A hot-dip galvanizing coating intelligent control system

Abstract

The application relates to the technical field of automatic control, and particularly discloses a hot galvanizing coating intelligent control system, which aims to solve the problems of large coating thickness fluctuation, high zinc consumption and poor quality stability. The system comprises a multi-modal sensing acquisition module, a dynamic coupling modeling module, a distributed parallel decision module and a multi-target collaborative execution module. Through real-time acquisition of process data, a nonlinear state space model is established, multi-target optimization decisions are generated in parallel, and the final control instruction is output based on dynamic weight fusion, so that the accurate control of coating quality and the improvement of production energy efficiency are realized.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of automatic control, and particularly relates to a hot galvanizing coating intelligent control system. BACKGROUND

[0002] In the field of metal material surface treatment technology, hot galvanizing, as an efficient and economical corrosion protection process, is widely used in the construction, automobile, home appliance and other industries. This process forms a zinc-iron alloy layer on the surface of the steel substrate, effectively isolating the substrate from the corrosive medium, and significantly prolonging the service life of the material. The hot galvanizing process involves multiple processes such as pretreatment, galvanizing, cooling, etc., and the accurate control of its process parameters has a decisive influence on the quality of the coating and production efficiency.

[0003] Among them, the intelligent control of the hot galvanizing coating process is the key technical direction to improve the uniformity of the coating, reduce zinc consumption and realize green production. This technology aims to realize accurate regulation and control of the coating thickness and microstructure by real-time monitoring and dynamic adjustment of key process parameters such as galvanizing temperature, zinc immersion time, air knife pressure, etc., so as to meet the differentiated needs of different application scenarios for coating performance.

[0004] The existing technology mostly adopts single-loop regulation systems based on traditional PID control, which is difficult to cope with the complex dynamic characteristics such as strong nonlinearity, large time delay and multivariable coupling in the galvanizing process; the collaborative control effect between air knife pressure, strip speed and zinc liquid temperature is not good, which easily leads to large fluctuations in coating thickness, over-thickness at the edge or zinc wave defects.

[0005] At the same time, the existing system lacks the ability to deeply mine the mapping relationship between process parameters and coating quality, and cannot adaptively optimize according to real-time working conditions, resulting in high zinc consumption and poor quality stability.

[0006] In addition, the traditional control system responds slowly to changes in production rhythm, and easily produces transition section waste during specification switching or speed adjustment, which seriously affects production efficiency and cost control.

[0007] Therefore, it is urgent to develop a hot galvanizing coating intelligent control system that can realize multi-parameter intelligent collaborative optimization and has self-learning and adaptive ability. SUMMARY

[0008] The technical problem to be solved by the present application is to overcome the defects of large fluctuations in coating thickness, high zinc consumption, poor quality stability and insufficient production rhythm adaptability caused by strong nonlinearity, large time delay and multivariable coupling in the existing hot galvanizing coating control process. The purpose of the present application is to provide a hot galvanizing coating intelligent control system, which realizes accurate and stable control of coating quality and significant improvement of production efficiency by constructing a full-process intelligent control architecture that integrates multi-modal perception, dynamic coupling modeling, distributed parallel decision-making and multi-objective collaborative optimization.

[0009] The technical scheme of the present application is a hot galvanizing coating intelligent control system, comprising:

[0010] A multi-modal sensor acquisition module is configured to acquire in real time the strip steel speed, zinc liquid temperature, air knife pressure, air knife distance, strip steel tension, and post-coating plate surface temperature distribution data during the production line operation;

[0011] A dynamic coupling modeling module is coupled to the output of the multi-modal sensor acquisition module and is configured to establish a nonlinear state space model describing the dynamic coupling relationship between the air knife pressure, strip steel speed, zinc liquid temperature, and coating thickness based on the real-time data provided by the multi-modal sensor acquisition module;

[0012] A distributed parallel decision module is coupled to the output of the dynamic coupling modeling module and is configured to receive the state prediction information output by the dynamic coupling modeling module and perform parallel calculation based on a plurality of preset performance indicators to generate a plurality of sets of optimized decision schemes for the air knife pressure set value, strip steel speed set value, and zinc liquid temperature set value;

[0013] A multi-target cooperative execution module is coupled to the output of the distributed parallel decision module and is configured to comprehensively evaluate and fuse the plurality of sets of optimized decision schemes output by the distributed parallel decision module to generate final execution instructions and issue them to the corresponding execution mechanisms.

[0014] Preferably, the multi-modal sensor acquisition module comprises a speed detection unit, a temperature detection unit, a pressure detection unit, a distance detection unit, a tension detection unit, and a face array infrared temperature measurement unit;

[0015] The speed detection unit uses a laser speed meter to measure the strip steel running speed in real time;

[0016] The temperature detection unit uses an immersion thermocouple array to measure the zinc liquid temperature in different regions inside the zinc pot;

[0017] The pressure detection unit uses a high-precision pressure transmitter to monitor the pressure values of the air knife gas supply main pipe and the branch pipes on both sides;

[0018] The distance detection unit uses a laser displacement sensor to detect the distance between the air knife lip and the strip steel surface in real time;

[0019] The tension detection unit obtains real-time tension data of the strip steel through the tension sensor attached to the tension roller;

[0020] The face array infrared temperature measurement unit is installed at the entrance of the post-coating cooling section and is configured to obtain two-dimensional temperature field distribution data of the strip steel surface.

