomenon of zinc layer peeling has become a technical problem to be solved in the zinc plating industry. Among them, the proportioning of alloy elements in the zinc plating solution as a core factor affecting the formation and bonding performance of the zinc layer, the correlation mechanism between it and the zinc layer peeling and the control method have attracted much attention.

[0003] In the prior art, the control of zinc layer peeling relies mainly on empirical process adjustment. The concentration of several main alloy elements in the zinc plating solution is monitored manually, combined with the basic parameters of the production line such as temperature and speed, and an adjustment scheme is formulated based on historical production data and operating experience. Some processes use a single diffusion model or a simple data analysis method to calculate and evaluate some parameters in the zinc layer formation process, and then use the results to guide the optimization of process parameters in order to try to reduce the probability of zinc layer peeling.

[0004] The prior art has obvious deficiencies. On the one hand, the complex correlation between the diffusion reactions of multiple alloy elements in the zinc plating process and the bonding performance of the zinc layer is not fully considered, and only a single model or local parameter analysis is used, which is difficult to fully reflect the influence of the interaction of various elements on the zinc layer peeling, resulting in limited pertinence and precision of process adjustment. On the other hand, there is a lack of mechanism for effectively integrating the atomic motion characteristics at the microscale, the process parameter changes at the mesoscale, and the macro production execution process. The data transmission and analysis of each link are disconnected, and it is impossible to realize dynamic closed-loop management from parameter calculation to process control, so that the control effect of zinc layer peeling is unstable and difficult to adapt to complex and variable production conditions. SUMMARY

[0005] In order to overcome the shortcomings and deficiencies of the prior art, the present application provides a zinc layer peeling control method based on alloy element proportioning of a zinc plating solution.

[0006] The technical solution adopted by the present application is a zinc layer peeling control method based on alloy element proportioning of a zinc plating solution, which comprises the following steps:

[0007] Step S1, the diffusion coefficients, reaction kinetics parameters of different types of alloying elements in the galvanizing process under different temperature, concentration gradient conditions are calculated by the CALPHAD-based multi-component diffusion reaction coupling model, the diffusion reaction phase diagram of the multi-component system is constructed, and the influence law of the interaction between different types of alloying elements on the formation of zinc layer is determined;

[0008] Step S2, the atomic motion trajectory, interface binding energy and zinc layer crystal structure evolution process of the alloying elements in the galvanizing solution are simulated and analyzed by using the molecular dynamics-evolutionary neural network analysis algorithm, the calibrated atomic scale characteristic parameters affecting the bonding force of the zinc layer and the substrate are extracted, and the parameters are used as the input variables of the analysis algorithm, and the mapping relationship between the characteristic parameters and the bonding performance of the zinc layer is established through the training process of the neural network;

[0009] Step S3, the real-time running data of the galvanizing production line are collected by using the Siemens SIMATIC PCS 7 system, including the galvanizing solution temperature, flow rate, alloying element concentration, steel strip running speed, and coating thickness process parameters, the collected data are transmitted to the data processing unit for real-time storage and preliminary analysis through the communication module built-in the system;

[0010] Step S4, the multi-component diffusion reaction calculation results obtained in step S1, the atomic scale characteristic parameters and the mapping relationship obtained in step S2, and the real-time process parameters collected in step S3 are imported into the MES for Metals platform, the correlation between different parameters is mined through the data analysis module of the platform, and the process parameter range affecting the zinc layer peeling is identified.

[0011] Step S5, based on the process parameter range determined in step S4, a multi-component galvanizing solution alloying element ratio adjustment scheme is generated through the process optimization module of the MES for Metals platform, which includes the concentration adjustment value of different types of alloying elements, and the corresponding temperature and flow rate matching parameters;

[0012] Step S6, the optimal ratio adjustment scheme generated in step S5 is converted into specific control instructions through the control module of the Siemens SIMATIC PCS 7 system, and is issued to the execution mechanism of the galvanizing production line to dynamically control the galvanizing solution alloying element ratio and related process parameters;

[0013] Further, in step S1, during the calculation process of the CALPHAD-based multi-component diffusion reaction coupling model, a zinc layer peeling sensitivity factor calculation model is introduced, and the formula of the model is: wherein, is the zinc layer peeling sensitivity factor, is the model correction coefficient, is the actual concentration of the i th alloying element, is the reference concentration of the i th alloying element, is the diffusion coefficient of the i th alloying element, is the diffusion activation energy of the i th alloying element, is the gas constant, is the absolute temperature, is the Gibbs free energy change of the multi-component reaction; the zinc layer exfoliation sensitivity factor under different alloy element ratios is calculated by the model, which is used as a quantitative index to judge the influence degree of multi-component diffusion reaction on zinc layer exfoliation, and the construction process of multi-component diffusion reaction phase diagram is optimized according to the index.

[0014] Further, step S2 includes the following steps: step S2.1, using molecular dynamics method to simulate the adsorption, diffusion and deposition process of alloy element atoms in the galvanizing liquid on the substrate surface, setting the simulation time step to 1 fs to 10 fs, the simulation temperature range covers the actual galvanizing process temperature interval, recording the spatial coordinates, velocity and interaction force data of different atoms at different times, and generating the atomic trajectory file; step S2.2, based on the generated atomic trajectory file, calculating the atomic coordination number, bond length distribution and binding energy of the interface region between the zinc layer and the substrate, determining the micro area with weak interface bonding force and the corresponding atomic arrangement characteristics by analyzing the variation law of the parameters; step S2.3, after the interface binding energy, atomic coordination number and atomic scale characteristic parameters are standardized, they are input into the input layer of the evolutionary neural network, the hidden layer of the neural network uses 10-20 neurons, the output layer is the zinc layer bonding performance evaluation value, and the initial weight and threshold value of the neural network are optimized by genetic algorithm; step S2.4, using the verification data set to test the trained evolutionary neural network, adjusting the structure parameters of the neural network according to the test results, until the prediction error of the network for the zinc layer bonding performance is controlled within the preset range.

[0015] Further, in step S3, the Siemens SIMATIC PCS 7 system adopts a process parameter fluctuation coefficient calculation model in the process of collecting and processing real-time running data of the galvanizing production line, and the formula of the model is: , wherein, is the fluctuation coefficient of the j th process parameter, is the number of data collection, is the k th collection value of the j th process parameter, is the average value of the j th process parameter; the fluctuation coefficient of different process parameters is calculated by the model, when the fluctuation coefficient exceeds the set threshold value, the Siemens SIMATIC PCS 7 system starts the abnormal data marking mechanism, and the corresponding abnormal data and normal data are stored separately, and abnormal marking information is added in the data transmission process.

[0016] Further, in step S4, when the data analysis module of the MES for Metals platform mines the correlation between different parameters, a zinc layer peeling risk assessment model is introduced, and the formula of the model is: wherein, is the zinc layer peeling risk value, is the weight coefficient, is the influence weight of the jth process parameter, is the actual value of the jth process parameter, is the maximum allowable value of the jth process parameter, is the zinc layer peeling sensitivity factor; by calculating the zinc layer peeling risk value under different process parameter combinations, the parameter range corresponding to the process parameter combination with a risk value higher than the set critical value is identified as the process parameter range affecting the zinc layer peeling.

[0017] Further, step S4 includes the following sub-steps: step S4.1, classifying the calculation results of the multi-group diffusion reaction coupling model in step S1, the mapping relationship data obtained by the molecular dynamics-evolutionary neural network analytical algorithm in step S2, and the real-time process parameters collected by the Siemens SIMATIC PCS 7 system in step S3 according to the data types, and establishing a comprehensive database including element concentration, temperature, diffusion coefficient, binding energy, and process parameter dimension; step S4.2, using the association rule algorithm in the data analysis module of the MES for Metals platform to mine the data in the comprehensive database, calculate the correlation coefficients between different parameters, and select the parameter pairs with an absolute value of the correlation coefficient greater than 0.6 to determine the strong correlation relationship between the parameters; step S4.3, constructing a parameter influence network model based on the strong correlation relationship, taking the zinc layer peeling phenomenon as the target node, analyzing the direct and indirect influence degree of different parameters on the zinc layer peeling through the connection strength between nodes, and generating an influence degree ranking table; step S4.4, according to the influence degree ranking table, selecting the top 8-10 parameters as the calibration influence parameters, and combining the controllability of the parameters in actual production to determine the final process parameter range affecting the zinc layer peeling.

[0018] Further, in step S5, when the process optimization module of the MES for Metals platform generates the galvanizing liquid alloy element ratio adjustment scheme, a ratio optimization objective function model is used, and the formula of the model is: wherein, is the optimization objective function value, is the weight coefficient, is the zinc layer peeling risk value, is the adjusted concentration of the ith alloy element, is the ideal concentration of the ith alloy element, The energy consumption coefficient is adjusted for the process. The maximum value of the objective function is solved by a genetic algorithm. The combination of the different kinds of alloy element concentrations, temperatures and flow rates corresponding to the maximum value of the objective function is taken as the optimal matching adjustment scheme.

[0019] Further, step S5 includes the following sub-steps: step S5.1, setting the adjustment step length of the different kinds of alloy element concentrations to 0.01%-0.1%, the temperature adjustment step length to 1-5℃, and the flow rate adjustment step length to 0.1-0.5m / s, and constructing a process parameter adjustment space with the process parameter range determined in step S4 as a constraint; step S5.2, randomly generating an initial matching adjustment scheme set in the process parameter adjustment space, the number of schemes in the set being 50-100 groups, and each group of schemes including complete alloy element concentrations, temperatures and flow rates; step S5.3, calling the process simulation module of the MES for Metals platform to simulate the galvanizing process for each group of schemes in the initial scheme set, and calculating the zinc layer adhesion and the risk value evaluation index corresponding to each group of schemes; step S5.4, screening the initial scheme set according to the evaluation index, retaining 30%-50% of the schemes with better evaluation indexes, and generating new matching adjustment schemes through crossover and mutation operations on the schemes, repeating the simulation and screening processes until multiple adjustment schemes meeting the preset evaluation standard are obtained.

