A method and system for controlling the temperature of a cold heading steel rod on a cooling roller

Abstract

This invention relates to the field of temperature control technology and discloses a method and system for temperature control of air-cooled roller conveyors for cold heading steel wire rod. The method includes: analyzing the phase transformation law of the current wire rod steel grade during the cooling process using a multi-field coupling system, and dividing the phase transformation temperature range of the current wire rod from austenite to pearlite / sorbite based on the phase transformation law; dividing the target cooling path into a high-temperature segment, a phase transformation segment, and a low-temperature segment based on the phase transformation temperature range; using strong convection cooling control in the high-temperature segment, isothermal control in the phase transformation segment, and low-stress slow cooling control in the low-temperature segment as temperature control targets, and solving to generate an optimal segmented cooling strategy; and converting the optimal segmented cooling strategy into control commands for the cooling actuator to drive its operation. This invention accurately divides the high-temperature segment, phase transformation segment, and low-temperature segment according to the phase transformation temperature range, and sets targeted temperature control targets based on the characteristics of each stage, enabling refined control of the cooling process.

Description

Technical Field

[0001] This invention relates to the field of temperature control technology, and more specifically, to a method and system for temperature control of air-cooled roller conveyors for cold heading steel wire rods. Background Technology

[0002] The Stellmore air-cooled roller conveyor is an air-cooled roller conveyor system widely used in the metal processing industry. Its core function is to uniformly and stably force-cool steel and other metal workpieces during the cooling process, controlling the cooling rate of the workpiece and preventing physical deformation caused by excessive temperature differences. Temperature control of the air-cooled roller conveyor is a precise and critical process, requiring the maintenance of accurate and stable temperature during high-speed continuous production to ensure the final quality of the workpiece.

[0003] The phase transformation law of steel is a crucial basis for temperature control, determining the microstructure transformation and property evolution characteristics of steel in different temperature ranges. However, traditional air-cooled roller conveyor temperature control methods often fail to fully consider this factor. If the specific phase transformation law of steel is ignored in temperature control, it is difficult to accurately correlate the phase transformation critical point with the dynamic cooling process, and it is also impossible to formulate the optimal cooling curve for different steel types, thus affecting the precise control effect of the cooling process.

[0004] In the absence of phase transformation laws, the resulting temperature control strategies are usually simplistic and fail to deeply integrate the dynamic characteristics of steel microstructure transformation in different temperature ranges. They often rely solely on uniform control standards, ignoring the individual differences of different steel grades and their temperature-phase transformation-performance correlation mechanisms. This control method cannot adapt to the actual process requirements of different steel grades under different cooling conditions, and can easily lead to local overcooling or undercooling of the workpiece during the cooling process, which in turn has an adverse effect on the physical properties and surface quality of the finished product.

[0005] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0006] In view of the problems in the related technologies, the present invention proposes a method and system for temperature control of air-cooled roller conveyor for cold heading steel wire rod, so as to overcome the above-mentioned technical problems existing in the existing related technologies.

[0007] Therefore, the specific technical solution adopted by the present invention is as follows:

[0008] According to a first aspect of the present invention, a method for temperature control of a cold heading steel wire rod air-cooling roller conveyor is provided, the method comprising:

[0009] A multi-source time-series dataset was collected on the cooling process of wire rod after high-speed wire rod rolling on the Steyrmo air-cooled roller conveyor. The multi-source time-series dataset includes wire rod surface temperature change data, cooling actuator operating parameters, ambient temperature and humidity data, and material property data of the current wire rod steel grade.

[0010] Multi-source time-series datasets are input into a multi-field coupling system integrating thermo-dynamic-microstructure. The phase transformation law of the current wire rod steel grade during the cooling process is analyzed using the multi-field coupling system. Based on the phase transformation law, the phase transformation temperature range of the current wire rod from austenite to pearlite / sorbite structure is divided.

[0011] Based on the phase change temperature range, the target cooling path of wire rod in the Steyrmore air-cooled roller cooling process is divided into a high-temperature section, a phase change section, and a low-temperature section.

[0012] The optimal segmented cooling strategy is generated by using high-temperature section strong convection cooling control, phase change section isothermal control and low-stress slow cooling control as temperature regulation targets, and the temperature boundaries of high-temperature section, phase change section and low-temperature section as constraints.

[0013] The optimal segmented cooling strategy is converted into control commands for the cooling actuator, which drives the cooling actuator to operate, thereby achieving temperature control of the wire rod on the Steyrmo air-cooled roller conveyor.

[0014] Preferably, the step of inputting multi-source time-series datasets into a multi-field coupling system integrating thermo-dynamic-microstructure analysis, using the multi-field coupling system to analyze the phase transformation law of the current wire rod steel grade during the cooling process, and dividing the phase transformation temperature range of the current wire rod from austenite to pearlite / sorbite based on the phase transformation law includes:

[0015] Multi-source time-series datasets are input into an echo state network that integrates sparse reservoirs. The sparse reservoirs in the echo state network are used to mine phase transition correlation features between wire rod surface temperature change data, material property data, and environmental temperature and humidity data.

[0016] A multi-field coupled system integrating thermodynamics, kinetics, and microstructure was constructed. The phase transformation correlation characteristics were input into the multi-field coupled system for thermodynamic, kinetic, and microstructure analysis in sequence. The results of thermodynamic, kinetic, and microstructure analysis were integrated to obtain the phase transformation law of the current wire rod steel grade during the cooling process.

[0017] Among them, the microstructure analysis of multi-field coupled systems includes:

[0018] Initialize the cellular automaton simulation space, and simulate the growth and evolution of preset crystal nuclei in the cellular automaton simulation space to realize the dynamic evolution simulation of the microstructure of the current wire rod steel grade during the cooling process.

[0019] Based on the phase transformation law of the current wire rod steel grade during the cooling process, the phase transformation initiation temperature and phase transformation termination temperature are calculated. Using the phase transformation initiation temperature and phase transformation termination temperature as boundaries, the phase transformation temperature range of the current wire rod from austenite to pearlite / sorbite is divided.

[0020] Preferably, the initialization of the cellular automaton simulation space, in which the growth and evolution process of preset crystal nuclei is simulated to realize the dynamic evolution simulation of the microstructure of the current wire rod steel grade during the cooling process, includes:

[0021] Initialize the cellular automata simulation space based on grid cells, set the initial nucleus positions in the austenitic matrix, and configure the initial cooling conditions;

[0022] The phase transition driving force of each grid cell is calculated based on the temperature field formed by the temperature change of the wire rod surface and the composition field formed by the material properties.

[0023] Based on the phase transition driving force of each grid cell and combined with the predefined growth competition rules of adjacent phase regions, the phase state transition probability of adjacent grid cells is calculated. The co-growth of pearlite clusters and the lamellar coarsening process are simulated through a probability selection mechanism, and the growth morphology of pearlite clusters and the distribution of phase transition units are output.

[0024] The temperature field and composition field are dynamically updated based on the current pearlite cluster growth morphology and phase transition unit distribution. The phase transition driving force of each grid unit is iteratively calculated based on the updated temperature field and composition field until the iteration number is met or the cooling termination condition is met.

[0025] The output is the simulation results of pearlite cluster morphology and phase distribution, which describe the dynamic evolution of the microstructure of the current wire rod steel during the cooling process.

[0026] Preferably, the calculation of the phase state transition probability of adjacent grid cells based on the phase transition driving force of each grid cell and in combination with predefined adjacent phase region growth competition rules includes:

[0027] Calculate the phase transition driving energy difference between each grid cell and its adjacent grid cells; based on the crystallographic orientation relationship between adjacent grid cells, calculate the phase transition driving energy difference corrected by the interface energy change caused by the interface orientation difference.