[0021] Preferably, the nonlinear state space model constructed by the dynamic coupling modeling module contains four state variables, three control inputs, and two main outputs.

[0022] The four state variables are respectively coating thickness deviation, strip speed variation rate, zinc liquid temperature gradient and gas knife pressure fluctuation amount;

[0023] The three control inputs are respectively gas knife pressure set value, strip speed set value and zinc liquid temperature set value;

[0024] The two main outputs are respectively predicted coating thickness and predicted coating uniformity index.

[0025] Preferably, the model establishment process of the dynamic coupling modeling module comprises:

[0026] Based on historical production data and real-time sensing data, the recursive least squares method is used to identify the model parameters online;

[0027] By introducing an adaptive mechanism with a forgetting factor, it is ensured that the model can track the slow time-varying characteristics of the process conditions;

[0028] And the Kalman filtering technology is used to make optimal estimation on the model state to suppress the interference of measurement noise.

[0029] Preferably, the distributed parallel decision module is composed of three parallel running decision sub-modules, which are respectively energy efficiency optimal decision sub-module, quality optimal decision sub-module and stability optimal decision sub-module;

[0030] The energy efficiency optimal decision sub-module performs optimization calculation with the minimum of unit area zinc consumption as the objective function;

[0031] The quality optimal decision sub-module performs optimization calculation with the minimum of the weighted sum of coating thickness deviation and uniformity index as the objective function;

[0032] The stability optimal decision sub-module performs optimization calculation with the minimum of the square sum of variation rates of key process parameters as the objective function.

[0033] Preferably, each decision sub-module adopts the model predictive control framework, and generates the optimization sequence of the corresponding gas knife pressure set value, strip speed set value and zinc liquid temperature set value by solving a rolling optimization problem in a finite time domain.

[0034] Preferably, the decision fusion mechanism of the multi-objective collaborative execution module comprises:

[0035] A dynamic weight coefficient is respectively assigned to the decision schemes output by the energy efficiency optimal decision sub-module, the quality optimal decision sub-module and the stability optimal decision sub-module; the dynamic weight coefficient is adaptively adjusted according to the current production stage and the target priority;

[0036] In the normal steady production stage, the weight coefficient of the quality optimal decision sub-module is set to 0.5, and the weight coefficients of the energy efficiency optimal decision sub-module and the stability optimal decision sub-module are each set to 0.25;

[0037] In the specification switching or acceleration / deceleration transition stage, the weight coefficient of the stability optimal decision sub-module is raised to 0.5, the weight coefficient of the quality optimal decision sub-module is adjusted to 0.3, and the weight coefficient of the energy efficiency optimal decision sub-module is adjusted to 0.2.

[0038] Finally, the linear weighted average of the weighted multiple groups of decision schemes is performed to calculate the final gas knife pressure setting value, strip speed setting value and zinc liquid temperature setting value.

[0039] Preferably, it further comprises a process knowledge base and an adaptive learning engine.

[0040] The process knowledge base is used to store historical optimal process parameter combinations, process window limits of different steel specifications, and characteristic data and coping strategies of various defects.

[0041] The adaptive learning engine is used to iteratively optimize and update the parameters of the dynamic coupling modeling module, the weight coefficients of the objective functions of the distributed parallel decision module and the dynamic weight coefficients of the multi-objective collaborative execution module based on the actual results of each production task.

[0042] Preferably, the update strategy of the adaptive learning engine adopts a reinforcement learning algorithm based on gradient descent, with the maximization of cumulative production benefit as the long-term optimization goal.

[0043] Preferably, a command smooth transition unit is arranged between the multi-objective collaborative execution module and the underlying execution mechanism.

[0044] The command smooth transition unit is used to perform ramp processing on the step change of the setting value, and the ramp rate is adaptively calculated according to the strip speed and the specification change amount to ensure smooth action of the execution mechanism and avoid impact on the strip running and the plating layer formation process.

[0045] In summary, the present application includes at least one of the following beneficial technical effects:

[0046] 1. The present application fundamentally solves the control problems caused by strong nonlinearity, large time delay and multivariable coupling in the hot galvanizing process by constructing a full-process intelligent control architecture integrating multi-modal perception, dynamic coupling modeling, distributed parallel decision and multi-objective collaborative optimization, and the dynamic coupling modeling module can accurately describe the complex dynamic relationship between key process parameters and plating quality indicators, providing a reliable model basis for advanced control strategies.