[0020] Further, step S6 includes the following sub-steps: step S6.1, the control module of the SIMATIC PCS 7 system receives the optimal matching adjustment scheme sent by the MES for Metals platform, analyzes the different kinds of alloy element concentration adjustment values, temperature and flow rate matching parameters included in the scheme, and converts them into digital signals recognizable by the system; step S6.2, the control module sends control instructions to the galvanizing liquid alloy element adding device according to the converted digital signals, adjusts the valve opening and adding frequency of the adding device to adjust the concentrations of different kinds of alloy elements; step S6.3, at the same time, sends corresponding control signals to the temperature control device and the flow rate control device to adjust the heating power and the pump body speed, so that the temperature and flow rate of the galvanizing liquid reach the values set in the adjustment scheme; step S6.4, during the regulation process, the control module receives the feedback signals of each actuator in real time, compares them with the target values of the adjustment scheme, dynamically adjusts the control instructions according to the deviation values, and stabilizes the different process parameters in the target range.

[0021] A zinc layer peeling control method based on the matching of galvanizing liquid alloy elements, which is realized by different units and includes:

[0022] A multi-component diffusion reaction parameter calculation unit is connected with the molecular dynamics-evolutionary neural network analysis unit through a data transmission bus, and is used to transmit the calculated multi-component diffusion reaction parameters to the molecular dynamics-evolutionary neural network analysis unit;

[0023] A molecular dynamics-evolutionary neural network analysis unit is connected with the production line real-time data acquisition and processing unit through a wireless communication module, and is used to transmit the analyzed atomic scale characteristic parameters and mapping relationship to the production line real-time data acquisition and processing unit;

[0024] A production line real-time data acquisition and processing unit is connected with the metal manufacturing execution platform data correlation analysis unit through an Ethernet, and is used to transmit the acquired and processed real-time process parameters to the metal manufacturing execution platform data correlation analysis unit;

[0025] A metal manufacturing execution platform data correlation analysis unit is connected with the process parameter optimization scheme generation unit through a fiber communication, and is used to transmit the identified process parameter range to the process parameter optimization scheme generation unit;

[0026] A process parameter optimization scheme generation unit is connected with the production process dynamic regulation and control execution unit through an industrial bus, and is used to transmit the generated optimal proportioning adjustment scheme to the production process dynamic regulation and control execution unit;

[0027] A production process dynamic regulation and control execution unit is connected with the multi-component diffusion reaction parameter calculation unit through a feedback signal line, and is used to feed back the regulated process parameters to the multi-component diffusion reaction parameter calculation unit.

[0028] Beneficial effects: the present application proposes a zinc layer peeling control method based on the alloy element proportioning of galvanizing liquid, which is based on the multi-component diffusion reaction coupling model of CALPHAD, comprehensively analyzes the diffusion and reaction rules of various alloy elements under different conditions, combines the molecular dynamics-evolutionary neural network analysis algorithm to mine the correlation between atomic scale characteristics and zinc layer bonding performance, overcomes the defects of insufficient consideration of multi-component complex correlation in the prior art, and improves the pertinence and accuracy of process adjustment. Meanwhile, through real-time acquisition of production data by the SIMATIC PCS 7 system of Siemens, the MES for Metals platform is used to realize correlation analysis and process optimization of multi-scale data, and a closed-loop management from parameter calculation to dynamic regulation and control is formed, solving the problem of disconnection of data in each link in the prior art and inability to realize dynamic closed-loop management, so that the zinc layer peeling control effect is more stable and can adapt to complex production conditions. In addition, the method realizes intelligent regulation and control of the alloy element proportioning of galvanizing liquid through multi-model cooperation and system integration, reduces the dependence on empirical adjustment, and greatly improves the stability and reliability of the zinc layer quality. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 A method step flow chart for the present application;

[0030] Figure 2 A method implementation unit composition chart for the present application. DETAILED DESCRIPTION

[0031] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict, and the present application will be further described in detail below in combination with the drawings and specific embodiments.

[0032] As shown in Figure 1 A zinc layer peeling control method based on galvanizing liquid alloy element ratio, the method comprising:

[0033] Step S1, through the multi-component diffusion reaction coupling model based on CALPHAD, the diffusion coefficients and reaction kinetics parameters of different kinds of alloying elements in the galvanizing process under different temperature and concentration gradient conditions are calculated, the diffusion reaction phase diagram of the multi-component system is constructed, and the influence law of the interaction between different kinds of alloying elements on the formation of zinc layer is determined;

[0034] Specifically, the multi-component diffusion reaction coupling model based on CALPHAD is the basic link to realize the zinc layer peeling control, and the diffusion and reaction characteristics of multiple alloying elements in the galvanizing process are accurately calculated. This model can cover common alloying elements such as zinc, iron, aluminum, nickel, and silicon, and by calculating the diffusion coefficients and reaction kinetics parameters of these elements under different temperatures (usually 420-480℃) and concentration gradients (such as zinc element concentration gradient can reach 0.05-0.2% / mm), it provides key data support for subsequent process optimization. The diffusion reaction phase diagram of the multi-component system can clearly present the interaction rules between elements, such as the types and distribution of intermetallic compounds formed by different element combinations, which is of great significance to understand the bonding mechanism of zinc layer and substrate and predict the risk of zinc layer peeling. Through this step, the influence of the concentration change of each element on the diffusion rate and reaction degree can be determined, which provides a theoretical basis for subsequent adjustment of alloy element ratio.

[0035] In the implementation process, first, the types and initial concentration ranges of alloying elements involved in the galvanizing solution need to be determined, generally including zinc (concentration range 95-98%), aluminum (0.1-0.3%), iron (0.05-0.15%), and other trace elements (such as nickel, silicon, etc., total concentration not exceeding 0.5%). Input these element parameters into the multi-component diffusion reaction coupling model based on CALPHAD, set the temperature parameters to 420°C, 440°C, 460°C, and 480°C four gradients, and the concentration gradient to 0.05% / mm, 0.1% / mm, 0.15% / mm, and 0.2% / mm four levels. When the model runs, it will calculate the diffusion coefficient of each element (the diffusion coefficient of zinc in this temperature range is usually 1×10⁻¹²-5×10⁻¹¹ m² / s), reaction activation energy, and other kinetic parameters according to different temperature and concentration gradient conditions. Subsequently, according to these calculation results, a multi-component diffusion reaction phase diagram is constructed using a phase diagram drawing tool, and the phase region distribution, reaction products, and stability of each element under different conditions need to be marked in the phase diagram. Through analysis of the phase diagram, it is determined that when the aluminum element concentration exceeds 0.3%, too much brittle intermetallic compound may be formed, increasing the risk of zinc layer shedding; and appropriately increasing the nickel element concentration (0.05-0.1%) can improve the bonding force between the zinc layer and the substrate, etc. key laws, providing clear parameter adjustment direction for subsequent steps.

[0036] In step S2, the atomic motion trajectory, interface binding energy, and zinc layer crystal structure evolution process of the alloying elements in the galvanizing solution are simulated and analyzed using a molecular dynamics-evolutionary neural network analysis algorithm, the calibrated atomic scale feature parameters that affect the bonding force between the zinc layer and the substrate are extracted, and the parameters are used as input variables of the analysis algorithm. Through the training process of the neural network, the mapping relationship between the feature parameters and the bonding performance of the zinc layer is established.

[0037] Specifically, the molecular dynamics-evolutionary neural network analysis algorithm plays an important role in the microscale analysis of zinc layer shedding control. Its main function is to explore the internal relationship between the atomic motion behavior of alloying elements in the zinc plating solution and the bonding performance of the zinc layer. Through molecular dynamics simulation, the algorithm can capture the atomic trajectory at the nanoscale, analyze the changes in interfacial bonding energy (usually in the range of 0.5-2.0 J / m²), and the evolution process of zinc layer crystal structure, such as crystal orientation, grain size (usually 5-20 nm) changes on the bonding force. The extracted key atomic scale feature parameters include atomic coordination number (common range 6-12), bond length distribution (usually between 0.25-0.3 nm), diffusion activation energy, etc., which can accurately reflect the micro bonding state of the zinc layer and the substrate interface. Through the training and learning of these parameters, the evolutionary neural network establishes a mapping relationship between the feature parameters and the zinc layer bonding performance, providing an effective tool for predicting the zinc layer shedding possibility from a micro perspective, and helping to improve the micro-targetedness of process control.

[0038] In specific implementation, first, an atomic model of the zinc plating solution and the substrate interface is constructed using molecular dynamics simulation software. The model includes atoms of zinc, iron, and other alloying elements, with a number of atoms typically between 10,000-50,000, and the simulation box size is set to 10 nm x 10 nm x 20 nm. The simulation time step is set to 1 fs, and the total simulation time is 100 ps, with the temperature controlled within the range of 420-480°C, consistent with the actual zinc plating process temperature. During the simulation, the atomic spatial coordinates, velocities, and interaction force data are recorded every 1000 steps to generate atomic motion trajectory files. Based on the trajectory files, the atomic coordination number in the interface area (such as the coordination number of zinc atoms at the interface, usually 8-10), bond length distribution (zinc-iron bond length about 0.28 nm), and interfacial bonding energy (when the bonding energy is lower than 0.8 J / m², the zinc layer bonding performance is poor) are calculated. After sorting these feature parameters, they are imported into the evolutionary neural network as input variables. The input layer node number of the neural network is 8-12 (determined according to the number of feature parameters), the hidden layer is set to 2 layers, each containing 15 neurons, and the output layer is the zinc layer bonding performance evaluation value. Genetic algorithm is used to optimize the initial weights and thresholds of the neural network, the training data set contains 500-1000 sample data, the training iteration number is 1000 times, and the learning rate is set to 0.01. After training, 200 validation data are used to test the network to ensure that the network prediction error is controlled within 5%, and finally a stable mapping relationship between feature parameters and zinc layer bonding performance is established.

