A method and system for closed-loop control of cold heading steel wire rod rolling temperature based on PID

Abstract

The application provides a kind of PID-based cold heading steel wire rod rolling temperature closed loop control method and system, relating to cold heading steel wire rod rolling process control technical field, comprising the following steps: according to the process requirement of cold heading steel wire rod, the temperature control parameter of rolling mill rolling area in different rolling stages of cold heading steel wire rod is determined;Dynamic thermodynamic model of different rolling stages is constructed;Real-time state data and real-time temperature data are preprocessed;The extracted temperature control parameter is compared with the preprocessed real-time temperature data;Based on the dynamic thermodynamic model and the real-time temperature deviation, the control amount is calculated by PID algorithm, and the control instruction is output to the actuator of rolling mill;The actuator of rolling mill receives instruction and carries out temperature adjustment action.The application can effectively guarantee the temperature stability of each rolling stage, improve the product quality consistency of cold heading steel wire rod, and adapt to the dynamic change requirement of rolling process.

Description

Technical Field

[0001] This invention relates to the field of cold heading steel wire rod rolling process control technology, and more specifically, to a PID-based closed-loop temperature control method and system for cold heading steel wire rod rolling. Background Technology

[0002] Cold heading steel wire rod is mainly used for cold heading to produce fasteners such as bolts and nuts, or cold-formed parts, and has a wide range of applications. Rolling temperature is the core process parameter that determines the metallographic structure, mechanical properties, and forming quality of cold heading steel wire rod. The temperature control precision requirements vary significantly at different rolling stages (heating, roughing, finishing, and controlled cooling). Excessive temperature fluctuations can lead to defects such as cracks, uneven hardness, and cold heading cracks in the wire rod, seriously affecting the product qualification rate.

[0003] Currently, temperature control in cold heading steel wire rod rolling mostly adopts traditional PID open-loop or simple closed-loop control strategies, which have the following key technical problems: The differences in heat transfer dominance at different rolling stages are not considered (e.g., radiation conduction is dominant in the heating stage, while convection and frictional heat generation are dominant in the finishing stage). Using uniform control parameters and models fails to adapt to the temperature change patterns at each stage, resulting in insufficient control accuracy. Furthermore, there is a complex coupling relationship between mill status parameters (heating furnace power, cooling water volume, rolling speed, etc.) and temperature parameters. Traditional methods struggle to quantify the correlation strength between parameters and cannot accurately identify the sources and propagation paths of interference factors such as raw material composition fluctuations and equipment operating condition changes, easily leading to temperature runaway.

[0004] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0005] In view of this, the present invention provides a closed-loop control method and system for cold heading steel wire rod rolling temperature based on PID, so as to solve the above-mentioned problems.

[0006] To solve the above problems, the specific technical solution adopted by the present invention is as follows:

[0007] A closed-loop control method for temperature control in cold heading steel wire rod rolling based on PID control includes the following steps:

[0008] S1. Based on the process requirements of cold heading steel wire rod, determine the temperature control parameters of the rolling zone of the rolling mill for different rolling stages of cold heading steel wire rod;

[0009] S2. Based on the parametric gravity analysis method, combined with the historical state data of the rolling mill and the historical temperature data of the rolling area of ​​the rolling mill, a parametric gravity network is constructed. According to different rolling stages, the heat transfer mechanism and interference factors of the rolling process are analyzed in segments. Based on the segmented analysis results, dynamic thermodynamic models of different rolling stages are constructed to achieve temperature adjustment in different rolling stages.

[0010] S3. Obtain the real-time status data of the rolling mill and the real-time temperature data of the rolling area of ​​the rolling mill, and preprocess the real-time status data and real-time temperature data to obtain the preprocessed real-time status data and real-time temperature data.

[0011] S4. Based on the preprocessed real-time status data, determine the rolling stage of cold heading steel wire rod and extract the temperature control parameters of the rolling stage. Compare the extracted temperature control parameters with the preprocessed real-time temperature data to obtain the real-time temperature deviation.

[0012] S5. Based on the dynamic thermodynamic model and real-time temperature deviation, the control quantity is calculated through the PID algorithm, and the control command is output to the actuator of the rolling mill.

[0013] S6. The rolling mill's actuator receives instructions and performs temperature adjustment actions to achieve closed-loop stable control of the rolling temperature of cold heading steel wire rod.

[0014] Preferably, the parametric gravity analysis method, combined with pre-acquired historical state data of the rolling mill and historical temperature data of the rolling zone, constructs a parametric gravity network. Based on different rolling stages, it performs segmented analysis of the heat transfer mechanism and interference factors in the rolling process. Based on the segmented analysis results, it constructs dynamic thermodynamic models for different rolling stages, thereby achieving the following steps for different rolling stages:

[0015] S21. Perform timestamp alignment processing on the historical state data of the rolling mill and the historical temperature data of the rolling area of ​​the rolling mill, and construct a parametric gravity network based on the alignment results;

[0016] S22. Based on different rolling stages, the parametric gravitational network is segmented to obtain several network communities, and the heat transfer mechanism and interference factors of the network community corresponding to each rolling stage are analyzed.

[0017] S23. Based on the heat transfer mechanism and interference factors of each rolling stage, construct a dynamic thermodynamic model for each rolling stage.

[0018] Preferably, the step of performing timestamp alignment processing on the pre-acquired historical state data of the rolling mill and the historical temperature data of the rolling area of ​​the rolling mill, and constructing a parametric gravity network based on the alignment results and in conjunction with niche theory, includes the following steps:

[0019] S211. Using a time synchronization algorithm, the historical state data of the rolling mill and the historical temperature data of the rolling area of ​​the rolling mill are timestamped and a time-aligned data matrix containing temperature sequence and state parameter sequence is constructed.

[0020] S212. Based on the time-aligned data matrix, calculate the niche width of each state parameter, and filter out key state parameters that meet the preset width threshold to obtain a set of key state parameters.

[0021] S213. Based on the set of key state parameters, calculate the niche overlap between any two sets of state parameters, and select strongly correlated parameters that meet the preset overlap threshold to obtain a set of strongly correlated parameters.

[0022] S214. Based on the set of strongly correlated parameters and the niche width of each state parameter, construct a parameter gravity network with time points as nodes and the gravitational strength between parameters as edge weights.

[0023] Preferably, the construction of a parametric gravity network based on a set of strongly correlated parameters and the niche width of each state parameter, with time points as nodes and the gravitational strength between parameters as edge weights, includes the following steps:

[0024] S2141. Take each sampling time point as a network node. For any two network nodes, calculate the comprehensive state characterization value of any two network nodes based on the niche width of the state parameters.