[0028] The corrected driving energy difference is input into the predefined adjacent phase region growth competition rule. The adjacent phase region growth competition rule is used to compare the phase state, relative supercooling and historical evolution stage of the current grid cell with each neighboring cell, and to comprehensively determine the priority change direction.

[0029] Based on the preferred transition direction, a competitive weight is assigned to each adjacent grid cell. After normalizing the competitive weights of all adjacent grid cells, they are converted into probability values, thus obtaining the phase state transition probability of each adjacent grid cell in the next simulation step.

[0030] Preferably, the step of dividing the target cooling path of the wire rod in the Stellmore air-cooled roller conveyor cooling process into a high-temperature section, a phase change section, and a low-temperature section based on the phase change temperature range; using strong convection cooling control in the high-temperature section, isothermal control in the phase change section, and low-stress slow cooling control in the low-temperature section as temperature control targets, and using the temperature boundaries of the high-temperature section, phase change section, and low-temperature section as constraints, to solve for and generate the optimal segmented cooling strategy includes:

[0031] Based on the phase change temperature range, the target cooling path of wire rod in the Steyrmore air-cooled roller cooling process is divided into a high-temperature section, a phase change section, and a low-temperature section.

[0032] Among them, the temperature range of the high-temperature section is from the rolling exit temperature to the phase transformation initiation temperature, the temperature range of the phase transformation section is the temperature within the phase transformation temperature range, and the temperature range of the high-temperature section is from the phase transformation termination temperature to the target coiling temperature.

[0033] Based on the phase transition correlation features, a spatiotemporal probability distribution map of the phase transition temperature range is generated. The spatiotemporal probability distribution map is then mapped to a low-dimensional space using manifold learning technology to extract low-dimensional manifold lines that characterize the essence of the cooling process.

[0034] The target temperature regulation is to use high-temperature section strong convection cooling control, phase change section isothermal control and low-stress slow cooling control as the temperature regulation objectives, and the temperature boundaries of the high-temperature section, phase change section and low-temperature section as the constraints. The target cooling path optimization problem is transformed into a problem to be solved on a low-dimensional manifold.

[0035] The optimal continuous smooth path that satisfies the temperature control objectives at each stage is found on the low-dimensional manifold. The optimal continuous smooth path is then mapped back to the original feature space to generate a segmented cooling strategy that satisfies the temperature control objectives.

[0036] Preferably, the step of solving for the optimal continuous smooth path that satisfies the temperature control objectives at each stage on the low-dimensional manifold, mapping the optimal continuous smooth path back to the original feature space, and generating a piecewise cooling strategy that satisfies the temperature control objectives includes:

[0037] On the low-dimensional manifold line that characterizes the essence of the cooling process, the rolling exit temperature and the target coil temperature are mapped to the start and end points on the low-dimensional manifold line. The temperature control targets are strong convection rapid cooling in the high-temperature section, isothermal maintenance in the phase change section, and low-stress slow cooling in the low-temperature section. The temperature boundaries of the high-temperature section, the phase change section, and the low-temperature section are used as constraints. The optimal continuous smooth path is solved using the Riemann geometric optimization algorithm.

[0038] The optimal continuous smooth path obtained by the solution is restored to the continuous process parameter sequence in the original feature space through inverse mapping. Based on the pre-divided temperature boundaries of the high temperature section, phase transition section and low temperature stage, the optimal temperature control parameters corresponding to each stage are identified and extracted from the continuous process parameter sequence.

[0039] Preferably, the step of using the Riemann geometric optimization algorithm to solve for the optimal continuous smooth path includes:

[0040] The original low-dimensional manifold is given an equivalent Riemannian metric structure. A path planning problem that satisfies the temperature control objective and boundary constraints is established on the low-dimensional manifold. The problem is then mapped to a discrete numerical optimization model with an equivalent Riemannian metric structure in Euclidean space through coordinate transformation.

[0041] The objective function is to minimize the total length of the path under the equivalent Riemannian metric structure. The temperature control objective and boundary constraints are transformed into equivalent constraints of discrete coordinate points. The discrete numerical optimization model is solved by the solver to obtain a discrete coordinate sequence that satisfies the temperature control objective and boundary constraints.

[0042] The optimal continuous smooth path is obtained by interpolating the discrete coordinate sequence using the smooth interpolation method.

[0043] According to a second aspect of the present invention, a temperature control system for a cold heading steel wire rod air-cooled roller conveyor is provided, the system comprising:

[0044] The data acquisition module is used to collect multi-source time-series datasets of wire rod after high-speed wire rolling during the cooling process of the Stellmore air-cooled roller table. The multi-source time-series datasets include wire rod surface temperature change data, cooling actuator operating parameters, ambient temperature and humidity data, and material property data of the current wire rod steel grade.

[0045] The temperature range division module is used to input multi-source time-series datasets into a multi-field coupling system integrating thermo-dynamic-microstructure. The multi-field coupling system is used to analyze the phase transformation law of the current wire rod steel grade during the cooling process, and the phase transformation temperature range of the current wire rod from austenite to pearlite / sorbite structure is divided according to the phase transformation law.

[0046] The temperature control solution module is used to divide the target cooling path of wire rod in the Stellmore air-cooled roller cooling process into high-temperature section, phase change section and low-temperature section according to the phase change temperature range; with strong convection cooling control in high-temperature section, isothermal control in phase change section and low-stress slow cooling control in low-temperature section as temperature control targets, and with the temperature boundaries of high-temperature section, phase change section and low-temperature section as constraints, the optimal segmented cooling strategy is generated.

[0047] The temperature control execution module is used to convert the optimal segmented cooling strategy into control commands for the cooling actuator, driving the cooling actuator to operate, so as to achieve temperature control of the wire rod on the Steyrmo air-cooled roller conveyor.

[0048] According to a third aspect of the present invention, a computer device is provided.

[0049] In some embodiments, the computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described above.

[0050] According to a fourth aspect of the present invention, a computer-readable storage medium is provided.

[0051] In one embodiment, a computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of the above method.

[0052] The beneficial effects of this invention are as follows:

[0053] 1. This invention collects multi-source time-series data and combines it with a multi-field coupling system to accurately analyze the phase transformation law of wire rod and divide the phase transformation temperature range. Based on this, it divides the cooling stages and formulates targeted segmented cooling strategies, which can achieve precise temperature control at each stage. This effectively ensures the stable transformation of wire rod from austenite to pearlite / sorbite, improves the uniformity of wire rod microstructure and mechanical properties, and optimizes the operation of cooling actuators through optimal cooling strategies, reducing energy consumption and production losses, thereby improving the stability and production efficiency of the air-cooled roller cooling process.

[0054] 2. This invention uses cellular automata simulation to accurately initialize the simulation space and dynamically iteratively calculate the phase transformation driving force and simulate the crystal nucleus growth and evolution process. This can realistically restore the dynamic evolution process of microstructure. By combining crystallographic orientation relationships to correct the driving energy difference and calculating the phase state transition probability through competitive weight allocation and normalization, the accuracy and rationality of phase transformation simulation can be improved. Ultimately, it can achieve precise division of the phase transformation temperature range, providing a precise basis for the subsequent formulation of segmented cooling strategies, ensuring the stability and uniformity of microstructure transformation during the cooling process, and thus improving the quality of cold heading steel wire rod products.