[0039] Step S3, real-time running data of the galvanizing production line is collected by using the Siemens SIMATIC PCS 7 system, including galvanizing liquid temperature, flow rate, alloy element concentration, steel strip running speed, and coating thickness process parameters, and the collected data is transmitted to a data processing unit for real-time storage and preliminary analysis through a communication module built in the system;

[0040] Specifically, the Siemens SIMATIC PCS 7 system as the core of real-time data acquisition and control of the production line, its function implementation is crucial to the real-time and accuracy of the zinc layer peeling control. The system can comprehensively and efficiently collect key process parameters of the galvanizing production line, including galvanizing liquid temperature (control accuracy needs to reach ±1℃), flow rate (usually 1-3m / s), alloy element concentration (detection accuracy is 0.01%), steel strip running speed (100-300m / min), and coating thickness (5-20μm) etc. Real-time changes of these parameters directly affect the formation quality of the zinc layer. Through the communication module (such as PROFINET module) built in the system, the collected data can be transmitted to the data processing unit at a frequency of 100ms / second, realizing real-time storage and preliminary analysis of the data. The data processing unit will process the data by filtering, outlier detection, etc., to ensure the accuracy and effectiveness of the data, providing reliable basic data support for subsequent integration and analysis with other model data, and is the prerequisite for realizing dynamic regulation of process parameters.

[0041] In the implementation process, first, corresponding sensors are installed at key positions of the galvanizing production line, such as platinum resistance temperature sensors (measurement range 400-500℃, accuracy ±0.5℃) are installed in the galvanizing tank to monitor the galvanizing liquid temperature in real time; electromagnetic flow meters (measurement range 0.5-5m / s, accuracy ±0.5%) are installed to monitor the galvanizing liquid flow rate; X-ray fluorescence spectrometers (detection accuracy 0.01%) are installed to analyze the alloy element concentration online; laser thickness gauges (measurement range 0-50μm, accuracy ±0.1μm) are installed at the steel strip inlet and outlet to detect the coating thickness; encoders are installed to measure the steel strip running speed (measurement range 50-500m / min, accuracy ±0.1m / min). These sensors are connected with the input modules of the SIMATIC PCS 7 system of Siemens, and the system sets the sampling frequency of each sensor to 10Hz through hardware configuration. The collected data are processed by the OB35 cycle organization block inside the system, the execution cycle of the organization block is set to 100ms, the data are smoothed and filtered to remove high-frequency noise. At the same time, the built-in outlier detection algorithm of the system will mark the data that exceed the preset range (such as temperature exceeding 480℃ or below 420℃) as abnormal, and store them in a special abnormal database, while the normal data are stored in the real-time database, and the storage capacity of the database can meet the historical data storage requirement of at least 30 days. The processed data are transmitted to the MES for Metals platform through the communication module of the system, the transmission protocol uses TCP / IP to ensure the stability and real-time performance of data transmission.

[0042] Step S4, the multi-group element diffusion reaction calculation results obtained in step S1, the atomic scale characteristic parameters obtained in step S2 and the mapping relationship, and the real-time process parameters collected in step S3 are imported into the MES for Metals platform, the correlation between different parameters is mined through the data analysis module of the platform, and the process parameter range affecting the zinc layer peeling is identified;

[0043] Specifically, the MES for Metals platform plays a key role in data integration and analysis in the whole zinc layer peeling control method. It can effectively fuse multi-dimensional and multi-source data and deeply mine the correlation between parameters. The platform receives multiple sets of calculation results of multi-component diffusion reaction (such as diffusion coefficient, reaction phase diagram data) from S1, atomic scale characteristic parameters and mapping relationship from S2, and real-time process parameters from S3, forming a comprehensive data set. Through the data analysis module of the platform (including algorithms such as association rule analysis and principal component analysis), these data are deeply processed, and the main process parameter range affecting zinc layer peeling can be identified, such as the optimal concentration range of aluminum element in the zinc plating solution and the suitable temperature interval. This step breaks down the data barriers between different models and systems, realizes the correlation analysis from microcosmic calculation to macroscopic process parameters, provides a scientific basis for subsequent generation of accurate process optimization scheme, and is an important bridge connecting theoretical analysis and actual production control.

[0044] In specific implementation, first, the multi-component diffusion reaction phase diagram data, element diffusion coefficients and reaction kinetics parameters obtained in S1 are imported into the database of the MES for Metals platform, and the data format is converted to CSV format compatible with the platform. At the same time, the zinc layer bonding performance evaluation value output by the evolutionary neural network in S2 and the corresponding atomic characteristic parameters (atomic coordination number, interface binding energy, etc.) are uploaded to the platform to establish a microcosmic parameter database. The real-time process parameters (one record every 100 ms) collected by the Siemens SIMATIC PCS 7 system in S3 are synchronized in real time to the real-time database of the platform through the OPC interface, and the platform summarizes and calculates the statistical values of average temperature, average flow rate, and average element concentration every hour. The data analysis module of the platform starts the association rule algorithm, sets the minimum support degree to 0.2 and the minimum confidence degree to 0.7, analyzes the statistical data of the last 30 days, and calculates the correlation degree between each process parameter and the zinc layer peeling phenomenon (marked by subsequent detection data). Through the principal component analysis algorithm, parameters with a contribution degree to zinc layer peeling exceeding 10% are extracted, such as zinc plating solution temperature (440-460℃), aluminum element concentration (0.15-0.25%), and steel strip running speed (150-250m / min), and the main influence range of these parameters is determined. The analysis results are displayed in the form of charts and tables, such as parameter correlation thermodynamic diagram and main influence parameter ranking table, providing clear parameter adjustment targets for the process optimization scheme generation of S5.

[0045] Step S5, based on the process parameter range determined in step S4, a multi-component zinc plating solution alloy element ratio adjustment scheme is generated through the process optimization module of the MES for Metals platform. The scheme includes concentration adjustment values of different types of alloy elements, corresponding temperature and flow rate matching parameters.

[0046] Specifically, the process optimization module of the MES for Metals platform is the core link for generating the adjustment scheme of the alloying element ratio of the galvanizing solution. Based on the main process parameter range identified by S4, the module generates multiple feasible adjustment schemes in combination with a multi-objective optimization algorithm. These schemes not only contain the concentration adjustment values of each alloying element (such as the specific concentration changes of elements such as zinc, aluminum, nickel, etc.), but also match the corresponding temperature and flow rate parameters to ensure the synergy between parameters. The module takes into account multiple factors such as zinc layer bonding performance, production efficiency, cost, etc., and optimizes the production conditions as much as possible under the premise of meeting the zinc layer shedding control requirements. The multiple generated schemes provide sufficient selection space for subsequent selection of the optimal scheme, and the diversity of the schemes can adapt to different production conditions and quality requirements. Through evaluation and screening of the schemes, the final optimal scheme can effectively reduce the zinc layer shedding rate while ensuring the stability and economy of production.

[0047] In the specific implementation process, the process optimization module of the MES for Metals platform first sets the constraint conditions according to the main process parameter range determined by S4, such as limiting the temperature of the galvanizing solution to 440-460℃, the concentration of aluminum to 0.15-0.25%, and the running speed of the steel strip to 150-250m / min. The optimization target is set to be the zinc layer shedding rate below 0.5% while the coating thickness fluctuation is controlled within ±1μm. The module uses a genetic algorithm for multi-objective optimization, with the initial population size set to 80 groups. Each scheme contains the concentration values of zinc (95-98%), aluminum (0.15-0.25%), iron (0.05-0.15%), and nickel (0-0.1%), as well as the corresponding temperature (440-460℃) and flow rate (1.5-2.5m / s) parameters. The algorithm iteration number is set to 50 generations, the crossover probability is 0.8, and the mutation probability is 0.05. The schemes in each generation population are evaluated by calling the built-in process simulation model of the platform, which calculates and predicts the zinc layer shedding rate and coating thickness fluctuation value based on the input parameters. After the evaluation is completed, the top 50% of the schemes in terms of fitness value are retained to enter the next generation iteration, and new random schemes are introduced to supplement the population size. After 50 generations of iteration, 30 groups of optimization schemes that meet the constraint conditions are generated, each containing detailed adjustment values of the concentration of each alloying element (such as adjusting the aluminum element from the current 0.2% to 0.22%), temperature adjustment value (such as adjusting from 450℃ to 455℃), and flow rate adjustment value (such as adjusting from 2m / s to 2.2m / s), and are sorted in order of zinc layer shedding rate from low to high for subsequent selection of the optimal scheme.

[0048] Step S6, the optimal matching adjustment scheme generated in step S5 is converted into specific control instructions by the control module of the Siemens SIMATIC PCS 7 system, and is issued to the actuator of the galvanizing production line to dynamically regulate the galvanizing liquid alloy element matching and related process parameters.

[0049] Specifically, converting the optimal matching adjustment scheme into control instructions by the Siemens SIMATIC PCS 7 system and issuing it to the actuator is the final link to realize dynamic regulation of zinc layer peeling, and directly determines the landing effect of the process optimization scheme in actual production. This step needs to convert the abstract scheme parameters generated by the MES for Metals platform (such as element concentration target value, temperature target value, etc.) into specific control signals (such as valve opening, motor speed, etc.) that can be recognized by the actuator. The control module (such as the PID control module) of the Siemens SIMATIC PCS 7 system can dynamically adjust the control instructions according to the deviation between the target value and the actual measured value, to ensure that each process parameter is stable within the target range. This process realizes closed-loop control from scheme to execution, can respond to parameter fluctuations in the production process in real time, correct deviations in a timely manner, and ensure that the galvanizing liquid alloy element matching and related process parameters are always in the optimal state, thereby effectively controlling the occurrence of zinc layer peeling.