[0047] 2. The distributed parallel decision mechanism adopted by the application realizes the special optimization and collaborative consideration of different performance targets through the parallel operation of the three decision sub-modules of energy efficiency optimization, quality optimization and stability optimization. Combined with the dynamic weight distribution strategy based on the production stage self-adaptation in the multi-target collaborative execution module, the system can flexibly consider zinc consumption and production stability while ensuring the core requirements of coating quality, significantly improving the comprehensive operation efficiency of the system.

[0048] 3. The process knowledge base and adaptive learning engine introduced in the application enable the system to have continuous evolution capability, learn and optimize its model parameters and control strategies from historical data and real-time operation results, thereby gradually adapting to changes in production line equipment characteristics and process requirements of new product specifications, realizing the leap from static control to dynamic adaptation, and effectively ensuring the quality control stability and production efficiency in long-term operation.

[0049] 4. Through the design of the instruction smooth transition unit, the impact in the setting value adjustment process is effectively suppressed, especially in the specification switching and speed adjustment stage, significantly reducing the generation of transition section waste, improving production continuity and yield, and at the same time protecting the actuator, prolonging the service life of the equipment. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 is the overall technical scheme architecture schematic diagram of the application.

[0051] Figure 2 is the interaction relationship and data flow schematic diagram of the distributed parallel decision and multi-target collaborative execution module of the application. DETAILED DESCRIPTION

[0052] In order to further illustrate the technical means and effects adopted by the application to achieve the predetermined invention purpose, the specific embodiments according to the application are described in detail as follows in combination with the drawings and preferred embodiments.

[0053] Please refer to the accompanying Figure 1 , the embodiment details a hot-dip galvanizing coating intelligent control system, aiming to solve the technical problems of large coating thickness fluctuation, high zinc consumption, poor quality stability and insufficient production rhythm adaptability caused by strong nonlinearity, large time delay and multivariable coupling in the hot-dip galvanizing coating control process. It mainly realizes accurate and stable control of coating quality and significant improvement of production efficiency by building a full-process intelligent control architecture integrating multi-modal perception, dynamic coupling modeling, distributed parallel decision and multi-target collaborative optimization.

[0054] Specifically, the system comprises a multi-modal sensor acquisition module, a dynamic coupling modeling module, a distributed parallel decision module, and a multi-target collaborative execution module, the output of the multi-modal sensor acquisition module is coupled to the dynamic coupling modeling module, the output of the dynamic coupling modeling module is coupled to the distributed parallel decision module, and the output of the distributed parallel decision module is coupled to the multi-target collaborative execution module.

[0055] The multi-modal sensor acquisition module is configured to acquire in real time the strip steel speed, the zinc liquid temperature, the air knife pressure, the air knife distance, the strip steel tension, and the post-plating plate surface temperature distribution data during the operation of the production line; the dynamic coupling modeling module is configured to establish a nonlinear state space model describing the dynamic coupling relationship between the air knife pressure, the strip steel speed, the zinc liquid temperature, and the coating thickness based on the real-time data provided by the multi-modal sensor acquisition module; the distributed parallel decision module is configured to receive the state prediction information output by the dynamic coupling modeling module, and perform parallel calculation based on a plurality of preset performance indicators to generate a plurality of sets of optimized decision schemes for the air knife pressure set value, the strip steel speed set value, and the zinc liquid temperature set value; and the multi-target collaborative execution module is configured to comprehensively evaluate and fuse the plurality of sets of optimized decision schemes output by the distributed parallel decision module, generate final execution instructions, and issue the final execution instructions to corresponding execution mechanisms.

[0056] The multi-modal sensor acquisition module is the basis of the system perception layer, and comprises a speed detection unit, a temperature detection unit, a pressure detection unit, a distance detection unit, a tension detection unit, and a face array infrared temperature measurement unit.

[0057] The speed detection unit adopts a laser speed meter to measure the strip steel running speed in real time. The laser speed meter is installed after the entry loop of the galvanizing line, and the laser beam emitted by the laser speed meter irradiates the strip steel surface at a specific angle. The instantaneous linear speed of the strip steel is accurately calculated by analyzing the Doppler frequency shift of the reflected laser beam. The calculated speed data is transmitted to a data acquisition card at a rate of one hundred samples per second through an industrial Ethernet. The data acquisition card performs filtering and denoising processing on the original signal, eliminates the measurement noise introduced by the strip steel jitter or surface reflectivity change, and finally generates the real-time digital quantity of the strip steel speed.