[0055] 3. This invention precisely divides the phase transition temperature range into high-temperature, phase transition, and low-temperature segments, and sets targeted temperature control targets based on the characteristics of each stage, enabling refined control of the cooling process. By employing the Riemannian geometric optimization algorithm to endow the low-dimensional manifold with an equivalent Riemannian metric structure, the path planning problem is transformed into a discrete numerical optimization model in Euclidean space. With the goal of minimizing the total path length and integrating temperature control targets and boundary constraints, the rationality and feasibility of the optimal continuous smooth path can be guaranteed. The segmented cooling strategy obtained through smoothing interpolation and inverse mapping restoration can ensure the precise matching of temperature control parameters at each stage, achieving precise implementation of strong convection cooling in the high-temperature segment, isothermal control in the phase transition segment, and low-stress slow cooling in the low-temperature segment, effectively improving the stability and controllability of the cold heading steel wire rod cooling process. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 This is a flowchart of a method for controlling the temperature of a cold heading steel wire rod air-cooled roller conveyor according to an embodiment of the present invention;

[0058] Figure 2 This is a schematic diagram of a temperature control system for an air-cooled roller conveyor of cold heading steel wire rod according to an embodiment of the present invention;

[0059] Figure 3 This is a schematic diagram of the structure of a computer device;

[0060] Figure 4 This is a flowchart illustrating the construction of a multi-field coupling system in a method for controlling the temperature of an air-cooled roller conveyor for cold heading steel wire rod according to an embodiment of the present invention.

[0061] Figure 5 This is a flowchart of the dynamic evolution of microstructure in a method for temperature control of air-cooled roller conveyor for cold heading steel wire rod according to an embodiment of the present invention.

[0062] In the picture:

[0063] 1. Data acquisition module; 2. Temperature range division module; 3. Temperature control solution module; 4. Temperature control execution module. Detailed Implementation

[0064] To further illustrate the various embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of the present invention. The components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0065] According to an embodiment of the present invention, a method and system for controlling the temperature of a cold heading steel wire rod air-cooled roller conveyor are provided.

[0066] Figure 1 An embodiment of the method for controlling the temperature of a cold heading steel wire rod air-cooled roller conveyor according to the present invention is shown.

[0067] In this optional embodiment, a method for controlling the temperature of a cold heading steel wire rod air-cooled roller conveyor includes:

[0068] S1. Collect multi-source time-series datasets of wire rods after high-speed wire rolling during the cooling process on the Steyrmo air-cooled roller conveyor. The multi-source time-series datasets include wire rod surface temperature change data, cooling actuator operating parameters, ambient temperature and humidity data, and material property data of the current wire rod steel grade.

[0069] It should be noted that the data includes wire rod surface temperature change data, cooling actuator operating parameters, ambient temperature and humidity data, and material property data of the current wire rod steel grade. Among them, the wire rod surface temperature change data includes real-time temperature values, temperature change rates, and temperature gradient time-series data of the wire rod surface collected from multiple measuring points along the length of the Stellmore air-cooled roller conveyor. The cooling actuator operating parameters include time-series data of the start-stop status, operating frequency, wind speed level, damper opening of the section fans, as well as the roller conveyor speed, speed regulation frequency, and start-stop linkage parameters of the section roller conveyor. The ambient temperature and humidity data includes time-series data of real-time ambient temperature, ambient relative humidity, and temperature and humidity change trends at different monitoring points within the roller conveyor area. The material property data of the current wire rod steel grade includes time-series correlation data of thermodynamic parameters such as the chemical composition content, austenite grain size, original microstructure, thermal conductivity, specific heat capacity, and latent heat of phase transformation of the steel grade.

[0070] S2. Input the multi-source time series dataset into the multi-field coupling system integrating thermal-dynamic-microstructure, use the multi-field coupling system to analyze the phase transformation law of the current wire rod steel grade during the cooling process, and divide the phase transformation temperature range of the current wire rod from austenite to pearlite / sorbite structure according to the phase transformation law.

[0071] The multi-source time-series dataset is input into a multi-field coupled system integrating thermo-dynamic-microstructure analysis. This system is used to analyze the phase transformation behavior of the current wire rod steel grade during the cooling process. Based on this phase transformation behavior, the phase transformation temperature range for the current wire rod's transformation from austenite to pearlite / sorbite microstructure is defined, including:

[0072] like Figures 4-5 As shown, a multi-source time-series dataset is input into an echo state network that integrates sparse reservoirs. The sparse reservoirs in the echo state network are used to mine the phase transition correlation features between wire rod surface temperature change data, material property data, and environmental temperature and humidity data.

[0073] It should be noted that the Echo State Network (ESN) is a deep learning model based on recurrent neural networks. Its core feature is a large-scale, sparsely connected reservoir. The input layer is fully connected to the reservoir, and the neurons within the reservoir are randomly and sparsely connected with fixed weights. Only the output layer weights need to be trained. It utilizes the echo state characteristics of the reservoir to achieve feature extraction and dynamic modeling of time-series data. This eliminates the need to train complex recurrent weights, significantly reducing training difficulty and improving convergence speed. It is suitable for feature mining of multi-source time-series data, specifically including:

[0074] Preprocessing of multi-source time-series datasets: interpolation to complete missing values ​​and smoothing filtering for outliers; aligning the preprocessed wire rod surface temperature change data, material property data, and environmental temperature and humidity data by time step.

[0075] Initialize N independent sparse reservoirs, each with a different number of neurons, sparse connection probability, and spectral radius; construct an input layer-reservoir-output layer structure, where the number of neurons in the input layer matches the feature dimensions of the multi-source time series data, and the number of neurons in the output layer corresponds to the phase transition correlation feature dimensions to be mined; the input layer is fully connected to each reservoir, and each reservoir has random sparse connections with fixed weights, while each reservoir is independently connected to the output layer.

[0076] Preprocessed multi-source time-series data is input into an integrated sparse reservoir ESN. Each reservoir dynamically updates its internal neuron state (i.e., echo state) based on the input data and outputs its state matrix. Known disk phase transition marker data (such as time-series features corresponding to the start / end times of phase transition) are used as supervision signals to train the output layer weights of each reservoir, so that the associated features output by the model match the actual phase transition laws. Cross-validation is used to optimize the weight allocation of the integration strategy to ensure the effective fusion of features from each reservoir.

[0077] The multi-source time-series data collected in real time is input into the trained model. Each sparse reservoir outputs local correlation features, which are then fused to obtain global phase transition correlation features, including the correlation between surface temperature change and the time delay of phase transition initiation, the correlation between the influence of material thermal conductivity on temperature conduction rate, the nonlinear correlation between ambient humidity and phase transition rate, and so on.

[0078] A multi-field coupled system integrating thermodynamics, kinetics, and microstructure was constructed. The phase transformation correlation characteristics were input into the multi-field coupled system for thermodynamic, kinetic, and microstructure analysis in sequence. By integrating the results of thermodynamic, kinetic, and microstructure analysis, the phase transformation law of the current wire rod steel grade during the cooling process was obtained.

[0079] It should be noted that thermodynamic analysis is based on parameters such as wire rod surface temperature, material thermal conductivity, specific heat capacity, and latent heat of phase transformation from multi-source time-series data. Combined with the law of conservation of energy, a heat conduction equation for the wire rod cooling process is established. The spatiotemporal distribution of the temperature field inside the wire rod is calculated, and the coupling relationship between temperature change and heat transfer and the release of latent heat of phase transformation is analyzed. The characteristics of temperature change in different cooling stages and their impact on austenite stability are clarified. Kinetic analysis is based on the temperature field data obtained from thermodynamic analysis. Combined with parameters such as steel chemical composition and austenite grain size, the JMAK kinetic model or Arrhenius-type kinetic equation is used to quantify the relationship between the phase transformation rate of austenite to pearlite / sorbite and temperature and time. The phase transformation volume fraction change curves at different temperatures are calculated, the time thresholds for the start and end of the phase transformation are clarified, and the dynamic evolution law of the phase transformation process is revealed.