[0050] In implementation, first, from the 30 groups of optimization schemes generated by the MES for Metals platform, the scheme with the lowest zinc layer peeling rate and the smallest adjustment range of each parameter is selected as the optimal scheme. The parameters in the optimal scheme include: zinc concentration 97.5%, aluminum concentration 0.22%, iron concentration 0.1%, nickel concentration 0.05%, temperature 455℃, flow rate 2.2m / s. The control module of the SIMATIC PCS 7 system of Siemens receives these parameters, and converts them into control instructions through internal algorithms: for the alloy element adding device, the aluminum element concentration is adjusted from the current 0.2% to 0.22%, and the corresponding adding device valve opening is adjusted from 30% to 35%, and the adding frequency remains 5 times / min; the nickel element concentration is adjusted from 0.04% to 0.05%, and the valve opening is adjusted from 20% to 25%. For the temperature control device, the target temperature is 455℃, the current temperature is 450℃, and the control module outputs the heating power adjustment instruction, which increases the heating tube power from 80% to 85%. For the flow rate control device, the target flow rate is 2.2m / s, the current flow rate is 2m / s, and the control instruction increases the circulating pump motor speed from 1450r / min to 1550r / min. The control module uses the PID control algorithm, the proportional coefficient is set to 0.5, the integral time is set to 10s, and the differential time is set to 2s, the deviation is calculated and the control instruction is adjusted every 500ms. During the regulation process, the system collects the actual values of each parameter in real time, and when the aluminum element concentration reaches 0.22%, the temperature stabilizes at 455℃, and the flow rate stabilizes at 2.2m / s, the control module enters the stable control stage, maintains the current control instruction, ensures that each parameter fluctuates within the target value ±5%, and realizes the dynamic stable control of zinc layer peeling.

[0051] Preferably, in step S1, a zinc layer peeling sensitivity factor calculation model is introduced in the CALPHAD-based multi-component diffusion reaction coupling model calculation process, and the model formula is: wherein, is the zinc layer peeling sensitivity factor, is the model correction coefficient, is the actual concentration of the i-th alloy element, is the reference concentration of the i-th alloy element, is the diffusion coefficient of the i-th alloy element, is the diffusion activation energy of the i-th alloy element, is the gas constant, is the absolute temperature, is the Gibbs free energy change of the multi-component reaction; through the model, the zinc layer peeling sensitivity factor under different alloy element ratios is calculated, which is used as a quantitative index to judge the influence degree of multi-component diffusion reaction on zinc layer peeling, and the construction process of the multi-component diffusion reaction phase diagram is optimized according to the index.

[0052] Specifically, step S1 calculates the model of the zinc layer shedding sensitivity factor by introducing it, further quantifying the influence of multi-component diffusion reaction on zinc layer shedding. The model converts the originally complex multi-component diffusion reaction process into a directly assessable sensitivity factor, providing a clear quantitative index for judging the zinc layer shedding risk under the element ratio. The correction coefficient in the model 、 needs to be calibrated according to different galvanizing process types, for example, for continuous hot galvanizing process, the correction coefficient can be set to 0.6, and the correction coefficient of the hot-dip galvanizing process is set to 0.4 to adapt to its process characteristics. In actual implementation, first, the reference concentration of each alloying element needs to be determined , such as the reference concentration of zinc set to 96%, aluminum to 0.2%, and iron to 0.1%. These reference values are determined based on the process parameters of historical high-quality products. In the calculation process, the diffusion coefficient needs to be combined with the calculation results of the CALPHAD model in S1, for example, at 450℃, the diffusion coefficient of zinc is 3×10⁻¹¹ m² / s, and the diffusion coefficient of aluminum is 5×10⁻¹² m² / s. The diffusion activation energy refers to the relevant metal handbook, and the diffusion activation energy of zinc is about 120 kJ / mol, and that of aluminum is about 100 kJ / mol. The gas constant R is 8.314 J / (mol·K), and the absolute temperature T is converted according to the actual process temperature, such as 450℃ corresponding to 723K. The Gibbs free energy change of multi-component reaction is obtained by CALPHAD model calculation, when is negative, the reaction proceeds spontaneously, and intermetallic compounds that are conducive to bonding may be formed, otherwise it may have adverse effects. By calculating the sensitivity factor S under different element ratios, when the S value exceeds 0.8, it is determined as a high-risk ratio, which needs to be optimized in the subsequent process adjustment. The application of this model makes the multi-component diffusion reaction phase diagram constructed in S1 more targeted, and the risk level of the phase region in the phase diagram can be divided according to the S value. The element ratio corresponding to the high-risk phase region will be excluded first, thereby improving the efficiency and accuracy of subsequent process optimization.

[0053] Preferably, the step S2 comprises the following sub-steps: step S2.1, using molecular dynamics method to simulate the adsorption, diffusion and deposition process of alloying element atoms in the galvanizing liquid on the substrate surface, setting the simulation time step to 1 fs to 10 fs, the simulation temperature range covering the actual galvanizing process temperature interval, recording the spatial coordinates, velocity and interaction force data of different atoms at different times, and generating the atomic trajectory file; step S2.2, based on the generated atomic trajectory file, calculating the atomic coordination number, bond length distribution and binding energy of the interface region of the zinc layer and the substrate, determining the micro area with weak interface bonding force and the corresponding atomic arrangement characteristics by analyzing the variation law of the parameters; step S2.3, after the interface binding energy, atomic coordination number and atomic scale characteristic parameters are standardized, they are input into the input layer of the evolutionary neural network, the hidden layer of the neural network uses 10-20 neurons, the output layer is the zinc layer bonding performance evaluation value, and the initial weight and threshold value of the neural network are optimized through the genetic algorithm; step S2.4, using the verification data set to test the trained evolutionary neural network, adjusting the structure parameters of the neural network according to the test result, and until the prediction error of the network to the zinc layer bonding performance is controlled within the preset range.

[0054] Specifically, step S2 elaborates the implementation process of the molecular dynamics-evolutionary neural network analysis algorithm through four sub-steps, aiming to more accurately mine the correlation between atomic scale features and zinc layer bonding performance. In step S2.1, when constructing the atomic model, the atomic types and proportions in the model need to be determined according to the actual galvanizing liquid composition, such as 96% zinc atoms, 3.5% iron atoms, and 0.5% aluminum atoms. The size of the simulation box needs to ensure that the atoms have enough space to move while avoiding the influence of boundary effects. A size of 10 nm x 10 nm x 20 nm can balance the calculation accuracy and efficiency. The simulation time step is 1 fs to ensure that the rapid motion of atoms can be captured, and the total simulation time of 100 ps is sufficient to observe the diffusion and interface bonding process of atoms. In step S2.2, when calculating the interface parameters, a special molecular dynamics analysis software needs to be used to post-process the trajectory file. The atomic coordination number is calculated by the Voronoi tessellation method, the bond length distribution is analyzed by the radial distribution function, and the interface bonding energy is calculated by the stripping method, i.e., calculating the energy required after separating the zinc layer and the substrate. In step S2.3, when constructing the neural network, the number of input layer nodes needs to be consistent with the number of feature parameters, such as including 8 feature parameters such as atomic coordination number, average bond length, and interface bonding energy. The input layer node number is set to 8. The hidden layer is set to 2 layers with 15 neurons, which can better fit the non-linear mapping relationship. When optimizing with genetic algorithm, the population size is set to 50, and the iteration number is 100 times to quickly find the optimal weight and threshold. In step S2.4, when verifying the network, the validation dataset needs to cover samples under different process conditions to ensure the generalization ability of the network, and the prediction error is controlled within 5% to ensure the reliability of subsequent process evaluation. The entire process needs to be performed on a high-performance computing platform. The calculation time for a single simulation is about 24 hours, and the neural network training takes about 10 hours. By processing multiple samples in batches, the mapping relationship established can be directly used for predicting the bonding performance of new proportioning schemes in subsequent process optimization, providing micro-level decision-making basis for process optimization.

[0055] Preferably, in step S3, the Siemens SIMATIC PCS 7 system adopts a process parameter fluctuation coefficient calculation model in the process of collecting and processing real-time operation data of the galvanizing production line. The formula of the model is: wherein, is the fluctuation coefficient of the jth process parameter, is the number of data collection times, is the kth collection value of the jth process parameter, is the average value of the jth process parameter; the fluctuation coefficient of different process parameters is calculated by the model, and when the fluctuation coefficient exceeds the set threshold value, the abnormal data marking mechanism of the Siemens SIMATIC PCS 7 system is started, the corresponding abnormal data is stored separately from the normal data, and abnormal marking information is attached during data transmission.

[0056] Specifically, step S3 introduces a process parameter fluctuation coefficient calculation model for quantitative evaluation of the stability of each real-time process parameter, providing an effective means for identifying abnormal fluctuations and ensuring data quality. The model objectively reflects the dispersion degree of the parameter by calculating the fluctuation coefficient F_j of the parameter, avoiding the subjectivity of judging the stability of the parameter based on experience. When implemented, the data collection frequency m needs to be determined according to the parameter variation characteristics. For parameters such as temperature and flow rate that change rapidly, m can be set to 300 (i.e., data collected within 5 minutes), and for parameters such as element concentration that change slowly, m can be set to 60 (i.e., data collected within 10 minutes). When calculating, first extract the m collection values of the jth parameter from the real-time database of the Siemens SIMATIC PCS 7 system , such as 300 collection values of the temperature parameter, calculate the average value , and then calculate the ratio of the standard deviation to the average value to obtain the fluctuation coefficient . The threshold values of different parameters are different, the threshold value of temperature is set to 0.02 (i.e., 2%), the threshold value of flow rate is set to 0.03, and the threshold value of alloy element concentration is set to 0.05. When exceeds the corresponding threshold value, the system determines that the parameter fluctuation is abnormal. After the abnormal data marking mechanism is started, an abnormal identifier is added to the data record, and it is stored in a separate abnormal database, while normal data is stored in the real-time database for subsequent processing. When the data is transmitted to the MES for Metals platform, the abnormal marking information is sent with the data, and the platform can perform special processing on the abnormal data, such as replacing it with an interpolation method or reducing its weight, to avoid interference with the correlation analysis results. The application of the model makes the data processing of step S3 more scientific, improves the accuracy of subsequent data integration and analysis, and provides a reference for the maintenance of the production line equipment. When a parameter frequently appears fluctuation abnormality, the corresponding sensor or actuator can be prompted for maintenance.