[0058] The temperature detection unit adopts an immersed thermocouple array to measure the temperature of the zinc liquid in different regions of the zinc pot. The thermocouple array is composed of twelve K-type armored thermocouples arranged in a matrix of three rows and four columns. Six of the thermocouples are located in the middle region of the zinc pot, three are located on the inlet side of the zinc pot, and three are located on the outlet side of the zinc pot. The measuring end of each thermocouple is encapsulated in a ceramic protective sleeve resistant to zinc liquid corrosion, and its lead is connected to a multi-channel temperature transmitter through high-temperature compensation wires. The temperature transmitter amplifies and converts the microvolt-level potential signals generated by each thermocouple into a standard current signal of four to twenty milliamperes, which is then sent to the programmable logic controller (PLC) through an analog input module. The PLC has a built-in temperature gradient calculation algorithm that calculates the temperature gradient of the zinc liquid in the length and width directions in real time by comparing the readings of the thermocouples at different positions. This gradient data serves as an important indicator of the uniformity of the zinc liquid's thermal state.

[0059] The pressure detection unit uses high-precision pressure transmitters to monitor the pressure values of the gas knife supply manifold and the left and right branch pipes. An absolute pressure transmitter is installed on the gas knife supply manifold, with a range of zero to six hundred kilopascals and an accuracy of one percent. A gauge pressure transmitter is installed on each of the left and right branch pipes, with a range of zero to two hundred kilopascals and an accuracy of half a percent. All pressure transmitters are equipped with pulse dampers to suppress fluctuations in readings caused by airflow pulsations. The four to twenty milliamperes signals output by the pressure transmitters are transmitted to the distributed control system (DCS) through shielded cables. The DCS compares the pressures of the left and right branch pipes in real time, and if the difference between the two exceeds five pascals, it triggers the gas knife lateral pressure equalization adjustment mechanism.

[0060] The distance detection unit uses laser displacement sensors to detect the distance between the gas knife lip and the surface of the strip steel in real time. Two laser triangulation displacement sensors are installed on the left and right sides of the gas knife device, with their measurement beams perpendicular to the surface of the strip steel. The sensors emit six hundred and fifty nanometer wavelength laser light, which is detected by the built-in position sensitive device to calculate the absolute distance from the gas knife lip to the surface of the strip steel. The measurement data is transmitted to the industrial computer (IPC) at a frequency of one hundred hertz through the RS485 serial interface. The IPC calculates the arithmetic mean of the distance measurements from the left and right sides and compares it with the set distance value to generate a distance deviation signal for closed-loop control of the gas knife height.

[0061] The tension detection unit obtains the real-time tension data of the strip steel through the tension sensor attached to the tension roller. The tension sensor adopts the strain gauge type force measurement principle. The elastic body structure is directly integrated in the bearing seat of the tension roller. When the tension of the strip steel acts on the tension roller, the elastic body generates a micro-strain. The full-bridge strain gauge pasted on it converts the strain into resistance change. Then, the millivolt level voltage signal is converted through the Wheatstone bridge. After being amplified by the instrument amplifier, the millivolt level voltage signal is digitized by the twenty-four-bit high-resolution analog-to-digital converter. The digitized tension data is periodically transmitted to the main controller through the PROFIBUS DP field bus. The main controller normalizes the original tension value according to the material, width and thickness of the strip steel to obtain the tension value per unit width for subsequent modeling and control.

[0062] The area array infrared temperature measurement unit is installed at the entrance of the post-plating cooling section to obtain the two-dimensional temperature field distribution data of the strip steel surface. The core component of the area array infrared temperature measurement unit is a non-cooled microbolometer infrared focal plane array with three hundred and twenty by two hundred and forty pixels. Its spectral response range is eight to fourteen micrometers, the temperature measurement range is one hundred to six hundred degrees, and the thermal sensitivity is better than zero point one degree. A compressed air purge window is installed in front of the infrared lens to prevent zinc vapor condensation from contaminating the lens. The infrared temperature measurement unit collects thermal images of the strip steel surface at a rate of twenty-five frames per second. Each thermal image contains seventy-six thousand eight hundred independent temperature pixels. The original thermal image data collected is transmitted to a dedicated image processing computer through a gigabit Ethernet. The image processing computer first performs non-uniformity correction to eliminate the response differences of the detector's own pixels. Then, it performs radiation compensation to convert the measured radiation value into the true temperature value according to the strip steel surface emissivity curve. Finally, it removes random noise through a spatial filtering algorithm and calculates the temperature average, maximum, minimum and standard deviation of the entire image to form a feature data set of the post-plating plate surface temperature distribution.

[0063] The dynamic coupling modeling module is the core analysis engine of the entire system. Its function is to establish and maintain a nonlinear state space model that can accurately describe the dynamic coupling relationship between the air knife pressure, the strip steel speed, the zinc liquid temperature and the coating thickness based on the real-time data provided by the multi-modal sensing acquisition module.

[0064] The core of the dynamic coupling modeling module is to build a nonlinear state space model containing four state variables, three control inputs and two main outputs. The four state variables are the coating thickness deviation, the strip steel speed change rate, the zinc liquid temperature gradient and the air knife pressure fluctuation; the three control inputs are the air knife pressure set value, the strip steel speed set value and the zinc liquid temperature set value; the two main outputs are the predicted coating thickness and the predicted coating uniformity index.