[0025] S2142. Based on the comprehensive state characterization values ​​of any two network nodes, and combined with the weighted average of the niche overlap between their corresponding strongly correlated parameter pairs, calculate the basic gravitational strength between the two network nodes.

[0026] S2143. Based on the preset time-distance decay factor and the time interval between two network nodes, calculate the final gravitational strength.

[0027] S2144. Based on each network node, construct a parametric gravity network using the calculated gravitational strength as the weight of the directed edge.

[0028] Preferably, the step of segmenting the parametric gravitational network based on different rolling stages to obtain several network communities, and analyzing the heat transfer mechanism and interference factors of the network community corresponding to each rolling stage, includes the following steps:

[0029] S221. Determine the time interval for each rolling stage, and divide the parametric gravity network according to the time interval to obtain several network communities.

[0030] S222. Classify the network nodes in each network community and determine the heat transfer mechanism of each rolling stage through centrality analysis based on Shapley value.

[0031] S223. By combining the dynamic evolution characteristics of key state parameters and niche indicators in the network community, identify the interference factors in each rolling stage.

[0032] Preferably, the process of classifying network nodes in each network community and determining the heat transfer mechanism of each rolling stage through centrality analysis based on Shapley values ​​includes the following steps:

[0033] S2221. Divide the network nodes in each network community into heat transfer drive nodes and heat transfer response nodes.

[0034] S2222. A connectivity test algorithm is used to analyze the connectivity of network nodes in the network community, and network nodes that meet the connectivity threshold and contain heat transfer driving nodes and heat transfer response nodes are selected to obtain a feasible subset of heat transfer nodes.

[0035] S2223. Calculate the heat transfer utility of each feasible subset of heat transfer based on the preset utility function;

[0036] S2224. Calculate the total heat transfer utility of the current network community based on the heat transfer utility of each feasible subset of heat transfer.

[0037] S222. Based on the total heat transfer efficiency of the current network community, calculate the heat transfer contribution centrality of each network node based on the Shapley value, and determine the heat transfer mechanism of the rolling stage based on the heat transfer contribution centrality ranking results.

[0038] Preferably, the step of identifying interference factors in each rolling stage by combining the dynamic evolution characteristics of key state parameters in the network community with niche indicators includes the following steps:

[0039] S2231. For key state parameters in the network community, calculate the mean, variance and mutation rate of each key state parameter respectively.

[0040] S2232. Based on the mean, variance and mutation rate, combined with the historical normal fluctuation range and niche width of key state parameters, identify and mark key state parameters that show abnormal fluctuations, and preliminarily determine them as potential disturbance factors.

[0041] S2233. Analyze the niche overlap of potential interference sources within the network community, identify other key state parameters that cause changes in niche overlap due to their abnormal fluctuations, in order to determine the set of parameters affected by interference and the propagation path of interference.

[0042] S2234. By comprehensively considering potential interference factors, the set of parameters affected by interference, and the propagation path of interference, assess the intensity and scope of each interference factor in order to identify interference factors in the current rolling stage.

[0043] The beneficial effects of this invention are as follows:

[0044] 1. This invention achieves precise mechanism adaptation and interference identification by determining temperature control parameters in stages and constructing a dynamic thermodynamic model by combining parameter gravity analysis. After real-time data preprocessing, stage judgment and deviation calculation, the PID algorithm and the actuator are linked in a closed loop. It has the characteristics of stage adaptability, interference resistance and control accuracy, which can effectively ensure the temperature stability of each rolling stage, improve the consistency of cold heading steel wire rod product quality and adapt to the dynamic changes in the rolling process.

[0045] 2. This invention constructs a parametric gravity network through timestamp alignment and niche theory, and forms network communities by segmenting the rolling stages. Through node classification and Shapley value centrality analysis, the phased heat transfer mechanism is accurately analyzed. By identifying interference factors through the dynamic evolution characteristics of key parameters and niche indicators, a dynamic thermodynamic model adapted to each stage is constructed, realizing precise phased adaptation of rolling temperature control. It combines the scientific rigor of mechanism support with the targeted nature of interference identification, providing an accurate and reliable model foundation for subsequent closed-loop control, and effectively improving the stability and adaptability of temperature control. Attached Figure Description

[0046] 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 described 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. In the drawings:

[0047] Figure 1 This is a flowchart according to an embodiment of the present invention;

[0048] Figure 2 This is a flowchart illustrating the construction process of a dynamic thermodynamic model according to an embodiment of the present invention;

[0049] Figure 3 This is a flowchart illustrating the determination of the heat transfer mechanism according to an embodiment of the present invention;

[0050] Figure 4 This is a principle block diagram according to an embodiment of the present invention;

[0051] Figure 5 This is one of the performance test diagrams before optimization of the wire rod production process according to an embodiment of the present invention;

[0052] Figure 6This is the second performance test diagram of the wire rod production process before optimization according to an embodiment of the present invention;

[0053] Figure 7 This is the third performance test diagram of the wire rod production process before optimization according to an embodiment of the present invention;

[0054] Figure 8 This is one of the performance test diagrams after optimizing the wire rod production process according to an embodiment of the present invention;

[0055] Figure 9 This is the second performance test diagram after optimizing the wire rod production process according to an embodiment of the present invention;

[0056] Figure 10 This is the third performance test diagram after optimizing the wire rod production process according to an embodiment of the present invention.

[0057] In the picture:

[0058] 1. Temperature control parameter determination module; 2. Model building module; 3. Real-time data acquisition and processing module; 4. Real-time temperature deviation calculation module; 5. Control quantity calculation module; 6. Dynamic temperature adjustment module. Detailed Implementation

[0059] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0060] According to an embodiment of the present invention, a closed-loop control method and system for temperature control in cold heading steel wire rod rolling based on PID is provided.

[0061] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments, such as... Figure 1-3 As shown, according to a first embodiment of the present invention, a closed-loop control method for the rolling temperature of cold heading steel wire rod based on PID is provided, comprising the following steps:

[0062] S1. Based on the process requirements of cold heading steel wire rod, determine the temperature control parameters of the rolling zone of the rolling mill for different rolling stages of cold heading steel wire rod;

[0063] It should be noted that the process requirements for cold heading steel wire rod include the steel grade characteristics (such as the phase transformation temperature range, thermal conductivity, and critical temperature for plastic deformation of materials like SWRCH35KM and SCM435), cold heading performance requirements (such as yield strength, elongation after fracture, and cold heading cracking rate control standards), dimensional accuracy requirements, and equipment constraints of the rolling mill, including the maximum power of the heating furnace, the upper limit of the cooling system flow rate, and the rolling speed range.