[0057] Preferably, in step S4, when the data analysis module of the MES for Metals platform mines the correlation between different parameters, a zinc layer shedding risk assessment model is introduced, and the formula of the model is: , wherein is the zinc layer shedding risk value, is the weight coefficient, is the influence weight of the jth process parameter, is the actual value of the jth process parameter, is the maximum allowed value of the jth process parameter, is the zinc layer peeling sensitivity factor; by calculating the zinc layer peeling risk value under different process parameter combinations, the parameter range corresponding to the process parameter combination with a risk value higher than the set critical value is identified as the process parameter range affecting zinc layer peeling.

[0058] Specifically, the zinc layer peeling risk assessment model is introduced in step S4, which realizes the comprehensive quantitative evaluation of the zinc layer peeling risk by integrating the process parameters and the sensitivity factors, and provides a more accurate basis for identifying the main influence parameter range. This model combines macro process parameters with micro diffusion reaction characteristics, and comprehensively considers the influence of various factors on zinc layer peeling, making the risk assessment more comprehensive. The weight coefficients α and β in the model are determined according to historical data statistics. For hot galvanizing process, α can be set to 0.6 and β to 0.4, emphasizing the influence of process parameters; for electroplating zinc process, the weight of β can be appropriately increased. The influence weight of each process parameter is determined by the analytic hierarchy process, such as the influence weight of aluminum element concentration is set to 0.3, the influence weight of temperature is set to 0.2, the influence weight of flow rate is set to 0.15, the influence weight of steel strip speed is set to 0.15, the influence weight of coating thickness is set to 0.2, etc., ensuring that the total weight is 1. The actual value of the parameter is taken from the real-time data of S3, the maximum allowed value is determined according to the product quality standard, such as the maximum allowed value of aluminum element concentration is set to 0.3%, the maximum allowed value of temperature is set to 480℃, etc. In implementation, the data analysis module of the MES for Metals platform will calculate the risk value R once every hour (for example). When R value exceeds 0.7, it is determined as a high risk state. By analyzing the R value under different process parameter combinations, the parameter range corresponding to high risk can be identified, such as when the aluminum element concentration exceeds 0.25% and the temperature exceeds 460℃, the R value is generally higher than 0.7. These parameter ranges will be marked as the main influence range. The application of this model makes the identification of the main process parameter range in step S4 more objective and accurate, laying a foundation for subsequent generation of targeted optimization schemes, and at the same time can monitor the risk changes in the production process in real time, and early warning of potential zinc layer peeling problems.

[0059] Preferably, step S4 comprises the following sub-steps: step S4.1, classifying the calculation results of the multi-group diffusion reaction coupling model in step S1, the mapping relationship data obtained by the molecular dynamics-evolutionary neural network analytical algorithm in step S2, and the real-time process parameters collected by the Siemens SIMATIC PCS 7 system in step S3 according to data types, and establishing a comprehensive database including element concentration, temperature, diffusion coefficient, binding energy, and process parameter dimensions; step S4.2, using the association rule algorithm in the MES for Metals platform data analysis module to mine the data in the comprehensive database, calculate the correlation coefficients between different parameters, and select parameter pairs with correlation coefficients greater than 0.6 in absolute value to determine the strong correlation between parameters; step S4.3, constructing a parameter influence network model based on the strong correlation, taking the zinc layer peeling phenomenon as a target node, analyzing the direct and indirect influence of different parameters on the zinc layer peeling through the connection strength between nodes, and generating an influence degree ranking table; step S4.4, selecting the top 8-10 parameters as the calibration influence parameters according to the influence degree ranking table, and determining the final process parameter range affecting the zinc layer peeling in combination with the controllability of the parameters in actual production.

[0060] Specifically, step S4 describes the integration and analysis process of the MES for Metals platform for multi-dimensional data in four sub-steps, aiming to more systematically identify the main process parameter range affecting zinc layer peeling. In step S4.1, when establishing the comprehensive database, the multi-group diffusion reaction data (such as diffusion coefficient, phase diagram data) of S1, the atomic characteristic parameter and mapping relationship data of S2, and the real-time process parameters of S3 are stored in a unified data format, such as the diffusion coefficient data stored in the structure of "element-temperature-concentration-diffusion coefficient", the atomic characteristic parameter stored in the structure of "sample ID-coordination number-bond length-binding energy-evaluation value", and the real-time process parameter stored in the structure of "time-temperature-flow rate-element concentration-steel strip speed-coating thickness", which facilitates subsequent data correlation. In step S4.2, when mining parameter correlation, the minimum support of the association rule algorithm is 0.2, indicating that the probability of the simultaneous occurrence of parameters needs to exceed 20%, and the minimum confidence is 0.7, indicating that the probability of the occurrence of another parameter when a certain parameter occurs needs to exceed 70%, and through calculation, strong association rules such as "aluminum concentration > 0.25% → zinc layer peeling rate > 1%" can be obtained. In step S4.3, when constructing the parameter influence network model, the influence relationship between parameters is represented in the form of nodes and edges, and the weight of the edge represents the influence intensity. The influence score of each node is calculated by the PageRank algorithm, and an influence degree ranking table is generated, such as aluminum concentration, temperature, and steel strip speed, which usually rank in the top three. In step S4.4, when determining the main parameter range, the ranking table is combined with the controllability of the parameters, such as excluding some inherent parameters that are difficult to adjust, and finally selecting 8-10 controllable main parameters. According to the parameter value range corresponding to the high risk, the best control interval is determined, such as aluminum concentration 0.15-0.25%, temperature 440-460°C, etc. The entire process needs to be implemented through the data analysis tools of the platform, such as SPSS Modeler or Python's Scikit-learn library, with a processing period of about once a day, and the analysis is performed on the historical 30-day data to ensure that the identified parameter range meets the recent production conditions, providing clear target parameters for the optimization scheme generated in step S5.

[0061] Preferably, in step S5, when the process optimization module of the MES for Metals platform generates the adjustment scheme for the alloy element ratio of the galvanizing solution, a ratio optimization objective function model is used, and the formula of the model is: wherein, is the optimization objective function value, is the weight coefficient, is the zinc layer peeling risk value, is the adjusted concentration of the i-th alloy element, is the ideal concentration of the i-th alloy element, The energy consumption coefficient of process adjustment; the maximum value of the objective function is solved by genetic algorithm, and the different kinds of alloy element concentration, temperature and flow rate parameter combination corresponding to the maximum value of the objective function is taken as the optimal matching adjustment scheme.

[0062] Specifically, step S5 provides a quantitative optimization target for the process optimization module of the MES for Metals platform by introducing the matching optimization objective function model, so that the generated galvanizing liquid alloy element matching adjustment scheme is more scientific and practical. The model comprehensively considers the zinc layer shedding risk, the degree of alloy element matching deviation from the ideal value and the energy consumption of process adjustment, realizes the collaborative optimization of multiple objectives, and avoids the trade-off caused by single objective optimization. The weight coefficients γ1, γ2 and γ3 in the model need to be determined according to the production priority. When product quality is the primary target, γ1 can be set to 0.5, γ2 to 0.3 and γ3 to 0.2; when energy consumption control is more critical, the weight of γ3 can be increased to 0.4. The lower the value of the zinc layer shedding risk value R, the better the scheme performs in controlling the zinc layer shedding. The ideal concentration of each alloy element needs to be determined according to the product type and quality standard, such as the ideal concentration of zinc for galvanizing products for automobile plates is 97%, the ideal concentration of aluminum is 0.2%, the ideal concentration of nickel is 0.08%, etc. The energy consumption coefficient of process adjustment The energy consumption data corresponding to different parameter adjustment amplitudes is obtained by statistics, such as the energy consumption coefficient of temperature increase of 1°C is 0.02, and the energy consumption coefficient of aluminum element concentration adjustment of 0.01% is 0.015. In implementation, the process optimization module of the MES for Metals platform first takes the main process parameter range determined in S4 as the constraint condition, and then calls the genetic algorithm to solve the maximum value of the objective function. The population size of the genetic algorithm is set to 100, the iteration number is 80, the crossover probability is 0.85, and the mutation probability is 0.06, so as to ensure that the algorithm has enough search space and convergence speed. In each iteration, the objective function value of each scheme is calculated, and the excellent schemes are selected for crossover and mutation operation. When the algorithm converges, the scheme with the maximum objective function value is selected as the optimal matching adjustment scheme, which can effectively reduce the zinc layer shedding risk, make the alloy element matching as close to the ideal value as possible, and control the energy consumption within a reasonable range. Through the application of the model, the adjustment scheme generated in step S5 not only meets the requirements of zinc layer shedding control, but also considers the economy and energy consumption indicators of production, providing more comprehensive decision support for subsequent production control.