[0065] The establishment and updating of the dynamic coupling modeling module is a continuous and dynamic process. First, the dynamic coupling modeling module retrieves the historical data of the last one thousand coils from the process knowledge base, which includes the set values, actual measured values of various process parameters, and the coating thickness values detected by the final thickness gauge.

[0066] At the same time, the dynamic coupling modeling module receives the strip speed, zinc liquid temperature, air knife pressure, air knife distance, strip tension, and post-plating plate surface temperature distribution data from the multi-modal sensing acquisition module in real time. Based on these historical data and real-time data, the recursive least squares method is used to identify the model parameters online. The core recursive formula of the recursive least squares method is as follows: ;

[0067] Where k represents the current sampling time, P is the covariance matrix of parameter estimation error, φ is the data vector containing state variables and control inputs, and λ is the forgetting factor.

[0068] The core recursive formula is executed once every control period, usually one hundred milliseconds. By introducing an adaptive mechanism with a forgetting factor, the dynamic coupling modeling module can track the slow time-varying characteristics of the process conditions, such as the gradual change of the zinc composition in the zinc pot or the slight wear of the air knife nozzle. The typical value of the forgetting factor λ is between 0.998 and 0.9995. The system dynamically adjusts λ based on the adaptive covariance of the prediction error of the dynamic coupling modeling module. If the prediction error continues to increase, λ is automatically reduced to increase the model update speed. If the prediction error stabilizes at a low level, λ is increased to enhance the model's noise immunity.

[0069] Finally, the dynamic coupling modeling module uses Kalman filtering technology to optimally estimate the model state to suppress the interference of measurement noise. The Kalman filter includes two steps: prediction and update.

[0070] In the prediction step, the state at the current time is predicted based on the state estimation value at the previous time and the current control input through the state space model. In the update step, the predicted state is compared with the current actual measurement value, and the predicted state is corrected by calculating the Kalman gain matrix to obtain the optimal state estimation value.

[0071] This process continues to run, ensuring that the model state is always the unbiased estimate closest to the true process state. The predicted coating thickness, a scalar value with units of grams per square meter, represents the final coating weight formed after the strip leaves the zinc pot under the current process parameters. The predicted coating uniformity index is a dimensionless number between zero and one, calculated based on the distribution of the predicted coating thickness in the width direction of the strip. The closer the value is to one, the better the coating uniformity. These two output values, as key quality prediction indicators, are directly sent to the distributed parallel decision-making module.

[0072] The distributed parallel decision module is responsible for generating multiple sets of control decision schemes targeting different optimization objectives based on the state prediction information provided by the dynamic coupling modeling module.

[0073] Please refer to the attached Figure 2 The distributed parallel decision module is composed of three parallel running decision sub-modules, namely, the energy efficiency optimal decision sub-module, the quality optimal decision sub-module, and the stability optimal decision sub-module.

[0074] The three decision sub-modules can be deployed on different computing cores of the same industrial server or distributed in three independent embedded industrial computers, and exchange and synchronize data through shared memory or high-speed industrial networks. Each decision sub-module adopts a model predictive control framework to generate an optimized sequence of gas knife pressure set value, strip speed set value, and zinc liquid temperature set value by solving a rolling optimization problem within a finite time domain.

[0075] The energy efficiency optimal decision sub-module optimizes with the minimization of zinc consumption per unit area as the objective function. The mathematical expression of the objective function is to minimize the total zinc consumption per unit area of the strip within the prediction time domain. This optimization calculation not only considers the zinc liquid blown off by the gas knife but also comprehensively considers the evaporation loss of the zinc pot and the efficiency of the strip taking away the zinc liquid. The optimization process is subject to a series of constraint conditions, including the lower and upper limits of the coating thickness that must meet the product specification requirements, the gas knife pressure cannot exceed the equipment safety limit, the strip speed must be within the maximum design speed range of the production line, the zinc liquid temperature must be kept above the solidification point of the zinc liquid and avoid excessive oxidation.

[0076] The quality optimal decision sub-module optimizes with the minimization of the weighted sum of coating thickness deviation and uniformity index as the objective function. The coating thickness deviation is defined as the absolute value of the difference between the predicted coating thickness and the target thickness, and the uniformity index directly uses the output of the dynamic coupling modeling module. The optimization focus of the quality optimal decision sub-module is to ensure that the coating thickness accurately reaches the set value and is as uniform as possible across the entire plate surface. In addition to the device physical limit, the quality optimal decision sub-module emphasizes that the coating uniformity index must not be lower than 0.85, which is a red line for process quality.

[0077] The stability optimal decision submodule optimizes calculation with the square sum of the change rate of the key process parameters as the objective function, the key process parameters including the gas knife pressure set value, the strip speed set value and the zinc liquid temperature set value, the objective function aims to minimize the change range of these set values in adjacent control periods, so as to ensure the stability of the production process and reduce the impact on the mechanical actuator and the thermal system, especially suitable for the transition stage of strip specification switching or production speed adjustment, and the constraint condition ensures the smooth and continuous change of the process parameters and avoids the stepwise set value jump.