[0064] The rolling process of cold heading steel wire rod is divided into clearly defined rolling stages, typically including the heating stage, the roughing cooling stage, the intermediate cooling stage, the finishing rapid cooling stage, the sizing and temperature stabilization stage, and the wire drawing controlled cooling stage. The temperature control parameters for each stage need to be set according to their respective process functions. The specific content of the temperature control parameters includes the target temperature value, allowable fluctuation range, and control priority for each rolling stage. Some stages also require specific auxiliary control indicators. For example, the target value for the initial rolling temperature in the heating stage is set at 980-1020℃, with an allowable cross-sectional temperature difference of ≤15℃; the target value for the finishing exit temperature is 850-900℃, with an allowable fluctuation range of ±8℃; and the target value for the wire drawing temperature is 780-820℃, with an allowable fluctuation range of ±5℃. The determination of these parameters requires comprehensive reference to the steel grade's hot working process manual, historical production temperature data meeting quality requirements, and equipment performance test results.

[0065] S2. Based on the parametric gravity analysis method, combined with the historical state data of the rolling mill and the historical temperature data of the rolling area of ​​the rolling mill, a parametric gravity network is constructed. According to different rolling stages, the heat transfer mechanism and interference factors of the rolling process are analyzed in segments. Based on the segmented analysis results, dynamic thermodynamic models of different rolling stages are constructed to achieve temperature adjustment in different rolling stages.

[0066] As a preferred embodiment, the parametric gravity analysis method, combined with pre-acquired historical state data of the rolling mill and historical temperature data of the rolling zone, constructs a parametric gravity network. Based on different rolling stages, it performs segmented analysis of the heat transfer mechanism and interference factors in the rolling process. Based on the segmented analysis results, it constructs dynamic thermodynamic models for different rolling stages, thereby achieving the following steps for different rolling stages:

[0067] S21. Perform timestamp alignment processing on the historical state data of the rolling mill and the historical temperature data of the rolling area of ​​the rolling mill, and construct a parametric gravity network based on the alignment results;

[0068] As a preferred embodiment, the step of performing timestamp alignment processing on the pre-acquired historical state data of the rolling mill and the historical temperature data of the rolling area of ​​the rolling mill, and constructing a parametric gravity network based on the alignment results and in conjunction with niche theory, includes the following steps:

[0069] S211. Using a time synchronization algorithm, the historical state data of the rolling mill and the historical temperature data of the rolling area of ​​the rolling mill are timestamped and a time-aligned data matrix containing temperature sequence and state parameter sequence is constructed.

[0070] It should be noted that the historical status data of the rolling mill includes parameters reflecting the operating status of the equipment, such as furnace power, cooling water volume, rolling speed, and gas flow. The historical temperature data of the rolling mill's rolling area includes parameters reflecting the thermal state of the rolling process, such as the initial rolling temperature, roughing mill exit temperature, finishing mill exit temperature, and wire drawing temperature. The two types of data may have time misalignment due to differences in sensor acquisition frequencies and transmission delays. Time synchronization algorithms, such as the Dynamic Time Warping (DTW) algorithm and the benchmark calibration algorithm based on PLC system time, are used to uniformly calibrate the timestamps of the two types of raw data. After calibration, a time-aligned data matrix is ​​constructed: the rows of the matrix correspond to different sampling time points, the columns correspond to various status parameters and temperature parameters, and the matrix elements are the standardized values ​​of the corresponding time points and parameters, ultimately forming a two-dimensional structured data matrix of time and parameters.

[0071] S212. Based on the time-aligned data matrix, calculate the niche width of each state parameter, and select key state parameters that meet the preset width threshold (e.g., 0.4) to obtain a set of key state parameters.

[0072] Specifically, the formula for calculating niche breadth is:

[0073] ;

[0074] In the formula, W i T represents the niche width of each state parameter, N represents the total number of time points, i.e., the total number of samples taken during the entire analysis period, and T represents the total number of samples taken during the entire analysis period. j S represents the temperature value at the j-th time point. ij This represents the value of the i-th state parameter sequence at time j; max(S) i ) represents the maximum value in the i-th state parameter sequence, max(S) j ) represents the minimum value in the i-th state parameter sequence.

[0075] S213. Based on the set of key state parameters, calculate the niche overlap between any two sets of state parameters, and select strongly correlated parameters that meet the preset overlap threshold (e.g., 0.6) to obtain a set of strongly correlated parameters.

[0076] For the key parameter set {S k1 ,S k2 ,...,S km} Calculate any two sets of state parameters S ki With S kjNiche overlap O ij The formula is:

[0077] ;

[0078] In the formula, O ij Represents two sets of state parameters S ki With S kj Ecological niche overlap.

[0079] S214. Based on the set of strongly correlated parameters and the niche width of each state parameter, construct a parameter gravity network with time points as nodes and the gravitational strength between parameters as edge weights.

[0080] Specifically, a parametric gravitational network G=(V,E) is constructed with time point t as the node, where:

[0081] Node set V={t1,t2,...,t N} represents the discrete time points of the time series; the weights F of the edge set E are... ij Due to niche overlap O ij The temperature correlation strength of the parameter pair is defined jointly.

[0082] As a preferred embodiment, the construction of a parameter gravity network based on a set of strongly correlated parameters and the niche width of each state parameter, with time points as nodes and the gravitational strength between parameters as edge weights, includes the following steps:

[0083] S2141. Take each sampling time point as a network node. For any two network nodes, calculate the comprehensive state characterization value of any two network nodes based on the niche width of the state parameters.

[0084] It should be noted that each sampling time point is directly used as a node in the parameter gravity network because the influence of parameters on temperature during the rolling process is highly time-sensitive. For example, the cooling water volume at the current time point only has a significant impact on the finishing rolling temperature in the short period of time that follows. Using time points as nodes allows for the precise capture of the temporal dynamic characteristics of parameter correlations. Each node is naturally associated with the specific values ​​of all strongly correlated parameters at that time point, forming a correlation carrier between time points and multiple parameters.