[0063] Preferably, step S5 comprises the following sub-steps: step S5.1, setting the adjustment step length of different kinds of alloying elements concentration as 0.01%-0.1%, the temperature adjustment step length as 1-5℃, the flow rate adjustment step length as 0.1-0.5m / s, and constructing the process parameter adjustment space with the process parameter range determined in step S4 as the constraint condition; step S5.2, randomly generating an initial matching adjustment scheme set in the process parameter adjustment space, the number of schemes in the set being 50-100 groups, and each group of schemes including complete alloying element concentration, temperature and flow rate parameters; step S5.3, calling the process simulation module of the MES for Metals platform to simulate the galvanizing process for each group of schemes in the initial scheme set, and calculating the zinc layer adhesion and the falling risk value evaluation indexes corresponding to each group of schemes; step S5.4, screening the initial scheme set according to the evaluation indexes, retaining 30%-50% of the schemes with better evaluation indexes, and performing cross and mutation operations on the schemes to generate new matching adjustment schemes, repeating the simulation and screening process until multiple adjustment schemes meeting the preset evaluation standard are obtained.

[0064] Specifically, step S5 clearly illustrates the specific process of generating the alloying element ratio adjustment scheme of the galvanizing solution by the process optimization module of the MES for Metals platform through four sub-steps, ensuring the systematicness and operability of the scheme generation process. In step S5.1, when constructing the process parameter adjustment space, the main process parameter range determined in S4 is taken as the boundary, and the adjustment step of each parameter is set according to the control accuracy requirement. The alloying element concentration adjustment step is set to 0.01%-0.1%, which can ensure the accuracy of the adjustment and will not cause excessive calculation due to too small step size; the temperature adjustment step is 1-5°C, which can adapt to the requirements of different processes on temperature sensitivity; the flow rate adjustment step is 0.1-0.5 m / s, which takes into account the actual accuracy of pump body adjustment. In step S5.2, when generating the initial ratio adjustment scheme set, a random sampling method is used to select 50-100 schemes in the adjustment space, and each scheme needs to include complete alloying element concentration, temperature and flow rate parameters to ensure the diversity and representativeness of the scheme and avoid the initial scheme being too concentrated in a certain area and affecting the optimization effect. In step S5.3, when simulating and evaluating the initial scheme, the process simulation module of the MES for Metals platform calls the built-in galvanizing process prediction model, which is based on historical production data and physical and chemical principles. The model can predict key indicators such as zinc layer adhesion and peeling risk value according to the input parameters, and the simulation time is set to 10 minutes for each scheme to control the calculation time while ensuring accuracy. In step S5.4, when screening and iterative optimization, first, the schemes with the top 30%-50% evaluation indicators are retained, then these schemes are subjected to crossover operation (such as exchanging part of the parameters of two schemes) and mutation operation (such as randomly changing the value of a parameter), new schemes are generated and supplemented to the scheme set, and the simulation evaluation and screening process is repeated until the evaluation indicators of the schemes reach the preset standard (such as the zinc layer peeling risk value being less than 0.5) or the iteration number reaches the set upper limit (usually 50 times). Through this series of steps, the final multiple adjustment schemes not only cover different parameter combinations, but also ensure the effectiveness of each scheme in controlling zinc layer peeling, providing sufficient alternative space for selecting the optimal scheme in step S6.

[0065] Preferably, the step S6 comprises the following sub-steps: step S6.1, the control module of the Siemens SIMATIC PCS 7 system receives the optimal matching adjustment scheme sent by the MES for Metals platform, analyzes the different kinds of alloy element concentration adjustment values, temperature and flow rate matching parameters included in the scheme, and converts them into digital signals recognizable by the system; step S6.2, the control module sends control instructions to the galvanizing liquid alloy element adding device according to the converted digital signals, adjusts the valve opening and adding frequency of the adding device to adjust the concentration of different kinds of alloy elements; step S6.3, corresponding control signals are sent to the temperature control device and the flow rate control device at the same time, the heating power and the pump body speed are adjusted to make the temperature and flow rate of the galvanizing liquid reach the values set by the adjustment scheme; step S6.4, during the regulation process, the control module receives the feedback signals of each executing mechanism in real time, compares them with the target values of the adjustment scheme, dynamically adjusts the control instructions according to the deviation values, and stops until the different process parameters are stabilized in the target range.

[0066] Specifically, step S6 describes in detail the process of the Siemens SIMATIC PCS 7 system converting the optimal matching adjustment scheme into control instructions and achieving dynamic regulation through four sub-steps, ensuring that the process optimization scheme can be accurately implemented. In step S6.1, when analyzing and converting the optimal scheme, the control module of the Siemens SIMATIC PCS 7 system receives the optimal scheme sent by the MES for Metals platform, and first analyzes and converts the alloy element concentration adjustment value, temperature, and flow rate matching parameters in the scheme into digital signals recognizable by the system. Unit conversion and range conversion are required during the conversion process, such as converting the percentage value of aluminum element concentration to a 4-20 mA current signal and converting the temperature value to a corresponding voltage signal. In step S6.2, when regulating the alloy element adding device, the control module calculates the valve opening and adding frequency according to the converted digital signal. The valve opening has a linear relationship with the element concentration adjustment value, such as adjusting the valve opening from 25% to 35% when the aluminum element concentration needs to be increased from 0.18% to 0.22%. The adding frequency is set according to the concentration change rate, usually 3-8 times per minute, to ensure the stability of the concentration adjustment. In step S6.3, when regulating the temperature and flow rate, the temperature control device adjusts the target temperature by adjusting the power of the heating pipe, such as setting the heating power to 90% when the target temperature is 460℃, and using the PID control algorithm for real-time adjustment, with a proportional coefficient of 0.6, an integral time of 8s, and a differential time of 1.5s. The flow rate control device adjusts the flow rate by changing the motor frequency of the circulating pump, such as adjusting the motor frequency from 50Hz to 55Hz when the target flow rate is 2.3m / s, to ensure that the flow rate is stable within the target value ±0.1m / s range. In step S6.4, when dynamically feeding back and adjusting, the control module receives feedback signals (such as actual concentration, temperature, and flow rate values) from each execution mechanism every 200ms, compares them with the target values of the scheme, calculates the deviation, and adjusts the control instructions in time when the deviation exceeds the allowed range (such as a concentration deviation of ±0.02% and a temperature deviation of ±2℃), until the deviation returns to the allowed range. Through this series of steps, accurate conversion from the optimal scheme to production execution and dynamic regulation are achieved, ensuring that the alloy element matching of the galvanizing liquid and related process parameters are always in the optimal state, effectively controlling the occurrence of zinc layer peeling.

[0067] In the CALPHAD-based multi-component diffusion reaction coupling model of the application, CALPHAD stands for "Calculation of Phase Diagrams". Through the phase diagram calculation technology, the model combines the thermodynamic and kinetic parameters of multiple alloy elements to construct the phase diagram and diffusion equation of the multi-component system, so as to analyze the interaction and phase change rule of elements in the galvanizing process. The CALPHAD-based multi-component diffusion reaction coupling model is a thermodynamic and kinetic integrated model for analyzing the diffusion and reaction characteristics of multiple alloy elements in the galvanizing process. By collecting and evaluating the thermodynamic parameters (such as enthalpy, entropy, Gibbs free energy, etc.) and kinetic parameters (such as diffusion coefficient, reaction rate constant, etc.) of different alloy elements, the phase diagram and diffusion equation of the multi-component system are established to describe the interaction and phase change rule of zinc, iron, aluminum, nickel and other elements under different temperature and concentration conditions. The implementation process of the model first needs to determine the types of alloy elements involved in the reaction and the initial concentration range, such as zinc concentration 95-98%, aluminum concentration 0.1-0.3%, etc., and then input the process parameters such as temperature (420-480℃), concentration gradient (0.05-0.2% / mm), etc. Using the database and calculation module in the CALPHAD method, the diffusion coefficient, reaction activation energy and other key parameters of each element are calculated, and then the multi-component diffusion reaction phase diagram is constructed to determine the conditions and stability of the formation of intermetallic compounds between elements. The role of the model is to provide basic data support for zinc layer shedding control. By quantitatively analyzing the influence of different element ratios on diffusion rate and reaction degree, the element concentration range that may cause the decrease of zinc layer adhesion is identified, such as high aluminum concentration may form brittle compounds. It breaks through the limitations of traditional single element analysis, realizes the systematic study of multi-component complex system, provides a theoretical basis for subsequent process parameter optimization, and helps to fundamentally understand the micro mechanism of zinc layer shedding, and improves the scientificity and accuracy of control method.

[0068] The molecular dynamics-evolutionary neural network analysis algorithm of the application is a comprehensive analysis tool combining micro atomic simulation and intelligent algorithm, which is used to explore the internal correlation between the atomic motion of alloying elements in galvanizing liquid and the bonding performance of zinc layer. The molecular dynamics part simulates the motion trajectory of atoms at the nanometer scale by constructing an atomic model containing zinc, iron and other alloying elements, calculates micro characteristic parameters such as interface binding energy, atomic coordination number and bond length distribution, so as to reflect the bonding state of the zinc layer and the substrate interface; the evolutionary neural network part takes these micro parameters as input, establishes the mapping relationship between the micro parameters and the bonding performance of the zinc layer (such as bonding force and shedding risk) through the training and learning of the neural network, and optimizes the network structure and parameters by using the genetic algorithm to improve the prediction accuracy. The implementation of the algorithm first needs to construct an atomic model (containing 10000-50000 atoms, simulation box size 10nm×10nm×20nm), set the simulation time step to 1fs and the total simulation time to 100ps, and record the atomic motion data; then extract the micro characteristic parameters, input them into the neural network containing 8-12 input nodes, 2 layers of hidden layers (15 neurons per layer), train 500-1000 groups of sample data (1000 iterations, learning rate 0.01), and verify 200 groups of data to control the error within 5%. The role is to reveal the influencing factors of the bonding performance of the zinc layer from the atomic scale, and to provide quantitative indicators at the micro level for evaluating the shedding risk of the zinc layer. It makes up for the defects that the macro process parameter analysis cannot reflect the micro mechanism, realizes the cross-scale correlation from atomic behavior to macro performance, provides more detailed guidance for process optimization, and helps to develop more targeted zinc layer shedding control strategies.