[0078] The model predictive control algorithm of each decision submodule takes two hundred and fifty milliseconds as a control period, and the prediction time domain length is twenty periods, i.e. five seconds. At the beginning of each control period, each decision submodule obtains the latest state prediction value from the dynamic coupling modeling module, and then independently solves the optimization problem of each decision submodule. The optimization problem is solved by using an interior point method algorithm, which converts the constraint optimization problem into a series of unconstrained optimization problems by introducing an obstacle function, and uses Newton iteration method to quickly converge to the optimal solution. After the solution is completed, each decision submodule outputs an optimized sequence of gas knife pressure set value, strip speed set value and zinc liquid temperature set value in the next five sampling periods, but only the set value at the current time is submitted to the multi-objective collaborative execution module for fusion.

[0079] The core function of the multi-objective collaborative execution module is to comprehensively evaluate and fuse the three groups of optimized decision schemes output by the distributed parallel decision modules to generate the final execution instruction. The decision fusion mechanism of the multi-objective collaborative execution module is as follows:

[0080] First, a dynamic weight coefficient is assigned to the decision schemes output by the energy efficiency optimal decision submodule, the quality optimal decision submodule and the stability optimal decision submodule. The dynamic weight coefficient is not fixed but is adaptively adjusted according to the current production stage and the target priority. The system divides the production stage into normal steady-state production stage and specification switching or acceleration / deceleration transition stage.

[0081] In the normal steady-state production stage, i.e. after the strip specification and speed have been running stably for more than thirty seconds, the weight coefficient of the quality optimal decision submodule is set to zero point five, and the weight coefficients of the energy efficiency optimal decision submodule and the stability optimal decision submodule are each set to zero point two five.

[0082] In the specification switching or acceleration / deceleration transition phase, i.e. within a time window of sixty seconds after the system detects the target specification change or speed instruction change, the weight coefficient of the stability optimal decision submodule is raised to zero point five, the weight coefficient of the quality optimal decision submodule is adjusted to zero point three, and the weight coefficient of the energy efficiency optimal decision submodule is adjusted to zero point two. The adjustment process of the weight coefficient is smoothed by a first-order inertia link to avoid the oscillation of the final set value caused by the sudden change of the weight.

[0083] Finally, the three groups of decision schemes after weighting are linearly weighted and averaged to calculate the final gas knife pressure set value, strip speed set value and zinc liquid temperature set value. The linear weighted average calculation is independent for each control parameter. For example, the final gas knife pressure set value is equal to the gas knife pressure recommended by the energy efficiency optimal decision submodule multiplied by the energy efficiency weight, plus the gas knife pressure recommended by the quality optimal decision submodule multiplied by the quality weight, plus the gas knife pressure recommended by the stability optimal decision submodule multiplied by the stability weight. The final set value obtained by calculation needs to be processed by the instruction smoothing transition unit before being sent to the underlying actuator.

[0084] The instruction smoothing transition unit is a key interface connecting the high-level control decision and the underlying actuator, and its core function is to slope process the step change of the set value to ensure smooth action of the actuator. The instruction smoothing transition unit receives the final gas knife pressure set value, strip speed set value and zinc liquid temperature set value from the multi-target collaborative execution module.

[0085] For each set value, the instruction smoothing transition unit maintains a current actual output value inside. When a new set value instruction arrives, the instruction smoothing transition unit compares the difference between the new set value and the current actual output value. If the difference is less than the preset dead zone threshold, the new set value is directly output. If the difference exceeds the dead zone threshold, a ramp function is started. The ramp rate of the ramp function is not fixed, but is adaptively calculated according to the strip speed and the specification change amount. For the strip speed set value, its ramp rate is proportional to the current strip speed, and the basic rate is set to zero point five meters per second, but the maximum is two meters per second. For the gas knife pressure set value, its ramp rate is calculated according to the strip width change amount. For every one hundred millimeters of width change, the ramp rate increases by one thousand pascals per second. For the zinc liquid temperature set value, due to its large thermal inertia, the ramp rate is fixed at zero point five degrees per minute.

[0086] Through this adaptive ramp processing, the impact of set value mutation on the stability of strip running and the formation process of plated layer is effectively avoided, the thickness out-of-tolerance products generated in the specification switching transition section are significantly reduced, and the key execution devices such as the gas knife regulating valve, the main drive motor and the zinc pot heater are protected, prolonging their service life.

[0087] The system also contains a process knowledge base and an adaptive learning engine, which together constitute the system's self-optimization and evolution capabilities. The process knowledge base is a structured database that stores data of three types.