[0085] For any two network nodes, denoted as nodes t1 and t2, corresponding to two different sampling time points, we first need to extract the values ​​of all strongly correlated parameters corresponding to these two nodes. Then, using the niche width of each strongly correlated parameter as a weight, we perform a weighted summation of the strongly correlated parameter values ​​for each node to obtain the comprehensive state characterization value of a single node. The niche width directly reflects the range and regulatory capacity of the parameter on temperature changes; the larger the parameter width, the higher its contribution to the overall effect of the node. The final comprehensive state characterization value is essentially a quantitative result of the synergistic effect of all strongly correlated parameters at each time point. The larger the value, the stronger the potential impact of the parameter combination at that time point on subsequent temperature changes.

[0086] S2142. Based on the comprehensive state characterization values ​​of any two network nodes, and combined with the weighted average of the niche overlap between their corresponding strongly correlated parameter pairs, calculate the basic gravitational strength between the two network nodes.

[0087] It should be noted that the comprehensive state characterization value of two nodes corresponds to the mass basis of the node, analogous to gravitational mass in physics. The larger the value, the stronger the potential correlation between the node and other nodes. For the strongly correlated parameter pairs shared by the two nodes, the weighted average of the overlap of all strongly correlated parameter pairs is calculated based on the niche overlap of each pair of parameters, with the sum of the niche widths of the pair of parameters as the weight. The gravitational basis strength is calculated by multiplying the comprehensive state characterization values ​​of the nodes by the weighted average of the overlap. The gravitational basis strength represents the inherent correlation between the two time points based on the interaction of strongly correlated parameters. The larger the value, the closer the parameter combinations at the two time points are on the path affecting temperature change, and the stronger the correlation.

[0088] S2143. Based on the preset time-distance decay factor and the time interval between two network nodes, calculate the final gravitational strength.

[0089] It should be noted that during the rolling process, the influence of the state parameters at the previous time point on the temperature at the next time point gradually weakens as the interval between the two time points increases. For example, the heating power at t1 has a significant impact on the initial rolling temperature at t2, but a smaller impact on the initial rolling temperature at t2. 10 The roughing temperature has a negligible effect. Final gravitational strength calculation formula and explanation:

[0090] Final gravitational strength F = fundamental gravitational strength × λ Δt ;

[0091] In the formula, F represents the final gravitational strength between two network nodes (i.e., the weight of subsequent network edges); λ ΔtThe decay term is time decay. When Δt=0 (at the same time point, it has no practical significance), the decay term is 1. As Δt increases, the decay term gradually decreases, causing F to decrease with the increase of time interval, which perfectly matches the time-dependent effect of parameters in the rolling process.

[0092] S2144. Based on each network node, construct a parametric gravity network using the calculated gravitational strength as the weight of the directed edge.

[0093] Specifically, constructing a parametric gravity network includes the following steps:

[0094] All sampling time points are imported into the network model as independent nodes in chronological order, with each node accompanied by its corresponding comprehensive state representation value.

[0095] The direction of the edge is defined according to the temporal sequence, pointing from the node with earlier time to the node with later time. The core reason is that the rolling process has an irreversible temporal sequence. The parameters at the previous time point will only affect the state at the subsequent time point, and will not have a reverse effect. Directed edges can accurately capture this unidirectional influence relationship.

[0096] The calculated final gravitational strength is directly used as the weight of the corresponding directed edge. To avoid network redundancy, such as weakly correlated edges interfering with the core rules, weakly correlated edges with weights below a preset threshold, such as 0.3, are removed, and only strongly correlated edges are retained.

[0097] The completed parametric gravity network can intuitively show the distribution of the strength of parameter correlations at different time stages. For example, the gravity strength between nodes is generally high during the heating stage (the parameter correlation is close), while the gravity strength between nodes fluctuates greatly during the finishing rolling stage.

[0098] S22. Based on different rolling stages, the parametric gravitational network is segmented to obtain several network communities, and the heat transfer mechanism and interference factors of the network community corresponding to each rolling stage are analyzed.

[0099] As a preferred embodiment, the step of segmenting the parametric gravitational network based on different rolling stages to obtain several network communities, and analyzing the heat transfer mechanism and interference factors of the network community corresponding to each rolling stage, includes the following steps:

[0100] S221. Determine the time interval for each rolling stage, and divide the parametric gravity network according to the time interval to obtain several network communities.

[0101] It should be noted that, in conjunction with the standard process flow for cold heading steel wire rod rolling, including heating, roughing, intermediate rolling, finishing, sizing, and wire cutting and controlled cooling, the time stamps of key process nodes are used as markers. For example, the time when the billet enters the heating furnace is the starting point of the heating stage, and the time when the billet leaves the heating furnace and enters the roughing mill is the ending point of the heating stage and the starting point of the roughing stage. At the same time, the typical parameter characteristics of each stage in historical production data are referenced, such as the heating furnace power remaining at a high level during the heating stage and the cooling water volume fluctuating frequently during the finishing stage, to assist in calibrating the time interval and ensure that the interval division is completely matched with the actual process.

[0102] Since the nodes of the parametric gravity network are themselves sampling time points, the segmentation essentially involves extracting corresponding nodes and associated edges according to time intervals. For each rolling stage's time interval, all network nodes within that interval are selected, i.e., time points, and all directed edges between these nodes are extracted, i.e., corresponding gravity strength weights, forming independent network communities. Each community corresponds to a rolling stage and includes the temporal relationships of all strongly correlated parameters within that stage. For example, the heating stage community focuses on the relationship between heating power, gas flow rate, and initial rolling temperature, while the finishing rolling stage community focuses on the relationship between cooling water volume, rolling speed, and finishing exit temperature. Each network community becomes a structured carrier of the parameter-temperature relationships for that rolling stage.

[0103] S222. Classify the network nodes in each network community and determine the heat transfer mechanism of each rolling stage through centrality analysis based on Shapley value.

[0104] In a preferred embodiment, the process of classifying network nodes in each network community and determining the heat transfer mechanism of each rolling stage through centrality analysis based on Shapley values ​​includes the following steps:

[0105] S2221. Divide the network nodes in each network community into heat transfer drive nodes and heat transfer response nodes.

[0106] It should be noted that the heat-driven node is a mill state parameter node that directly affects the rolling temperature change, including at least one of the following: furnace power node, cooling water flow node, rolling speed node, and gas flow node; the heat transfer response node is a temperature parameter node that reflects the rolling temperature change result, including at least one of the following: initial rolling temperature node, roughing mill exit temperature node, finishing mill exit temperature node, and wire drawing temperature node.