[0069] In the Siemens SIMATIC PCS 7 system of the application, SIMATIC is the abbreviation of German "Siemens Automatic Control", which means Siemens automatic control; PCS is the abbreviation of "Process Control System", that is, process control system, so the system is called Siemens process control system 7 in full. The Siemens SIMATIC PCS 7 system is an integrated process control system for real-time data acquisition, processing and control of galvanizing production line. The system is composed of sensors, data acquisition modules, communication modules, control modules and human-computer interface, etc., and can realize real-time monitoring and regulation of key process parameters such as galvanizing liquid temperature, flow rate, alloy element concentration, steel strip running speed, coating thickness, etc. The implementation process is as follows: firstly, corresponding sensors are installed at key positions of the production line, such as platinum resistance temperature sensor (measurement range 400-500℃, accuracy ±0.5℃), electromagnetic flowmeter (0.5-5m / s, accuracy ±0.5%), X-ray fluorescence spectrometer (detection accuracy 0.01%) and the like, the sensors convert physical quantities into electrical signals and transmit them to the data acquisition module; the system collects data at a sampling frequency of 10Hz, and after filtering, abnormal value detection (such as temperature exceeding 420-480℃ is marked as abnormal) and other treatments, the data is transmitted in real time to the data processing unit and the MES for Metals platform through the PROFINET communication protocol; the control module generates control instructions through PID algorithm according to the received optimization scheme, and adjusts the alloy element adding device, heating device, circulating pump and other actuators. The function of the system is to build a data acquisition and control closed loop of the production process, to ensure real-time monitoring and accurate regulation of process parameters, and to provide a stable production operation environment for zinc layer shedding control. Real-time and automatic management of production data is realized, replacing traditional manual monitoring and operation, improving the accuracy and timeliness of data, and providing reliable basic data for data analysis and process optimization of the upper platform. It is the key link connecting theoretical analysis and actual production.

[0070] In the MES for Metals platform of the present application, MES is the abbreviation of "Manufacturing Execution System", i.e. manufacturing execution system. The platform is specially designed for the metal manufacturing industry, aiming to realize multi-dimensional data integration, analysis and process optimization, and is an important tool for production management of metal manufacturing enterprises, which can integrate various model calculation results and real-time production data, and help enterprises improve production efficiency and product quality. The MES for Metals platform is a manufacturing execution system specially designed for the metal manufacturing industry, which undertakes the core functions of multi-dimensional data integration, analysis and process optimization in the present application. The platform can integrate the calculation results from the multi-component diffusion reaction coupling model based on CALPHAD, the micro-characteristic parameters and mapping relationship of the molecular dynamics-evolutionary neural network analysis algorithm, and the real-time process parameters of the Siemens SIMATIC PCS 7 system, forming a unified production database. The implementation process first receives and stores various data through a data interface (such as an OPC interface), and classifies management according to data types (such as diffusion parameters, micro-characteristics, and real-time processes); then uses the data analysis module of the platform (including algorithms such as association rule analysis and principal component analysis) to mine the correlation between different parameters, such as the relationship between aluminum concentration and zinc layer shedding rate, and identifies the main process parameter range affecting zinc layer shedding (such as aluminum concentration 0.15-0.25%, temperature 440-460°C); the process optimization module generates multiple alloy element ratio adjustment schemes based on these parameter ranges using optimization methods such as genetic algorithm, and evaluates the feasibility of the schemes. The role of the platform is to realize the deep integration and intelligent analysis of data, provide decision support for process parameter optimization, generate targeted adjustment schemes, and guide the dynamic regulation and control of the production process. It breaks down the data barriers between different models and systems, realizes integrated management from micro-computing to macro-control, improves the scientificity and efficiency of process optimization, and at the same time, generates multiple optimization schemes, enhances the adaptability to complex production conditions, and provides a strong guarantee for continuously improving the control effect of zinc layer shedding.

[0071] As shown in Figure 2 A zinc layer shedding control method based on galvanizing liquid alloy element ratio, the method is realized by different units, including:

[0072] Multi-component diffusion reaction parameter calculation unit, which is connected with the molecular dynamics-evolutionary neural network analysis unit through a data transmission bus, for transmitting the calculated multi-component diffusion reaction parameters to the molecular dynamics-evolutionary neural network analysis unit;

[0073] Molecular dynamics-evolutionary neural network analysis unit, which is connected with the production line real-time data acquisition and processing unit through a wireless communication module, for sending the analyzed atomic scale characteristic parameters and mapping relationship to the production line real-time data acquisition and processing unit;

[0074] A production line real-time data acquisition and processing unit is connected with the metal manufacturing execution platform data correlation analysis unit through an Ethernet connection, and is used to transmit the processed real-time process parameters to the metal manufacturing execution platform data correlation analysis unit;

[0075] A metal manufacturing execution platform data correlation analysis unit is connected with the process parameter optimization scheme generation unit through a fiber communication connection, and is used to send the identified process parameter range to the process parameter optimization scheme generation unit;

[0076] A process parameter optimization scheme generation unit is connected with the production process dynamic regulation and control execution unit through an industrial bus, and is used to transmit the generated optimal matching adjustment scheme to the production process dynamic regulation and control execution unit;

[0077] A production process dynamic regulation and control execution unit is connected with the multi-component diffusion reaction parameter calculation unit through a feedback signal line, and is used to feed back the regulated process parameters to the multi-component diffusion reaction parameter calculation unit.

[0078] A zinc layer peeling control method based on galvanizing liquid alloy element matching, which overcomes the defects of insufficient consideration of multi-component complex correlation in the prior art through multi-model collaborative analysis. Through a multi-component diffusion reaction coupling model based on CALPHAD, the diffusion rules and reaction characteristics of multiple alloy elements under different temperature and concentration conditions can be comprehensively captured, and the influence mechanism of element interaction on zinc layer formation can be accurately constructed. Combined with a molecular dynamics-evolutionary neural network analysis algorithm, the internal correlation between microscopic features such as atomic motion trajectory and interface binding energy and zinc layer bonding performance is deeply mined, forming a multi-scale analysis system from macro to micro, so that process adjustment is no longer limited to local parameters, greatly improving the pertinence and accuracy of regulation and control.

[0079] By constructing a data closed-loop management system, the method solves the problem of disconnection of data at each link in the prior art, and cannot realize dynamic closed-loop management. The Siemens SIMATIC PCS 7 system is used to collect key process parameters such as galvanizing liquid temperature, alloy element concentration, and steel strip speed in real time, ensuring the timeliness and integrity of the data; the MES for Metals platform is used to integrate and analyze multi-component diffusion reaction data, atomic scale characteristic parameters, and real-time process data, realize parameter correlation mining and process scheme optimization; and the control system is used to convert the optimized scheme into an execution instruction, forming a complete closed loop of "data acquisition-analysis optimization-regulation and execution-feedback correction", so that the zinc layer peeling control can dynamically respond to changes in production conditions, and the stability of the control effect is significantly improved.

[0080] The method realizes intelligent regulation and control, reduces the dependence on experience adjustment, converts the complex process optimization process into a standardized and automated operation process through multi-model cooperation and system integration, and avoids subjectivity and limitations caused by manual experience. From multi-component diffusion reaction calculation to alloy element ratio scheme generation, to dynamic regulation and control of production parameters, all are automatically completed through algorithm models and system modules, not only improving the efficiency of process adjustment, but also maintaining stable regulation and control accuracy in complex production environment, greatly reducing the probability of zinc layer falling off, and significantly improving the consistency and reliability of the quality of galvanized products.

[0081] In the description of the present application, it should be noted that unless otherwise explicitly specified and limited, the terms "arrangement", "installation", "connection", "connection", "fixing" should be understood broadly, for example, it can be fixed connection, or detachable connection, or integral connection; it can be mechanical connection, or electrical connection; it can be direct connection, or indirect connection through intermediate medium, or internal communication of two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0082] Although the embodiments of the present application have been shown and described, those skilled in the art can understand that various equivalent changes, modifications, replacements and variations of the embodiments can be made without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended claims and their equivalent scope.

Claims

1. A method for controlling flaking of a zinc layer based on the proportioning of liquid alloy elements for galvanizing, characterized by, The method comprises: Step S1, calculating the diffusion coefficient and reaction kinetics parameter of different types of alloying elements in the galvanizing process under different temperature and concentration gradient conditions by a CALPHAD-based multi-component diffusion reaction coupling model, constructing a diffusion reaction phase diagram of a multi-component system, and determining the influence law of the interaction between different types of alloying elements on the formation of a zinc layer; Step S2, simulating and analyzing the atomic motion trajectory, interface binding energy and zinc layer crystal structure evolution process of the alloying elements in the galvanizing solution by using a molecular dynamics-evolution neural network analysis algorithm, extracting the calibrated atomic scale characteristic parameters affecting the bonding force of the zinc layer and the substrate, and taking the parameters as input variables of the analysis algorithm, and establishing a mapping relationship between the characteristic parameters and the bonding performance of the zinc layer through the training process of the neural network; Step S3, collecting real-time running data of the galvanizing production line by using the Siemens SIMATIC PCS 7 system, including galvanizing solution temperature, flow rate, alloying element concentration, steel strip running speed, and coating thickness process parameters, and transmitting the collected data to the data processing unit for real-time storage and preliminary analysis through the built-in communication module of the system; Step S4, importing the multi-component diffusion reaction calculation results obtained in step S1, the atomic scale characteristic parameters and the mapping relationship obtained in step S2, and the real-time process parameters collected in step S3 into the MES for Metals platform, and mining the correlation between different parameters through the data analysis module of the platform to identify the process parameter range affecting the peeling of the zinc layer.