[0080] Among them, the multi-field coupling system includes the following when performing microstructure analysis: initializing the cellular automata simulation space, and simulating the growth and evolution process of preset crystal nuclei in the cellular automata simulation space to realize the dynamic evolution simulation of the microstructure of the current wire rod steel grade during the cooling process.

[0081] The process includes initializing the cellular automaton simulation space, and then simulating the growth and evolution of preset crystal nuclei within that space to simulate the dynamic evolution of the microstructure of the current wire rod steel grade during the cooling process.

[0082] Initialize the cellular automata simulation space based on grid cells, set the initial nucleus positions in the austenitic matrix, and configure the initial cooling conditions;

[0083] The phase transition driving force of each grid cell is calculated based on the temperature field formed by the temperature change of the wire rod surface and the composition field formed by the material properties.

[0084] It should be noted that the cellular automata simulation space is a digital space constructed based on discrete grid cells to simulate the evolution of microstructure. Each grid cell corresponds to a tiny volume element of the micro-region of the wire rod, and can be assigned physical properties such as temperature, composition, and phase state. The state changes of the grid cells follow preset phase transformation rules and neighborhood interaction rules. The purpose of this step is to build a simulation carrier for the microstructure evolution of the transformation of austenite to pearlite / sorbite during the wire rod cooling process. By setting the initial nucleus position, the starting point of the phase transformation is determined, and the initial cooling conditions are configured to restore the actual cooling conditions of the Stellmore air-cooled roller conveyor.

[0085] The Gibbs free energy difference of each grid cell is used as the core calculation index. First, the real-time temperature field of the grid cell is obtained by combining the surface temperature change data of the wire rod with the heat conduction equation. The composition field distribution of the grid cell is determined based on the material property data. Then, the Gibbs free energy of the austenite phase and the pearlite / sorbite phase under the current temperature and composition conditions is obtained through the thermodynamic database. The difference between the two is the phase transformation driving force of the grid cell.

[0086] Based on the phase transition driving force of each grid cell and combined with the predefined growth competition rules of adjacent phase regions, the phase state transition probability of adjacent grid cells is calculated. The co-growth of pearlite clusters and the lamellar coarsening process are simulated through a probabilistic selection mechanism, and the growth morphology of pearlite clusters and the distribution of phase transition units are output.

[0087] It should be noted that the target microstructure of cold-heading steel wire rod is mainly pearlite / sorbite, and sorbite is essentially fine lamellar pearlite. In the specific simulation process, by adjusting parameters such as the cooling rate of the temperature field and the phase transformation driving force threshold, the microstructure evolution from conventional pearlite to fine lamellar sorbite can be simulated. When the cooling rate increases, the nucleation rate increases and the interlamellar spacing decreases, and the simulated microstructure is sorbite. When the cooling rate is moderate, the interlamellar spacing is relatively large, and the corresponding microstructure is pearlite. Therefore, this invention uses pearlite as the core characterization object, which can achieve unified coverage of the simulation process of both pearlite and sorbite target microstructures.

[0088] This study establishes a standardized spatial carrier and initial benchmark for simulating the dynamic evolution of microstructure during the cooling process of wire rod. By dividing the grid into units, the spatial scale of microstructure evolution is clarified. Initial nucleus positions are set in the austenite matrix to determine the initiation core sites of phase transformation. Initial cooling conditions are configured to recreate the real initial working conditions of wire rod cooling on the Steyrmo air-cooled roller conveyor. This provides a realistic and quantifiable simulation basis for subsequent calculation of phase transformation driving force, simulation of phase state transition probability, and pearlite cluster growth process.

[0089] By constructing a cellular automata simulation space based on grid cells and setting initial conditions, the initial state and initial environmental parameters of phase transition can be accurately anchored. By combining the temperature field and composition field to calculate the phase transition driving force and simulating the phase transition process according to the growth competition rules, the collaborative growth of pearlite clusters and the dynamic behavior of lamellar coarsening during the cooling process of wire rod can be realistically reproduced. The growth morphology of pearlite clusters and the distribution of phase transition units can be accurately output, providing high-precision microstructure evolution data support for multi-field coupled systems, improving the accuracy of phase transition law analysis and phase transition temperature range division, and thus providing a reliable basis for the formulation of segmented cooling strategies.

[0090] The calculation of the phase state transition probability of adjacent grid cells, based on the phase transition driving force of each grid cell and combined with predefined adjacent phase region growth competition rules, includes:

[0091] Calculate the phase transition driving energy difference between each grid cell and its adjacent grid cells; based on the crystallographic orientation relationship between adjacent grid cells, calculate the phase transition driving energy difference corrected by the interface energy change caused by the interface orientation difference.

[0092] It should be noted that crystallographic orientation relationships refer to the orientational correspondence between the crystal lattices of the austenite matrix and the newly formed pearlite / sorbite phase in adjacent grid units. A commonly used example is the typical KS orientation relationship in the pearlite transformation (the {111} crystal plane of austenite is parallel to the {110} crystal plane of ferrite in pearlite, and the austenite...). <110> Crystal orientation parallel to ferrite <111> The orientation relationship (or NW orientation) directly determines the degree of atomic arrangement matching between adjacent phase interfaces and affects the magnitude of the interface energy.

[0093] Based on the material property data of wire rod steel, the crystallographic orientation characteristics of austenite and pearlite / sorbite during the phase transformation process are determined. Each grid cell is assigned a corresponding crystallographic orientation parameter, and the orientation difference angle between adjacent grid cells is calculated. The value of the interface energy caused by the interface orientation difference is determined by combining the orientation difference angle. The change of interface energy will directly affect the energy balance of the phase transformation process. The interface energy value is converted into a correction coefficient for the phase transformation driving energy difference. Finally, the correction coefficient is used to adjust the phase transformation driving energy difference between adjacent grid cells obtained in the initial calculation, so as to obtain an accurate phase transformation driving energy difference considering the influence of interface orientation, thereby improving the accuracy of subsequent phase state transition probability calculation.

[0094] The corrected driving energy difference is input into a predefined adjacent phase region growth competition rule. This rule is then used to compare the phase state, relative supercooling, and historical evolution stage of the current mesh element with those of its neighbors, comprehensively determining the preferred direction of change. Specifically, this includes:

[0095] Step 1: Determine whether each element is in the austenitic matrix state or the pearlitic phase state by comparing the phase state parameters of the current grid element and all adjacent grid elements. Calculate the phase state type of each element and assign the corresponding basic score according to the phase state priority rule.

[0096] Step 2: Calculate the relative supercooling value of each unit, which is the difference between the actual temperature and the critical temperature of phase change. According to the relative supercooling priority rule, map the relative supercooling value to the corresponding score and multiply it by the preset weight coefficient.

[0097] Specifically, the growth competition rules for adjacent phase regions are based on the principles of metal phase transformation kinetics and crystallography. They are quantitative rules used to determine the priority order of phase transformation between adjacent grid units. The core content includes the phase state priority rule, that is, the unit that has undergone phase transformation has a higher priority than the untransformed austenite matrix unit, the lamellar pearlite unit has a higher priority than the granular pearlite unit, and the relative undercooling priority rule, that is, the unit with a greater relative undercooling has a stronger phase transformation driving force and a higher priority.

[0098] Step 3: Retrieve the historical evolution data of each unit to determine whether it is in the early stage of crystal nucleation, the middle stage of growth, or the late stage of growth. Assign corresponding scores according to the priority rules of historical evolution stages and multiply by the preset weight coefficient.