[0107] Specifically, when making the division, the node affiliation is determined by judging one by one through three dimensions: functional attributes, controllability, and causal relationship. The three dimensions must simultaneously meet the corresponding judgment criteria, as shown in Table 1.

[0108] Table 1 Judgment Criteria

[0109]

[0110] Furthermore, if there are nodes with ambiguous functions, such as the roll temperature node, which is affected by both the heat generated by rolling friction and subsequent rolling temperatures, the following two methods should be used to supplement the determination to ensure accurate classification:

[0111] Determine the dominant role; analyze the source and object of influence of this parameter. If the source of influence is mostly driving parameters, such as the temperature of the roll being affected by the rolling speed and cooling water volume, and the object of influence is only the temperature parameter, and it cannot be directly controlled, then it is determined to be an indirect response node and classified as a heat transfer response node.

[0112] Process priority determination: In conjunction with the key points of cold heading steel rolling process, if the parameter is the core control object of the on-site process, such as heating furnace power and cooling water volume, it is given priority to be determined as the heat transfer drive node; if it is the core monitoring indicator, such as the initial rolling temperature and the wire drawing temperature, it is given priority to be determined as the heat transfer response node.

[0113] In addition, the attributes of the heat transfer drive node and the heat transfer response node can be quantified to obtain the heat transfer characteristic parameters of each node.

[0114] It should be noted that the heat transfer driving node lies in the ability to actively control the rolling temperature. Therefore, quantification needs to focus on two core dimensions: one is the potential driving capability for temperature changes, and the other is the stability and reliability of the driving effect, ultimately forming two key heat transfer characteristic parameters, including heat transfer contribution and heat transfer reliability.

[0115] The heat transfer contribution is used to quantify the potential impact of a driving node on rolling temperature changes. A higher value indicates a more significant core role for temperature control. The heat transfer contribution is adjusted by combining the node's niche width with the Pearson correlation coefficient between the driving node and the core temperature parameters of the corresponding rolling stage. The formula is: Heat Transfer Contribution = Niche Width × Absolute Value of Correlation Coefficient. Here, niche width reflects the range of influence of the driving node in the rolling temperature regulation resource space, while the absolute value of the correlation coefficient quantifies the degree of linear correlation between the node and temperature changes. Multiplying the two together comprehensively reflects the node's actual driving potential for temperature changes.

[0116] Heat transfer reliability is used to quantify the ability of a driving node to function stably during heat transfer. A higher value indicates stronger node resistance to interference and a lower likelihood of driving failure. The correlation failure threshold method is used to calculate this reliability by analyzing the distribution of strongly correlated edges of the driving node in the parametric gravity network. The formula is: Heat transfer reliability = Number of strongly correlated edges that need to be broken to cause node failure ÷ (Total number of nodes in the current network community - 1). Here, the number of strongly correlated edges that need to be broken to cause node failure refers to the number of edges that need to be severed to disconnect the node from all other strongly correlated nodes, preventing it from transmitting its control function. The total number of nodes in the current network community - 1 represents the maximum potential number of connections between the node and other nodes within the community. The ratio of these two values ​​directly reflects the stability of the node's driving function.

[0117] The heat transfer response node reflects the gap between the heat transfer effect and the process target. Therefore, quantifying the temperature deviation requirement is the core dimension, forming a single key heat transfer characteristic parameter. The temperature deviation requirement quantifies the difference between the actual rolling temperature and the process target temperature corresponding to the response node. The larger the value, the greater the deviation between the current heat transfer effect and the target. The formula is: Temperature Deviation Requirement = |Actual Acquired Temperature - Process Target Temperature|. Where, the actual acquired temperature is the time-series data of the response node in the time-aligned data matrix; the process target temperature is the determined temperature control parameter for each rolling stage. For example, in the finishing rolling stage, the process target value for the finishing mill exit temperature is 880℃, and the actual acquired temperature at a certain time point is 872℃, then the temperature deviation requirement for this node at that time is 8℃.

[0118] In addition, the quantification parameters of the heat transfer drive node and the heat transfer response node should adopt the same standardization method, such as min-max standardization, to ensure that the two types of parameters can be directly calculated and compared in the subsequent heat transfer effect calculation, so as to obtain the standardized heat transfer contribution, heat transfer reliability and temperature deviation requirements.

[0119] S2222. A connectivity test algorithm is used to analyze the connectivity of network nodes in the network community, and network nodes that meet the connectivity threshold and contain heat transfer driving nodes and heat transfer response nodes are selected to obtain a feasible subset of heat transfer nodes.

[0120] It should be noted that the connectivity analysis of network nodes in the network community employs graph theory connectivity evaluation algorithms, such as depth-first search and breadth-first search, traversing the nodes within the network community and the strongly connected edges between them. This is used to determine whether nodes are interconnected through strongly connected edges; that is, any two nodes must have at least one path consisting of strongly connected edges, ensuring that the control effect of the driving node can be transmitted to the responding node. The connectivity analysis is based on the edge weights in the parametric gravity network, i.e., the gravity strength, considering only strongly connected edges. In other words, the gravity strength is greater than or equal to a preset basic threshold; weakly connected edges, due to their weak transmission effect, are not included in the connectivity determination. Based on connectivity analysis, two core conditions must be met simultaneously to determine a feasible subset for heat transfer: First, connectivity must be met, meaning the proportion of strongly correlated edges between nodes in the subset is greater than or equal to a preset connectivity threshold. This threshold is usually calibrated based on historical valid heat transfer data, such as 0.8, which means that more than 80% of the potential correlations in the subset are strongly correlated, ensuring close connections. Second, the subset must contain both heat transfer driving nodes and heat transfer response nodes. The absence of either type of node will prevent the formation of a complete heat transfer causal chain. Driving nodes alone cannot reflect the heat transfer results, and response nodes alone cannot trace the cause of heat transfer.

[0121] S2223. Calculate the heat transfer utility of each feasible subset of heat transfer based on the preset utility function;

[0122] It should be noted that the pre-defined utility function construction needs to consider the quantification capability of the driving nodes and the deviation requirements of the response nodes. A typical function form is: Heat transfer utility = Total effective control capability of driving nodes ÷ Total temperature deviation requirements of response nodes. Here, the effective control capability of the driving nodes is calculated by multiplying the heat transfer contribution and heat transfer reliability, i.e., Effective control capability = Heat transfer contribution × Heat transfer reliability. This product comprehensively reflects the potential control capability and stable performance of the driving nodes. The temperature deviation requirements of the response nodes are directly adopted using the obtained standardized temperature deviation requirements, and the sum is the accumulation of the deviation values ​​of all response nodes in the subset. Based on the above function, the heat transfer utility value is calculated by substituting the quantification parameters of each feasible heat transfer subset.