[0088] The first type is the historical optimal process parameter combination. For each produced steel grade and size combination, such as low-carbon steel with a thickness of 1.2 mm and a width of 1250 mm, the system stores multiple sets of process parameter settings that have historically produced the best coating quality, the lowest zinc consumption, or the most stable production, including gas knife pressure, strip speed, zinc liquid temperature, gas knife distance, etc. The second type is the process window limits for different steel grades and sizes. For example, the upper limit of the zinc liquid temperature for a certain high-strength steel is 465 degrees, and the lower limit of the gas knife pressure is 20 kPa. The third type is the characteristic data of various defects and the corresponding strategies. For example, when the coating has an edge over-thickness defect, the characteristic data is that the edge temperature detected by the area array infrared temperature measurement unit is more than 10 degrees higher than the middle part, and the corresponding strategy is to appropriately increase the edge gas knife pressure setting value by 5 to 10 Pa. The process knowledge base provides data query and services to other modules in the system through a standard structured query language interface.

[0089] The adaptive learning engine is the intelligent core of the system. Its function is to iteratively optimize and update the parameters of the dynamic coupling modeling module, the weight coefficients of the objective functions in the distributed parallel decision-making module, and the dynamic weight coefficients in the multi-objective collaborative execution module based on the actual results of each production task. The iterative optimization and update strategy uses a gradient descent-based reinforcement learning algorithm. After each coil of strip steel is produced, the system obtains the actual coating thickness average, uniformity standard deviation, and actual zinc consumption data from the downstream quality detection station.

[0090] These actual data, together with the system's own predicted values and control targets, form a reward signal. The adaptive learning engine takes the maximization of cumulative production benefit as its long-term optimization goal. Production benefit is a comprehensive indicator that takes into account coating quality score, zinc consumption cost, and production stability score. The engine calculates the difference between the actual benefit and the expected benefit under the current strategy, calculates the gradient using the backpropagation algorithm, and then fine-tunes the model parameter matrix in the dynamic coupling modeling module, the weight coefficients of each objective function in the distributed parallel decision-making module, and the threshold parameters for stage judgment in the multi-objective collaborative execution module along the gradient direction. The step size of each update is strictly limited, usually not more than 1% of the original parameter value, to ensure the stability of the learning process and avoid drastic fluctuations in system performance caused by single bad data.

[0091] Through continuous small-step iterative learning, the system can gradually adapt to slow changes in production line equipment characteristics, such as gas knife nozzle wear, temperature measurement element drift, and process requirements for new product specifications, thereby realizing the transition from static program control to dynamic adaptation.

[0092] The operation of the whole hot galvanizing coating intelligent control system relies on a solid industrial computing and communication platform, and the main controller can adopt a high-performance programmable automation controller which is equipped with a multi-core processor, the frequency of each core is not less than 1.8 GHz, and the memory capacity is not less than 4 GB.

[0093] The operation tasks of the distributed parallel decision module can be deployed on an additional industrial server which is equipped with a graphics processing unit to accelerate the solution of the model predictive control optimization problem, and all modules of the system communicate through a deterministic industrial Ethernet network, the network protocol adopts the IEEE 802.1AS time-sensitive network standard, and the transmission of control instructions and feedback data is completed within a strictly specified time window, and the worst-case communication delay is less than 1 ms.

[0094] The system software architecture adopts a modular design, the control logic is written in accordance with the IEC 61131-3 standard, and a high-performance algorithm library written in C++ language is supplemented, the human-computer interface provides complete system state monitoring, parameter setting, trend recording and alarm management functions, the operator can clearly observe the real-time data of the multi-modal sensing acquisition module, the prediction output of the dynamic coupling modeling module, the optimization results of the three decision sub-modules and the final decision and weight distribution of the multi-objective collaborative execution module, so as to realize transparent monitoring and intelligent intervention of the production process.

[0095] The above is only a preferred embodiment of the present application, and does not limit the present application in any form, although the present application has been disclosed as above, however, it is not intended to limit the present application, any person skilled in the art, without departing from the scope of the technical scheme of the present application, can make some changes or modifications of the above disclosed technical content as equivalent embodiments, but as long as it does not deviate from the technical scheme content of the present application, any modification, equivalent change and modification of the above embodiments according to the technical essence of the present application, all still belong to the scope of the technical scheme of the present application.