[0099] Step 4: Weight the phase state score of each unit relative to the supercooling score and the historical evolution stage score to obtain the comprehensive competition score of each adjacent unit. Finally, compare the comprehensive competition scores of all adjacent units. The unit with the highest score is the preferred transformation direction. If there are multiple units with the same score, further determine the direction based on the crystallographic orientation relationship and select the unit with the lowest interface energy as the preferred transformation direction.

[0100] Specifically, the historical evolution stage priority rule means that units in the middle stage of crystal nucleus growth have higher priority than units in the early stage of crystal nucleation and the end stage of growth. The rule will preset the weight coefficient of each priority index and determine the final competitive priority through weighted calculation.

[0101] Based on the preferred transition direction, a competitive weight is assigned to each adjacent grid cell. After normalizing the competitive weights of all adjacent grid cells, they are converted into probability values, thus obtaining the phase state transition probability of each adjacent grid cell in the next simulation step.

[0102] It should be noted that by calculating the difference in phase transition driving energy between grid cells and adjacent cells and combining it with crystallographic orientation relationships to correct the interface energy change, the accuracy of phase transition driving energy calculation can be improved. Based on the corrected driving energy difference and the predefined growth competition rules of adjacent phase regions, the priority direction of phase state transition can be scientifically determined. The phase transition probability obtained through competition weight allocation and normalization processing can ensure the rationality and practicality of the phase transition process simulation. Ultimately, high-precision simulation of the co-growth of pearlite clusters and the coarsening process of lamellar layers can be achieved. The output data on the growth morphology of pearlite clusters and the distribution of phase transition units can provide reliable microstructure evolution basis for multi-field coupled systems and improve the accuracy of phase transition law analysis and phase transition temperature range division.

[0103] The temperature and composition fields are dynamically updated based on the current pearlite cluster growth morphology and phase transformation unit distribution. The phase transformation driving force of each grid unit is iteratively calculated based on the updated temperature and composition fields until the iteration number is met or the cooling termination condition is met. The simulation results of pearlite cluster morphology and phase distribution are output to describe the dynamic evolution of the microstructure of the current wire rod steel grade during the cooling process.

[0104] It should be noted that in each iteration of the cellular automata simulation, based on the phase state transition probability, a probability selection mechanism through random sampling is used to determine whether adjacent grid cells undergo phase transition. For cells that undergo phase transition, the nucleus growth process is simulated and the phase state of the cells is updated. The lamellar spacing parameter is adjusted in combination with temperature field change data. In multiple iterations, the dynamic simulation of the co-growth of pearlite clusters and the lamellar coarsening process is realized, and finally, the growth morphology of pearlite clusters and the distribution of phase transition cells at different simulation times are output.

[0105] Based on the phase transformation law of the current wire rod steel grade during the cooling process, the phase transformation initiation temperature and phase transformation termination temperature are calculated. Using the phase transformation initiation temperature and phase transformation termination temperature as boundaries, the phase transformation temperature range of the current wire rod from austenite to pearlite / sorbite is divided.

[0106] It should be noted that the operating status of the segmented fans, roller speed, ambient temperature and humidity of the Stellmore air-cooled roller conveyor are all dynamically changing. These changes will affect the phase change process by influencing the temperature field distribution of the wire rod. The general phase change temperature range cannot meet the dynamic adjustment requirements under different working conditions. The multi-field coupling system can deeply couple the cooling actuator parameters, environmental data, material properties with the temperature field and microstructure evolution field, and correct the phase change temperature range in real time. At the same time, it can make up for the lack of direct measurement of the core temperature and phase change process of the wire rod on site, accurately capture the phase change range difference along the length of the roller conveyor, and provide a precise basis for the dynamic control of the segmented cooling strategy.

[0107] In this embodiment, cold heading steel wire rod SWRCH35K is used as the controlled object. Its chemical composition (mass fraction) is C: 0.33%-0.38%, Si: 0.15%-0.35%, Mn: 0.60%-0.90%, P≤0.030%, S≤0.035%, the rolling exit temperature is 920℃, and the target coiling temperature is 600℃.

[0108] The specific implementation process for simulating the dynamic evolution of microstructures and calculating the probability of phase transitions is as follows:

[0109] The simulation space of the cellular automata based on a 200×200 grid cell was initialized. The grid cell size was set to 5μm×5μm, corresponding to a micro-region of 1mm×1mm in the wire rod cross section. In the austenitic matrix, 120 initial nucleus positions were randomly set according to the austenitic grain size of SWRCH35K steel (initial grain diameter of about 20μm) to ensure uniform distribution of nuclei. At the same time, the initial cooling conditions were configured as an initial ambient temperature of 28℃, an initial wind speed of 3m / s, and a thermal convection coefficient of 80W / (m²·K).

[0110] Based on the temperature field (real-time temperature at different positions along the length of the roller conveyor) obtained by inverting the temperature change data of the wire rod surface and the composition field determined by the material properties, the phase transformation driving force of each grid cell is calculated. Specifically, the Gibbs free energy difference between the austenite and pearlite phases at the current temperature and composition is calculated using Thermo-Calc software, and the basic driving force is obtained as 180 J / mol. Then, interface energy and strain energy are introduced for correction, and the effective phase transformation driving force of each grid cell is finally obtained as 152-160 J / mol. Subsequently, based on the phase transformation driving force of each grid cell and combined with the predefined adjacent phase region growth competition rule, the phase state transition probability of adjacent grid cells is calculated. First, the phase transformation driving energy difference between each grid cell and 8 adjacent cells (range 5-12 J / mol) is calculated. Then, based on the KS crystallographic orientation relationship (austenite {111} / / ferrite {110}, austenite...), the phase transformation driving force is calculated. <110> Ferrite <111> ) Calculate the interface orientation difference angle between adjacent units, for example, 5°-15°, and then obtain the interface energy change, and use the change to correct the phase change driving energy difference.

[0111] The corrected driving energy difference is input into a predefined adjacent phase region growth competition rule. This rule presets a phase state priority weight of 0.4, a relative supercooling priority weight of 0.3, and a historical evolution stage priority weight of 0.3. By comparing the phase state (austenite is untransformed, pearlite is transformed), relative supercooling (50-80℃, dynamically changing with the cooling process), and historical evolution stage (early nucleation, middle growth, and late growth) of the current grid cell with each neighboring cell, the preferred transformation direction is comprehensively determined. For example, when a neighboring cell is pearlite and its relative supercooling is more than 20℃ higher than that of the current grid cell, this direction is determined to be the preferred transformation direction. Based on the preferred transformation direction, a competition weight of 0.1-0.8 is assigned to each adjacent grid cell. The competition weights of all adjacent cells of the same current grid cell are normalized and converted into probability values ​​in the range of 0-1 to obtain the phase state transformation probability of each adjacent grid cell in the next simulation step.

[0112] The co-growth of pearlite clusters and lamellar coarsening process were simulated using a probabilistic selection mechanism. In each simulation step, some adjacent elements were randomly selected based on the phase state transition probability to complete the transformation from austenite to pearlite. Simultaneously, the lamellar spacing was adjusted according to the lamellar coarsening kinetic model (increasing from an initial 0.2 μm to 0.5 μm as the simulation progressed). The pearlite cluster growth morphology and phase transformation element distribution of each simulation step were output. The temperature field and composition field were dynamically updated based on the current pearlite cluster growth morphology and phase transformation element distribution. The phase transformation driving force of each grid element was iteratively calculated based on the updated temperature field and composition field. The iteration step size was set to 0.1 s. The simulation was stopped when 200 iterations were completed and the pearlite phase transformation volume fraction reached more than 95%, satisfying the cooling termination condition. Finally, the simulation results describing the pearlite cluster morphology and phase distribution of the microstructure during the cooling process of SWRCH35K steel were output. These results provide accurate microstructure data support for subsequent phase transformation law analysis and phase transformation temperature range division.