[0123] S2224. Calculate the total heat transfer utility of the current network community based on the heat transfer utility of each feasible subset of heat transfer.

[0124] Specifically, by integrating the utility values ​​of all feasible subsets of heat transfer, the total heat transfer utility of the current network community is obtained, which corresponds to the total heat transfer utility of a certain rolling stage.

[0125] S2225. Based on the total heat transfer efficiency of the current network community, calculate the heat transfer contribution centrality of each network node based on the Shapley value, and determine the heat transfer mechanism of the rolling stage based on the heat transfer contribution centrality ranking results.

[0126] It should be noted that by calculating the heat transfer contribution centrality of each node using the Shapley value, the average marginal contribution of each node to the total heat transfer efficiency of the stage can be quantified. Combined with contribution ranking, core heat transfer parameters and dominant paths can be identified, ultimately leading to the extraction of a heat transfer mechanism that is both data-supported and physically logical. The specific implementation steps include the following:

[0127] Viewing the network community as a heat transfer cooperative game system, nodes are game participants, and total heat transfer utility is the total game payoff. The Shapley value is obtained by calculating the utility increment (marginal contribution) of each node after joining any subset of nodes, and then taking a weighted average of all increments to obtain the heat transfer contribution centrality of each node. The weighting coefficients are the probabilities of subsets appearing, ensuring computational fairness and preventing any subset from excessively influencing the results. In actual calculations, the traversal range can be simplified to only consider feasible subsets and their actual subsets; in other words, infeasible subsets have zero utility and no impact on marginal contribution. Simultaneously, it is necessary to verify whether the sum of the centralities of all nodes equals the total heat transfer utility to ensure calculation accuracy.

[0128] Nodes are sorted based on their heat transfer contribution centrality. Then, based on this centrality ranking, the top three core nodes are extracted from highest to lowest value. The first two are heat transfer driving nodes, representing core control parameters, and the first is a heat transfer response node, representing a core feedback parameter. Strongly correlated edges between core driving and core response nodes are identified; in other words, the gravitational pull strength ≥ 0.7. The parameter transfer relationship corresponding to this edge is the dominant heat transfer path for that stage. Combining the core parameter functions, dominant paths, and the physical laws of the rolling process, the heat transfer mechanism is extracted. For example, in the heating stage, the core driving nodes are furnace power and gas flow rate, the core response node is the initial rolling temperature, and the dominant path is furnace power → initial rolling temperature. The mechanism is that the conductive and radiative heat generated by fuel combustion dominates the billet temperature rise, and furnace power is the core means of controlling the heating rate.

[0129] S223. By combining the dynamic evolution characteristics of key state parameters and niche indicators in the network community, identify the interference factors in each rolling stage.

[0130] As a preferred embodiment, the step of identifying interference factors in each rolling stage by combining the dynamic evolution characteristics of key state parameters in the network community with niche indicators includes the following steps:

[0131] S2231. For key state parameters in the network community, calculate the mean, variance and mutation rate of each key state parameter respectively.

[0132] It should be noted that the mean is the average value of the key state parameter across all sampling time points within the current rolling stage of the network community; the variance is used to quantify the overall fluctuation of the key state parameter within this rolling stage, reflecting the stability level of the parameter operation. The mutation rate is used to capture nodes of drastic fluctuations in the parameter over time, quantifying the frequency of abnormal mutations in the parameter.

[0133] S2232. Based on the mean, variance and mutation rate, combined with the historical normal fluctuation range and niche width of key state parameters, identify and mark key state parameters that show abnormal fluctuations, and preliminarily determine them as potential disturbance factors.

[0134] Specifically, the historical normal fluctuation range serves as the normal operating range of key state parameters in the corresponding rolling stage, obtained from statistical analysis of a large amount of historical production data. It is typically taken as the mean ± 3 standard deviations, used to determine whether the current parameter's mean and fluctuation deviate from the norm. If the current parameter's mean exceeds this range, or the fluctuation range corresponding to the variance covers values ​​outside the range, it indicates an abnormal parameter operating state and potential interference. Niche width reflects the range of influence of key state parameters on the rolling temperature regulation resource space; the larger the width, the stronger the parameter's potential influence on the heat transfer process. In summary, if a key state parameter simultaneously meets the criteria of abnormal fluctuation—in other words, a mean exceeding the normal range, excessive variance, or excessively high mutation rate—and the niche width meets the standard, it can be marked as a potential interference factor. For example, if the average gas flow rate in the heating stage exceeds the historical normal range, the variance increases by 50% compared to the historical mean, and the niche width is 0.65, it can be preliminarily identified as a potential interference factor.

[0135] S2233. Analyze the niche overlap of potential interference sources within the network community, identify other key state parameters that cause changes in niche overlap due to their abnormal fluctuations, in order to determine the set of parameters affected by interference and the propagation path of interference.

[0136] It should be noted that by comparing the changes in the niche overlap between potential interference factors and other key state parameters before and after abnormal fluctuations, if the overlap between a parameter and the potential interference factor changes significantly after the fluctuation (e.g., an increase or decrease of ≥30%), it indicates that the correlation between the two has been disrupted by the interference factor, and this parameter is the parameter affected by the interference. For example, after abnormal fluctuations in the cooling water volume, a potential interference factor, the niche overlap between it and the finishing mill exit temperature decreased from 0.72 to 0.41, indicating that the finishing mill exit temperature was affected by the cooling water volume, and the correlation weakened. Based on the changes in overlap, the path of interference propagation is traced. Starting from the potential interference factor, the affected parameters are sequentially tracked along the edges where the overlap changes significantly, forming the interference propagation path.

[0137] S2234. By comprehensively considering potential interference factors, the set of parameters affected by interference, and the propagation path of interference, assess the intensity and scope of each interference factor in order to identify interference factors in the current rolling stage.