Claims

1. A hot-dip galvanizing coating and plating intelligent control system, characterized by, The method comprises the following steps: A multi-modal sensor acquisition module is used to acquire strip steel speed, zinc liquid temperature, air knife pressure, air knife distance, strip steel tension and post-plating plate surface temperature distribution data in real time during the production line operation process; A dynamic coupling modeling module is coupled with the output of the multi-modal sensor acquisition module, and is used to establish a nonlinear state space model describing the dynamic coupling relationship between air knife pressure, strip steel speed, zinc liquid temperature and plated layer thickness based on the real-time data provided by the multi-modal sensor acquisition module; A distributed parallel decision module is coupled with the output of the dynamic coupling modeling module, and is used to receive state prediction information output by the dynamic coupling modeling module and perform parallel calculation based on a plurality of preset performance indicators to generate a plurality of sets of optimized decision schemes for air knife pressure set value, strip steel speed set value and zinc liquid temperature set value, wherein the distributed parallel decision module is composed of three parallel running decision sub-modules, namely, an energy efficiency optimal decision sub-module, a quality optimal decision sub-module and a stability optimal decision sub-module; the energy efficiency optimal decision sub-module performs optimization calculation with minimization of unit area zinc consumption as an objective function; the quality optimal decision sub-module performs optimization calculation with minimization of the weighted sum of plated layer thickness deviation and uniformity index as an objective function; and the stability optimal decision sub-module performs optimization calculation with minimization of the square sum of the change rates of key process parameters as an objective function; A multi-objective collaborative execution module is coupled with the output of the distributed parallel decision module, and is used to comprehensively evaluate and fuse the plurality of sets of optimized decision schemes output by the distributed parallel decision module to generate final execution instructions and issue them to corresponding execution mechanisms, wherein the decision fusion mechanism of the multi-objective collaborative execution module comprises: a dynamic weight coefficient is assigned to the decision schemes output by the energy efficiency optimal decision sub-module, the quality optimal decision sub-module and the stability optimal decision sub-module; the dynamic weight coefficient is adaptively adjusted according to the current production stage and the target priority; in the normal steady-state production stage, the weight coefficient of the quality optimal decision sub-module is set to a high value; in the specification switching or acceleration / deceleration transition stage, the weight coefficient of the stability optimal decision sub-module is increased to a high value; and finally, the linear weighted average of the plurality of sets of weighted decision schemes is calculated to obtain the final control set value.

2. The hot dip galvanizing coating and plating intelligent control system according to claim 1, characterized in that, The multi-modal sensor acquisition module comprises a speed detection unit, a temperature detection unit, a pressure detection unit, a distance detection unit, a tension detection unit and a surface array infrared temperature measurement unit; The speed detection unit uses a laser speed meter to measure the strip steel running speed in real time; The temperature detection unit uses an immersion type thermocouple array to measure the zinc liquid temperature in different regions inside the zinc pot; The pressure detection unit uses a high-precision pressure transmitter to monitor the pressure values of the air knife gas supply main pipe and the branch pipes on both sides; The distance detection unit uses a laser displacement sensor to detect the distance between the air knife lip and the strip steel surface in real time; The tension detection unit obtains real-time tension data of the strip steel through the tension sensor attached to the tension roller; The surface array infrared temperature measurement unit is installed at the entrance of the post-plating cooling section to obtain two-dimensional temperature field distribution data of the strip steel surface.

3. The hot dip galvanizing coating and plating intelligent control system according to claim 1, characterized in that, The nonlinear state space model constructed by the dynamic coupling modeling module contains four state variables, three control inputs and two main outputs; The four state variables are respectively coating thickness deviation, strip speed change rate, zinc liquid temperature gradient and air knife pressure fluctuation amount; The three control inputs are respectively air knife pressure set value, strip speed set value and zinc liquid temperature set value; The two main outputs are respectively predicted coating thickness and predicted coating uniformity index.

4. The hot dip galvanizing coating and plating intelligent control system according to claim 3, characterized in that, The model establishment process of the dynamic coupling modeling module includes: Based on historical production data and real-time sensing data, the recursive least squares method is used to identify the model parameters online; By introducing an adaptive mechanism with a forgetting factor, the model can track the slow time-varying characteristics of the process conditions; And the Kalman filtering technology is used to optimize the estimation of the model state to suppress the interference of measurement noise.

5. The hot dip galvanizing coating and plating intelligent control system according to claim 1, characterized in that, Each decision-making submodule adopts a model predictive control framework, which generates an optimized sequence of air knife pressure set value, strip speed set value and zinc liquid temperature set value by solving a rolling optimization problem in a finite time domain.

6. The hot dip galvanizing coating and plating intelligent control system according to claim 1, characterized in that, It also includes a process knowledge base and an adaptive learning engine; The process knowledge base is used to store historical optimal process parameter combinations, process window limits for different steel specifications, and characteristic data and coping strategies for various defects; The adaptive learning engine is used to iteratively optimize and update the parameters of the dynamic coupling modeling module, the objective function weights of the distributed parallel decision-making module, and the dynamic weight coefficients of the multi-objective collaborative execution module based on the actual results of each production task.

7. The hot dip galvanizing coating and plating intelligent control system according to claim 6, characterized in that, The update strategy of the adaptive learning engine adopts a reinforcement learning algorithm based on gradient descent, with the maximization of cumulative production benefit as the long-term optimization goal.

8. The hot dip galvanizing coating and plating intelligent control system according to claim 1, characterized in that, The multi-objective collaborative execution module is provided with an instruction smooth transition unit between the bottom execution mechanism; The instruction smooth transition unit is used to perform ramp processing on the step change of the set value, and the ramp rate is adaptively calculated according to the strip speed and specification change amount, ensuring smooth action of the execution mechanism and avoiding impact on the strip running and coating formation process.

Images