[0113] S3. Based on the phase change temperature range, the target cooling path of the wire rod in the Stellmore air-cooled roller conveyor cooling process is divided into a high-temperature section, a phase change section, and a low-temperature section. Using strong convection cooling control in the high-temperature section, isothermal control in the phase change section, and low-stress slow cooling control in the low-temperature section as temperature control targets, and using the temperature boundaries of the high-temperature section, phase change section, and low-temperature section as constraints, the optimal segmented cooling strategy is generated, specifically including:

[0114] Based on the phase change temperature range, the target cooling path of wire rod in the Steyrmore air-cooled roller cooling process is divided into a high-temperature section, a phase change section, and a low-temperature section.

[0115] The high-temperature section is the temperature range from the rolling exit temperature to the phase transformation initiation temperature, the phase transformation section is the temperature range within the phase transformation temperature range, and the high-temperature section is the temperature range from the phase transformation termination temperature to the target coiling temperature.

[0116] It should be noted that the high-temperature section can rapidly reduce the wire rod temperature to the phase transformation initiation temperature through strong convection cooling, avoiding abnormal growth of austenite grains and laying the foundation for subsequent phase transformation microstructure refinement. The phase transformation section uses isothermal control to precisely match the phase transformation temperature range, ensuring a full and uniform transformation of austenite to pearlite / sorbite and improving the consistency of microstructure. The low-temperature section uses low-stress slow cooling control to reduce the risk of internal stress and deformation cracking caused by excessive temperature difference in the wire rod.

[0117] Based on the phase transition correlation features, a spatiotemporal probability distribution map of the phase transition temperature range is generated. The spatiotemporal probability distribution map is then mapped to a low-dimensional space using manifold learning technology to extract low-dimensional manifold lines that characterize the essence of the cooling process.

[0118] It should be noted that the phase transition correlation features obtained from the echo state network mining based on the integrated sparse reservoir, combined with the phase transition law data output by the multi-field coupling system, are used to statistically analyze the frequency of the wire rod temperature falling within the phase transition initiation and termination temperature ranges under different spatiotemporal coordinates. The phase transition probability value corresponding to each spatiotemporal coordinate is calculated through the probability density function. These probability values ​​are then correlated and integrated with the spatiotemporal coordinates to generate a spatiotemporal probability distribution map of the phase transition temperature range with spatial location as the horizontal axis, time as the vertical axis, and probability values ​​as color gradients. Through the dimensionality reduction algorithm in manifold learning technology, the high-dimensional spatiotemporal probability distribution map data is mapped to a low-dimensional space. The continuous curve that reflects the cooling process is extracted from the low-dimensional mapping result, that is, the low-dimensional manifold line that characterizes the essence of the cooling process.

[0119] The target temperature regulation is to use high-temperature section strong convection cooling control, phase change section isothermal control and low-stress slow cooling control as the temperature regulation objectives, and the temperature boundaries of the high-temperature section, phase change section and low-temperature section as the constraints. The target cooling path optimization problem is transformed into a problem to be solved on a low-dimensional manifold.

[0120] The optimal continuous smooth path that satisfies the temperature control objectives at each stage is found on the low-dimensional manifold. The optimal continuous smooth path is then mapped back to the original feature space to generate a segmented cooling strategy that satisfies the temperature control objectives.

[0121] Among these steps, the optimal continuous smooth path satisfying the temperature control objectives at each stage is solved on the low-dimensional manifold. This optimal continuous smooth path is then mapped back to the original feature space to generate a piecewise cooling strategy that satisfies the temperature control objectives.

[0122] On the low-dimensional manifold line that characterizes the essence of the cooling process, the rolling exit temperature and the target coil temperature are mapped to the start and end points on the low-dimensional manifold line. The temperature control targets are strong convection rapid cooling in the high-temperature section, isothermal maintenance in the phase transformation section, and low-stress slow cooling in the low-temperature section. The temperature boundaries of the high-temperature section, the phase transformation section, and the low-temperature section are used as constraints. The optimal continuous smooth path is solved using the Riemann geometric optimization algorithm.

[0123] The method for finding the optimal continuous smooth path using the Riemann geometric optimization algorithm includes:

[0124] The original low-dimensional manifold is given an equivalent Riemannian metric structure. A path planning problem that satisfies the temperature control objective and boundary constraints is established on the low-dimensional manifold. The problem is then mapped to a discrete numerical optimization model with an equivalent Riemannian metric structure in Euclidean space through coordinate transformation.

[0125] The objective function is to minimize the total length of the path under the equivalent Riemannian metric structure. The temperature control objective and boundary constraints are transformed into equivalent constraints of discrete coordinate points. The discrete numerical optimization model is solved by the solver to obtain a discrete coordinate sequence that satisfies the temperature control objective and boundary constraints.

[0126] The optimal continuous smooth path is obtained by interpolating the discrete coordinate sequence using the smooth interpolation method.

[0127] It should be noted that the low-dimensional manifold lines extracted based on manifold learning are given an equivalent Riemannian metric structure. This structure is essentially a non-Euclidean geometric metric method adapted to the characteristics of the cooling process. Its core is to define the distance measurement rule between any two points on the manifold according to the physical meaning of the cooling path, and to convert the traditional Euclidean distance into a weighted distance that fits the temperature control constraints and phase transition laws. This constructs a geometric space that conforms to the essential characteristics of the cooling process. Subsequently, in this Riemannian metric space, the rolling exit temperature and the target coil temperature are mapped as the starting point and ending point of the path planning. The temperature control targets of strong convection in the high-temperature section, isothermal in the phase transition section, and slow cooling in the low-temperature section, as well as the temperature boundaries of each stage, are transformed into the constraints of the path on the manifold. The optimization problem of path planning is established, and then the optimization problem in the Riemannian metric space is mapped to Euclidean space through coordinate transformation, which is transformed into a discrete optimization model that can be solved numerically.

[0128] Using the minimization of the total path length under the Riemann metric structure as the objective function, the discrete optimization model is solved using a numerical solver to obtain a discrete coordinate sequence that satisfies all constraints. Finally, a smooth interpolation method is used to interpolate the discrete coordinate sequence to fill the gaps between coordinate points and generate a continuous and smooth optimal path. This ensures that the path can be directly mapped back to the original feature space and transformed into executable cooling process parameters.

[0129] Specifically, the equivalent Riemannian metric structure refers to introducing a metric tensor of Riemannian geometry on the basis of a low-dimensional manifold. Based on physical parameters such as the rate of temperature change and phase transition sensitivity during the cooling process, a local distance metric rule is defined for each point on the manifold, so that the distance between two points on the manifold can reflect the actual control cost and difficulty of the cooling process. For example, the metric weight is increased in the manifold region corresponding to the phase transition segment to ensure that the temperature change of the optimized path is more gradual in that region.

[0130] The optimal continuous smooth path obtained by the solution is restored to the continuous process parameter sequence in the original feature space through inverse mapping. Based on the pre-divided temperature boundaries of the high temperature section, phase transition section and low temperature stage, the optimal temperature control parameters corresponding to each stage are identified and extracted from the continuous process parameter sequence.