[0138] It should be noted that, based on the interference-related information obtained above, the impact intensity and scope of each potential interference factor are quantitatively assessed to ultimately determine the core interference factor in the current rolling stage, completing the interference identification loop. When assessing interference intensity, three core dimensions are used to quantify it: the degree of anomaly of the potential interference factor (i.e., the magnitude of the deviation of the mean from the normal range, the magnitude of the mutation rate, the number of parameters affected by the interference, and the degree of influence of the interference on the core heat transfer parameters). An interference intensity score can be calculated using a weighted scoring method; the higher the score, the greater the interference intensity. The assessment of the interference impact range is based on the interference propagation path and is defined by the distribution range of the affected parameters. If the affected parameters are limited to a specific local heat transfer stage, such as affecting only the heating stage, the impact range is small; if they cover multiple heat transfer stages, such as simultaneously affecting the heating and finishing rolling stages, the impact range is large. Furthermore, considering the niche width corresponding to the affected parameters, if the affected parameters are mostly core parameters with large niche widths, the importance of the impact range is higher.

[0139] Based on the comprehensive evaluation results of interference intensity and impact range, the top-scoring potential interference factors were selected as the final interference factors for the current rolling stage. Simultaneously, the interference characteristics of each interference factor were determined, including interference type, propagation path, and core affected parameters, providing a precise basis for setting interference correction terms when subsequently constructing the dynamic thermodynamic model.

[0140] S23. Based on the heat transfer mechanism and interference factors of each rolling stage, construct a dynamic thermodynamic model for each rolling stage.

[0141] Specifically, the dominant heat transfer mechanisms in different rolling stages, such as radiation and conduction in the heating stage and convection and frictional heat generation in the finishing stage, as well as core parameters such as furnace power and cooling water volume, differ significantly. Therefore, it is necessary to design a dedicated model framework for each stage and determine the model's inputs, outputs, and core variables: the model inputs are the core heat transfer driving node parameters and identified interference factor parameters corresponding to the stage; the model outputs are the core heat transfer response node parameters corresponding to the stage, i.e., the key temperature values ​​that the model needs to predict; the core variables include thermodynamic core variables such as heat flux density, heat transfer coefficient, and billet temperature field distribution, and the selection of variables must be consistent with the dominant heat transfer mechanism of the stage.

[0142] Once the basic model framework is established, based on the determined staged heat transfer mechanism, the core driving parameters, the causal relationship of temperature changes, and the dominant heat transfer path need to be transformed into thermodynamic equations and embedded into the model framework to form the core prediction part of the model. Specifically, this includes the following steps:

[0143] Based on the dominant heat transfer type with a clearly defined stage heat transfer mechanism, the corresponding classical heat transfer equations are selected as the foundation. For example, in the heating stage, which is dominated by radiation and conduction, the radiation heat transfer equation (Stephen Boltzmann law) is used to describe the radiation heating of the steel billet by the furnace, and Fourier's law is combined to describe the conduction heat transfer inside the steel billet, integrating them into a quantitative equation for the change of steel billet temperature with heating time and furnace power; in the finishing rolling stage, which is dominated by convection and frictional heat generation, Newton's cooling formula is used to describe the convective heat dissipation of the cooling water flow, and the frictional heat generation formula is combined to describe the heat generation during the rolling process, integrating them into a quantitative equation for the change of finishing mill exit temperature with cooling water flow and rolling speed.

[0144] By using the obtained heat transfer contribution centrality, weights are assigned to the core driving parameters to reflect the degree of influence of different parameters on temperature changes.

[0145] After the core predictive component of the model is built, it needs to be calibrated using historical data and verified through on-site trial rolling to optimize model parameters, such as heat transfer coefficient, weighting coefficient, and correction function coefficient. This ensures that the deviation between the model output and actual temperature data is within acceptable limits, typically ≤ ±5℃, meeting the requirements of cold heading steel rolling processes. Specifically, this includes:

[0146] Historical production data for this rolling stage, i.e., the time-aligned data matrix, is selected. The core driving parameters and interference factor parameters are input into the model. The model coefficients are optimized by the least squares method to minimize the mean square error between the model's predicted temperature and the historical actual temperature.

[0147] Small-batch trial rolling is carried out on the actual production line. Drive parameters, disturbance parameters and actual temperature data are collected in real time. The deviation between the model-predicted temperature and the actual temperature is compared. If the deviation exceeds the allowable range, the heat transfer equation coefficients or disturbance correction term parameters of the model are adjusted back until the accuracy requirements are met.

[0148] S3. Obtain the real-time status data of the rolling mill and the real-time temperature data of the rolling area of ​​the rolling mill, and preprocess the real-time status data and real-time temperature data to obtain the preprocessed real-time status data and real-time temperature data.

[0149] Specifically, real-time status data of the rolling mill is collected in real time by various sensors on the rolling mill equipment, including furnace power, cooling water volume, rolling speed, and gas flow. The data acquisition frequency needs to match the rolling rhythm, typically 10-20Hz, to ensure the capture of dynamic changes in parameters. Real-time temperature data of the rolling zone is collected by temperature sensors deployed at key locations, such as thermocouples and infrared thermometers, corresponding to core response parameters at each stage, such as the initial rolling temperature, roughing mill exit temperature, and finishing mill exit temperature. The acquisition frequency needs to be synchronized with the status data to ensure time alignment. The collected data is transmitted to the control center in real time via industrial Ethernet or a PLC system, forming a real-time data stream.

[0150] Data preprocessing includes data cleaning, dimensional standardization, time alignment, and so on.

[0151] S4. Based on the preprocessed real-time status data, determine the rolling stage of cold heading steel wire rod and extract the temperature control parameters of the rolling stage. Compare the extracted temperature control parameters with the preprocessed real-time temperature data to obtain the real-time temperature deviation.

[0152] It should be noted that the current rolling stage of the cold heading steel wire rod is determined based on the preprocessed real-time status data and the typical parameter characteristics of each rolling stage. The core status parameters of different rolling stages exhibit significant characteristics. For example, in the heating stage, the furnace power remains consistently high and the rolling speed is close to zero; in the finishing stage, the cooling water volume fluctuates frequently and the rolling speed remains in the high-speed range. Therefore, stage determination can be achieved through preset parameter feature thresholds or machine learning classification models. For instance, when the furnace power is ≥80% of the rated power and the rolling speed is ≤0.5m / s in the real-time data, it is determined to be in the heating stage; when the cooling water volume is ≥50m / s, it is determined to be in the finishing stage. 3 When the rolling speed is ≥10m / s, it is determined to be the finishing rolling stage.

[0153] After determining the current rolling stage, the target temperature value and allowable fluctuation range for the corresponding stage are extracted from the preset temperature control parameters for each stage. For example, the target temperature for the initial rolling stage is 980-1020℃, and the target temperature for the exit stage is 850-900℃. The difference between the preprocessed real-time temperature data and the extracted temperature control target value is then calculated to obtain the real-time temperature deviation.