2. The method for controlling zinc layer spalling based on the proportioning of liquid alloy elements for galvanizing according to claim 1, characterized in that, The method further comprises: Step S5, generating a multi-component galvanizing solution alloying element ratio adjustment scheme through the process optimization module of the MES for Metals platform based on the process parameter range determined in step S4, which includes concentration adjustment values of different types of alloying elements, corresponding temperature and flow rate matching parameters; Step S6, converting the optimal ratio adjustment scheme generated in step S5 into specific control instructions through the control module of the Siemens SIMATIC PCS 7 system, and issuing the instructions to the execution mechanism of the galvanizing production line to dynamically control the galvanizing solution alloying element ratio and related process parameters; In step S1, a zinc layer peeling sensitivity factor calculation model is introduced in the CALPHAD-based multi-component diffusion reaction coupling model calculation process, and the model formula is: wherein, is a zinc layer peeling sensitivity factor, is a model correction coefficient, is an actual concentration of the i th alloying element, is a reference concentration of the i th alloying element, is a diffusion coefficient of the i th alloying element, is a diffusion activation energy of the i th alloying element, is a gas constant, is an absolute temperature, is a Gibbs free energy change of the multi-component reaction; the zinc layer peeling sensitivity factor under different alloy element ratios is calculated through the model, which is used as a quantitative index for judging the influence degree of the multi-component diffusion reaction on the zinc layer peeling, and the construction process of the multi-component diffusion reaction phase diagram is optimized according to the index.

3. The method for controlling zinc layer peeling based on the proportioning of liquid alloy elements for galvanizing according to claim 1, characterized in that, Step S2 comprises the following steps: Step S2.1, simulating the adsorption, diffusion and deposition process of alloying element atoms in the galvanizing solution on the substrate surface by using the molecular dynamics method, setting the simulation time step to 1fs to 10fs, covering the actual galvanizing process temperature range, recording the spatial coordinates, velocities and interaction forces of different atoms at different times, and generating an atomic motion trajectory file; Step S2.2, based on the generated atomic motion trajectory file, calculating the atomic coordination number, bond length distribution and binding energy of the zinc layer and the substrate interface region, determining the micro area with weak interface bonding force and the corresponding atomic arrangement characteristics by analyzing the variation law of the parameters; Step S2.3, the interfacial binding energy, atomic coordination number atomic scale characteristic parameters are standardized and input into the input layer of the evolutionary neural network, the hidden layer of the neural network uses 10-20 neurons, the output layer is the zinc layer bonding performance evaluation value, and the initial weight and threshold of the neural network are optimized by the genetic algorithm; Step S2.4, the trained evolutionary neural network is tested by using the verification data set, and the structure parameters of the neural network are adjusted according to the test results until the prediction error of the network for the zinc layer bonding performance is controlled within the preset range.

4. The method for controlling zinc layer spalling based on the proportioning of liquid alloy elements for galvanizing according to claim 1, characterized in that, In step S3, the Siemens SIMATIC PCS 7 system adopts a process parameter fluctuation coefficient calculation model in the collection and processing process of the real-time running data of the galvanizing production line, and the model formula is: Wherein, is the fluctuation coefficient of the jth process parameter, is the data collection times, is the kth collection value of the jth process parameter, is the average value of the jth process parameter; the fluctuation coefficients of different process parameters are calculated through the model, when the fluctuation coefficient exceeds the set threshold value, the Siemens SIMATIC PCS 7 system starts the abnormal data marking mechanism, and the corresponding abnormal data is stored separately from the normal data, and abnormal marking information is attached in the data transmission process.

5. The method for controlling the peeling of a zinc layer based on the proportioning of liquid alloy elements for galvanizing according to claim 1, characterized in that, In step S4, the data analysis module of the MES for Metals platform introduces a zinc layer peeling risk assessment model when mining the correlation between different parameters, and the formula of the model is: wherein, is the zinc layer peeling risk value, is the weight coefficient, is the influence weight of the jth process parameter, is the actual value of the jth process parameter, is the maximum allowable value of the jth process parameter, is the zinc layer peeling sensitivity factor; by calculating the zinc layer peeling risk value under different process parameter combinations, the parameter range corresponding to the process parameter combination with a risk value higher than the set critical value is identified as the process parameter range affecting the zinc layer peeling.

6. The method for controlling zinc layer peeling based on the proportioning of liquid alloy elements for galvanizing according to claim 1, characterized in that, Step S4 includes the following sub-steps: Step S4.1, the calculation results of the multi-component diffusion reaction coupling model in step S1, the mapping relationship data obtained by the molecular dynamics-evolutionary neural network analytical algorithm in step S2, and the real-time process parameters collected by the SIMATIC PCS 7 system in step S3 are classified according to the data type, and a comprehensive database including element concentration, temperature, diffusion coefficient, binding energy and process parameter dimension is established; Step S4.2, using the association rule algorithm in the MES for Metals platform data analysis module, the data in the comprehensive database is mined, the correlation coefficient between different parameters is calculated, and the parameter pairs with absolute value greater than 0.6 are selected to determine the strong correlation between parameters; Step S4.3, based on the strong correlation relationship, a parameter influence network model is constructed, the zinc layer peeling phenomenon is taken as the target node, the direct and indirect influence degree of different parameters on the zinc layer peeling is analyzed through the connection strength between nodes, and an influence degree ranking table is generated; Step S4.4, according to the influence degree ranking table, the top 8-10 parameters are selected as the calibration influence parameters, and combined with the controllability of the parameters in actual production, the final process parameter range affecting the zinc layer peeling is determined.

7. The method for controlling the peeling of a zinc layer based on the proportioning of liquid alloy elements for galvanizing according to claim 2, characterized in that, In step S5, when the process optimization module of the MES for Metals platform generates the galvanizing liquid alloy element ratio adjustment scheme, a ratio optimization objective function model is used, and the formula of the model is: wherein, is an optimization objective function value, is a weight coefficient, is a zinc layer peeling risk value, is an adjusted concentration of the i th alloying element, is an ideal concentration of the i th alloying element, is an energy consumption coefficient of process adjustment; the maximum value of the objective function is solved by a genetic algorithm, and the combination of the concentrations of different kinds of alloying elements, temperature and flow rate parameters corresponding to the maximum objective function value is taken as the optimal matching adjustment scheme.

8. The method for controlling zinc layer spalling based on the proportioning of liquid alloy elements for galvanizing according to claim 2, characterized in that, Step S5 includes the following sub-steps: Step S5.1, setting the adjustment step length of different kinds of alloy element concentrations to 0.01%-0.1%, the temperature adjustment step length to 1-5℃, and the flow rate adjustment step length to 0.1-0.5m / s, and constructing a process parameter adjustment space with the process parameter range determined in step S4 as a constraint condition; Step S5.2, randomly generating an initial matching adjustment scheme set in the process parameter adjustment space, the number of schemes in the set being 50-100 groups, and each group of schemes including complete alloy element concentration, temperature and flow rate parameters; Step S5.3, calling the process simulation module of the MES for Metals platform to simulate the galvanizing process for each group of schemes in the initial scheme set, and calculating the zinc layer adhesion and falling risk value evaluation indexes corresponding to each group of schemes; Step S5.4, screening the initial scheme set according to the evaluation indexes, retaining 30%-50% of the schemes with better evaluation indexes, and performing cross and mutation operations on the schemes to generate new matching adjustment schemes, repeating the simulation and screening processes until multiple adjustment schemes meeting the preset evaluation standards are obtained.

9. The method for controlling the peeling of a zinc layer based on the proportioning of liquid alloy elements for galvanizing according to claim 2, characterized in that, Step S6 includes the following sub-steps: Step S6.1, the control module of the SIMATIC PCS 7 system receives the optimal matching adjustment scheme sent by the MES for Metals platform, analyzes the different kinds of alloy element concentration adjustment values, temperature and flow rate matching parameters included in the scheme, and converts them into digital signals recognizable by the system; Step S6.2, the control module sends control instructions to the galvanizing liquid alloy element adding device according to the converted digital signals, adjusts the valve opening and adding frequency of the adding device to adjust the concentrations of different kinds of alloy elements; Step S6.3, at the same time, corresponding control signals are sent to the temperature control device and the flow rate control device to adjust the heating power and pump body speed so that the temperature and flow rate of the galvanizing liquid reach the values set in the adjustment scheme; Step S6.4, during the regulation process, the control module receives the feedback signals of each execution mechanism in real time, compares them with the target values of the adjustment scheme, dynamically adjusts the control instructions according to the deviation values, and finally stabilizes the different process parameters in the target range.

10. The method for controlling the peeling of a zinc layer based on the proportioning of liquid alloy elements for galvanizing according to any one of claims 1 to 9, characterized in that, The method is realized by different units, including: A multi-component diffusion reaction parameter calculation unit connected with the molecular dynamics-evolutionary neural network analysis unit through a data transmission bus, used to transmit the calculated multi-component diffusion reaction parameters to the molecular dynamics-evolutionary neural network analysis unit; A molecular dynamics-evolutionary neural network analysis unit connected with the production line real-time data acquisition and processing unit through a wireless communication module, used to send the analyzed atomic-scale feature parameters and mapping relationships to the production line real-time data acquisition and processing unit; A production line real-time data acquisition and processing unit connected with the metal manufacturing execution platform data correlation analysis unit through an Ethernet, used to transmit the acquired and processed real-time process parameters to the metal manufacturing execution platform data correlation analysis unit; A metal manufacturing execution platform data correlation analysis unit is connected with the process parameter optimization scheme generation unit through a fiber communication connection, and is used to send the identified process parameter range to the process parameter optimization scheme generation unit; A process parameter optimization scheme generation unit is connected with the production process dynamic regulation and control execution unit through an industrial bus, and is used to transmit the generated optimal matching adjustment scheme to the production process dynamic regulation and control execution unit; The production process dynamic regulation and control execution unit is connected with the multi-component diffusion reaction parameter calculation unit through a feedback signal line, and is used to feed back the regulated process parameters to the multi-component diffusion reaction parameter calculation unit.

Images