[0131] It should be noted that by using manifold learning technology, high-dimensional, strongly coupled multi-source time-series cooling data is mapped into low-dimensional manifolds that characterize the essential features of the cooling process. This transforms the complex cooling path optimization problem from the high-dimensional original feature space to the low-dimensional manifold space, reducing the complexity of the solution. Then, an equivalent Riemannian metric structure is assigned to the low-dimensional manifold, and geometric metric rules that fit the physical laws of the cooling process are constructed. Next, the path planning problem in the Riemannian manifold space is transformed into a discrete numerical optimization model with an equivalent Riemannian metric structure in Euclidean space through coordinate transformation. The temperature control objectives of strong convection in the high-temperature section, isothermal in the phase transition section, and low-stress slow cooling in the low-temperature section, as well as the temperature boundaries of each stage, are transformed into constraints of discrete coordinate points. The objective function is to minimize the total path length under the Riemannian metric structure, thereby achieving the numerical solution of the optimal path.

[0132] Transforming the complex high-dimensional cooling path optimization problem into a low-dimensional manifold space can effectively eliminate redundant noise information in the original data, focus on the core features of the cooling process, and significantly reduce the complexity and computational power consumption of optimization calculations. By giving the low-dimensional manifold an equivalent Riemannian metric structure and mapping it to a discrete numerical optimization model in Euclidean space, the temperature control target and boundary constraints can be accurately quantified into mathematical constraints, ensuring that the optimization path meets the process requirements of each stage and avoids optimization results that deviate from actual production conditions.

[0133] This embodiment uses cold heading steel wire rod SWRCH35K as the controlled object. Its chemical composition (mass fraction) is C: 0.33%-0.38%, Si: 0.15%-0.35%, Mn: 0.60%-0.90%, P≤0.030%, S≤0.035%, rolling exit temperature 920℃, target coiling temperature 600℃. Previously, the phase transformation initiation temperature of 830℃ and the phase transformation termination temperature of 680℃ were determined through a multi-field coupling system. Example S3 (cooling path analysis) The process of segment division and optimal strategy generation is as follows: First, the cooling path is divided according to the phase transition temperature range: high temperature segment 920-830℃, phase transition segment 830-680℃, and low temperature segment 680-600℃. Based on the phase transition correlation characteristics and phase transition law data mined by the echo state network of the integrated sparse reservoir, the frequency of temperature falling into 830-680℃ at each spatiotemporal coordinate along the roller conveyor 0-45m and cooling 0-30s is statistically analyzed. The phase transition probability value is calculated using the probability density function to generate a spatiotemporal probability distribution map. The t-SNE algorithm is used to map the problem to a 2D low-dimensional space and extract the low-dimensional manifold representing the essence of cooling. Then, the cooling path optimization problem is transformed into a problem of solving on the low-dimensional manifold, mapping 920℃ and 600℃. With the temperature control targets and boundaries at each stage as constraints, the Riemannian geometric optimization algorithm is used to solve for the optimal path. An equivalent Riemannian metric structure is assigned to the low-dimensional manifold, establishing a path planning problem and mapping it to a discrete numerical optimization model in Euclidean space. The goal is to minimize the total path length, incorporating cooling in the high-temperature segment. Constraints such as a cooling rate ≥12℃ / s and a phase transition temperature fluctuation of ±5℃ are applied. The optimal continuous smooth path is obtained by solving the problem using a solver (e.g., the IPOPT solver) and cubic spline interpolation. Finally, the path is inversely mapped back to the original feature space process parameter sequence, and the optimal control parameters for each stage are extracted as shown in Table 1: high temperature section fan 45Hz, roller conveyor 1.5m / s; phase transition section fan 25Hz, roller conveyor 1.2m / s; low temperature section fan 10Hz, roller conveyor 1.0m / s, forming a segmented cooling strategy.

[0134] Table 1 Optimal control parameters for each stage

[0135]

[0136] To further verify the practical application effect of the cold heading steel wire rod air-cooled roller conveyor temperature control method of the present invention, the present invention was also applied and verified in the industrial production line of SCM435 cold heading steel wire rod (Φ12mm and Φ18mm specifications). At the same time, comparative experiments were carried out with the traditional fixed cooling process. The verification was completed from the dimensions of process control effect, microstructure and macroscopic mechanical properties. The test results are shown in Table 2 and Table 3.

[0137] Table 2 Comparison of Cooling Control Effects Between New and Old Processes

[0138]

[0139] Table 3 Comparison of Metallographic Structure and Properties between New and Old Processes

[0140]

[0141] As can be seen from the verification results in Tables 2 and 3, the segmented temperature control method of the present invention significantly improves the temperature control accuracy, wire rod structure uniformity, mechanical property stability and cold heading adaptability compared with the traditional fixed cooling process. It effectively solves the technical problems such as local overcooling / insufficient cooling, abnormal structure and large performance fluctuation in the traditional process.

[0142] S4. The optimal segmented cooling strategy is converted into control commands for the cooling actuator, which drives the cooling actuator to operate, thereby achieving temperature control of the wire rod on the Steyrmo air-cooled roller conveyor.

[0143] Figure 2 An embodiment of a temperature control system for a cold heading steel wire rod air-cooled roller conveyor according to the present invention is shown.

[0144] In this optional embodiment, a temperature control system for the cold heading steel wire rod air-cooled roller conveyor includes:

[0145] Data acquisition module 1 is used to collect multi-source time-series datasets of wire rod after high-speed wire rolling during the cooling process of the Stellmore air-cooled roller table. The multi-source time-series datasets include wire rod surface temperature change data, cooling actuator operating parameters, ambient temperature and humidity data, and material property data of the current wire rod steel grade.

[0146] Temperature range division module 2 is used to input multi-source time series datasets into the multi-field coupling system integrating thermo-dynamic-microstructure, analyze the phase transformation law of the current wire rod steel grade during the cooling process using the multi-field coupling system, and divide the phase transformation temperature range of the current wire rod from austenite to pearlite / sorbite structure according to the phase transformation law.

[0147] Temperature control solution module 3 is used to divide the target cooling path of wire rod in the Stellmore air-cooled roller cooling process into high temperature section, phase change section and low temperature section according to the phase change temperature range; with strong convection cooling control in high temperature section, isothermal control in phase change section and low stress slow cooling control in low temperature section as temperature control targets, and with the temperature boundaries of high temperature section, phase change section and low temperature section as constraints, the optimal segmented cooling strategy is generated.

[0148] Temperature control execution module 4 is used to convert the optimal segmented cooling strategy into control commands for the cooling actuator, driving the cooling actuator to operate, so as to achieve temperature control of the wire rod on the Steyrmo air-cooled roller conveyor.

[0149] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 3 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores static and dynamic information data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the above method embodiments.

[0150] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0151] In addition, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0152] In addition, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0153] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0154] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for temperature control of a cold heading steel wire rod air-cooled roller conveyor, characterized in that, The method includes: A multi-source time-series dataset was collected on the cooling process of wire rod after high-speed wire rod rolling on the Steyrmo air-cooled roller conveyor. The multi-source time-series dataset includes wire rod surface temperature change data, cooling actuator operating parameters, ambient temperature and humidity data, and material property data of the current wire rod steel grade. Multi-source time-series datasets are input into a multi-field coupling system integrating thermo-dynamic-microstructure. The phase transformation law of the current wire rod steel grade during the cooling process is analyzed using the multi-field coupling system. Based on the phase transformation law, the phase transformation temperature range of the current wire rod from austenite to pearlite / sorbite structure is divided. Based on the phase change temperature range, the target cooling path of wire rod in the Steyrmore air-cooled roller cooling process is divided into a high-temperature section, a phase change section, and a low-temperature section. The optimal segmented cooling strategy is generated by using high-temperature section strong convection cooling control, phase change section isothermal control and low-stress slow cooling control as temperature regulation targets, and the temperature boundaries of high-temperature section, phase change section and low-temperature section as constraints. The optimal segmented cooling strategy is converted into control commands for the cooling actuator, which drives the cooling actuator to operate, thereby achieving temperature control of the wire rod on the Steyrmo air-cooled roller conveyor. The process of inputting multi-source time-series datasets into a multi-field coupled system integrating thermo-dynamic-microstructure analysis, using this system to analyze the phase transformation law of the current wire rod steel grade during the cooling process, and dividing the phase transformation temperature range of the current wire rod from austenite to pearlite / sorbite based on the phase transformation law includes: Multi-source time-series datasets are input into an echo state network that integrates sparse reservoirs. The sparse reservoirs in the echo state network are used to mine phase transition correlation features between wire rod surface temperature change data, material property data, and environmental temperature and humidity data. A multi-field coupled system integrating thermodynamics, kinetics, and microstructure was constructed. The phase transformation correlation characteristics were input into the multi-field coupled system for thermodynamic, kinetic, and microstructure analysis in sequence. The results of thermodynamic, kinetic, and microstructure analysis were integrated to obtain the phase transformation law of the current wire rod steel grade during the cooling process. Among them, the microstructure analysis of multi-field coupled systems includes: Initialize the cellular automaton simulation space, and simulate the growth and evolution of preset crystal nuclei in the cellular automaton simulation space to realize the dynamic evolution simulation of the microstructure of the current wire rod steel grade during the cooling process. Based on the phase transformation law of the current wire rod steel grade during the cooling process, the phase transformation initiation temperature and phase transformation termination temperature are calculated. Using the phase transformation initiation temperature and phase transformation termination temperature as boundaries, the phase transformation temperature range of the current wire rod from austenite to pearlite / sorbite is divided.

2. The method for temperature control of a cold heading steel wire rod air-cooled roller conveyor according to claim 1, characterized in that, The initialization of the cellular automaton simulation space, in which the growth and evolution of preset crystal nuclei are simulated, realizes the dynamic evolution simulation of the microstructure of the current wire rod steel grade during the cooling process, including: Initialize the cellular automata simulation space based on grid cells, set the initial nucleus positions in the austenitic matrix, and configure the initial cooling conditions; The phase transition driving force of each grid cell is calculated based on the temperature field formed by the temperature change of the wire rod surface and the composition field formed by the material properties. Based on the phase transition driving force of each grid cell and combined with the predefined growth competition rules of adjacent phase regions, the phase state transition probability of adjacent grid cells is calculated. The co-growth of pearlite clusters and the lamellar coarsening process are simulated through a probability selection mechanism, and the growth morphology of pearlite clusters and the distribution of phase transition units are output. The temperature field and composition field are dynamically updated based on the current pearlite cluster growth morphology and phase transition unit distribution. The phase transition driving force of each grid unit is iteratively calculated based on the updated temperature field and composition field until the iteration number is met or the cooling termination condition is met. The output is the simulation results of pearlite cluster morphology and phase distribution, which describe the dynamic evolution of the microstructure of the current wire rod steel during the cooling process.

3. The method for temperature control of a cold heading steel wire rod air-cooled roller conveyor according to claim 2, characterized in that, The calculation of the phase state transition probability of adjacent grid cells based on the phase transition driving force of each grid cell and combined with the predefined adjacent phase region growth competition rule includes: Calculate the phase transition driving energy difference between each grid cell and its adjacent grid cells; based on the crystallographic orientation relationship between adjacent grid cells, calculate the phase transition driving energy difference corrected by the interface energy change caused by the interface orientation difference. The corrected driving energy difference is input into the predefined adjacent phase region growth competition rule. The adjacent phase region growth competition rule is used to compare the phase state, relative supercooling and historical evolution stage of the current grid cell with each neighboring cell, and to comprehensively determine the priority change direction. Based on the preferred transition direction, a competitive weight is assigned to each adjacent grid cell. After normalizing the competitive weights of all adjacent grid cells, they are converted into probability values, thus obtaining the phase state transition probability of each adjacent grid cell in the next simulation step.

4. The method for temperature control of a cold heading steel wire rod air-cooled roller conveyor according to claim 1, characterized in that, The target cooling path of the wire rod in the Stellmore air-cooled roller conveyor cooling process is divided into a high-temperature section, a phase change section, and a low-temperature section based on the phase change temperature range. The optimal segmented cooling strategy is generated by using strong convection cooling control in the high-temperature section, isothermal control in the phase change section, and low-stress slow cooling control in the low-temperature section as temperature control targets, and the temperature boundaries of the high-temperature section, phase change section, and low-temperature section as constraints. Based on the phase change temperature range, the target cooling path of wire rod in the Steyrmore air-cooled roller cooling process is divided into a high-temperature section, a phase change section, and a low-temperature section. Among them, the temperature range of the high-temperature section is from the rolling exit temperature to the phase transformation initiation temperature, the temperature range of the phase transformation section is the temperature within the phase transformation temperature range, and the temperature range of the high-temperature section is from the phase transformation termination temperature to the target coiling temperature. Based on the phase transition correlation features, a spatiotemporal probability distribution map of the phase transition temperature range is generated. The spatiotemporal probability distribution map is then mapped to a low-dimensional space using manifold learning technology to extract low-dimensional manifold lines that characterize the essence of the cooling process. The target temperature regulation is to use strong convection cooling control in the high-temperature section, isothermal control in the phase change section, and low-stress slow cooling control in the low-temperature section as the temperature regulation objectives, and the temperature boundaries of the high-temperature section, phase change section, and low-temperature section as the constraints. The target cooling path optimization problem is transformed into a solution problem on a low-dimensional manifold. The optimal continuous smooth path that satisfies the temperature control objectives at each stage is found on the low-dimensional manifold. The optimal continuous smooth path is then mapped back to the original feature space to generate a segmented cooling strategy that satisfies the temperature control objectives.

5. The method for temperature control of a cold heading steel wire rod air-cooled roller conveyor according to claim 4, characterized in that, The process of finding the optimal continuous smooth path that satisfies the temperature control objectives at each stage on a low-dimensional manifold, mapping the optimal continuous smooth path back to the original feature space, and generating a piecewise cooling strategy that satisfies the temperature control objectives includes: On the low-dimensional manifold line that characterizes the essence of the cooling process, the rolling exit temperature and the target coil temperature are mapped to the start and end points on the low-dimensional manifold line. The temperature control targets are strong convection rapid cooling in the high-temperature section, isothermal maintenance in the phase change section, and low-stress slow cooling in the low-temperature section. The temperature boundaries of the high-temperature section, the phase change section, and the low-temperature section are used as constraints. The optimal continuous smooth path is solved using the Riemann geometric optimization algorithm. The optimal continuous smooth path obtained by the solution is restored to the continuous process parameter sequence in the original feature space through inverse mapping. Based on the pre-divided temperature boundaries of the high temperature section, phase transition section and low temperature stage, the optimal temperature control parameters corresponding to each stage are identified and extracted from the continuous process parameter sequence.

6. The method for temperature control of a cold heading steel wire rod air-cooled roller conveyor according to claim 5, characterized in that, The method of using the Riemann geometric optimization algorithm to solve for the optimal continuous smooth path includes: The original low-dimensional manifold is given an equivalent Riemannian metric structure. A path planning problem that satisfies the temperature control objective and boundary constraints is established on the low-dimensional manifold. The problem is then mapped to a discrete numerical optimization model with an equivalent Riemannian metric structure in Euclidean space through coordinate transformation. The objective function is to minimize the total length of the path under the equivalent Riemannian metric structure. The temperature control objective and boundary constraints are transformed into equivalent constraints of discrete coordinate points. The discrete numerical optimization model is solved by the solver to obtain a discrete coordinate sequence that satisfies the temperature control objective and boundary constraints. The optimal continuous smooth path is obtained by interpolating the discrete coordinate sequence using the smooth interpolation method.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images