[0154] S5. Based on the dynamic thermodynamic model and real-time temperature deviation, the control quantity is calculated through the PID algorithm, and the control command is output to the actuator of the rolling mill.

[0155] Specifically, the preprocessed real-time status data is input into the dynamic thermodynamic model corresponding to the rolling stage. The model combines the heat transfer mechanism of the current stage with the identified disturbance factors to predict the temperature change trend in subsequent time periods, such as predicting whether the temperature deviation will increase or decrease within the next 5 sampling periods. The influence of the existing disturbance factors on the temperature is quantified, providing a basis for PID control compensation.

[0156] Among them, the PID algorithm (Proportional-Integral-Derivative) is a classic closed-loop control algorithm in industrial control. Its core is to calculate the optimal control quantity through the synergistic effect of the proportional (P), integral (I), and derivative (D) terms. In this scheme, the input to the PID algorithm not only includes the real-time temperature deviation but also incorporates the prediction deviation and disturbance compensation of the dynamic thermodynamic model, making the control more targeted.

[0157] The proportional term (P) is a control quantity that is directly proportional to the real-time temperature deviation, which is based on the magnitude of the deviation and quickly suppresses the current deviation. For example, the larger the deviation, the larger the adjustment range.

[0158] The integral term (I) accumulates historical deviations, eliminates static errors such as long-standing small deviations, and ensures that the temperature eventually stabilizes at the target value.

[0159] The differential term (D) is based on the rate of change of temperature deviation, predicts the direction of deviation change in advance, and outputs a control quantity to suppress deviation change and avoid overshoot. For example, if the deviation is predicted to increase rapidly, the adjustment intensity is increased in advance.

[0160] Specifically, the control quantity calculated by the PID algorithm needs to be converted into control commands that the rolling mill actuators can recognize, such as current signals or pulse signals. The content of the control command corresponds to the adjustment amount of the core drive parameters. For example, if the real-time temperature is too high or the deviation is positive, and the model determines that it is caused by insufficient cooling water, then the control command is to increase the cooling water volume by 5%. The control command is transmitted in real time to the corresponding actuators, such as flow valves, power regulators, and frequency converters, through the PLC system.

[0161] S6. The rolling mill's actuator receives instructions and performs temperature adjustment actions to achieve closed-loop stable control of the rolling temperature of cold heading steel wire rod.

[0162] It should be noted that the mill's actuators precisely execute corresponding adjustment actions based on the output control commands. Different control commands correspond to different actuators: for example, adjusting the heating furnace power corresponds to the action of the power regulator, adjusting the cooling water volume corresponds to the action of the flow control valve, and adjusting the rolling speed corresponds to the change in the speed of the frequency converter drive motor. The accuracy of the actuator's action must match the process requirements, such as the flow valve adjustment accuracy ≤1% and the power adjustment accuracy ≤0.5%, to ensure accurate transmission of the adjustment amount and avoid temperature fluctuations caused by execution deviations. After the actuator completes the adjustment, the mill's status sensors will collect the adjusted status parameters in real time, and the temperature sensors will collect the adjusted rolling temperature. After these data are preprocessed, stage judgment and deviation calculation are performed again. If the new temperature deviation still exceeds the allowable range, the calculation process of S5-S6 will be repeated until the temperature deviation is reduced to the allowable range; if the deviation has reached the standard, the current control parameters are maintained, and the mill enters a stable monitoring state.

[0163] In addition, this invention selected YSCM435 wire rod of specification 11 for performance testing before process optimization. As shown in Table 2, the metallographic structure of the wire rod produced before process optimization was mainly bainite with a small amount of ferrite. The overall performance fluctuated greatly, and the toughness and plasticity were poor. Specifically, as shown in Table 2... Figure 5-7 As shown.

[0164] Table 2 Performance test data before process optimization

[0165]

[0166]

[0167] As shown in Table 3, based on the temperature closed-loop control method of this invention, 11-specification YSCM435 wire rods were selected for performance testing after process optimization; Figure 8-10 As shown, the mechanical properties of the wire rod produced before process optimization were significantly reduced to within 800 MPa, the wire rod matrix was mainly ferrite, the area shrinkage performance was excellent, and the overall performance was significantly improved.

[0168] Table 3 Performance test data after process optimization

[0169]

[0170] like Figure 4 As shown, according to a second embodiment of the present invention, a PID-based closed-loop control system for cold heading steel wire rod rolling temperature is provided, the system comprising:

[0171] Temperature control parameter determination module 1 is used to determine the temperature control parameters of the rolling zone of the rolling mill at different rolling stages of cold heading steel wire rod according to the process requirements of cold heading steel wire rod;

[0172] Model building module 2 is used to construct a parametric gravity network based on parametric gravity analysis, combined with pre-acquired historical state data of the rolling mill and historical temperature data of the rolling area of ​​the rolling mill. According to different rolling stages, it performs segmented analysis on the heat transfer mechanism and interference factors of the rolling process. Based on the segmented analysis results, it constructs dynamic thermodynamic models for different rolling stages to achieve temperature adjustment at different rolling stages.

[0173] The real-time data acquisition and processing module 3 is used to acquire the real-time status data of the rolling mill and the real-time temperature data of the rolling area of ​​the rolling mill, and to preprocess the real-time status data and real-time temperature data to obtain the preprocessed real-time status data and real-time temperature data.

[0174] The real-time temperature deviation calculation module 4 is used to determine the rolling stage of cold heading steel wire rod based on the pre-processed real-time status data and extract the temperature control parameters of the rolling stage. The extracted temperature control parameters are compared with the pre-processed real-time temperature data to obtain the real-time temperature deviation.

[0175] The control quantity calculation module 5 is used to calculate the control quantity based on the dynamic thermodynamic model and real-time temperature deviation using a PID algorithm, and output the control command to the actuator of the rolling mill.

[0176] The temperature dynamic adjustment module 6 is used by the rolling mill's actuator to receive instructions and perform temperature adjustment actions to achieve closed-loop stable control of the rolling temperature of cold heading steel wire rod.

[0177] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A closed-loop control method for temperature control in cold heading steel wire rod rolling based on PID control, characterized in that, Includes the following steps: S1. Based on the process requirements of cold heading steel wire rod, determine the temperature control parameters of the rolling zone of the rolling mill for different rolling stages of cold heading steel wire rod; S2. Based on the parametric gravity analysis method, combined with the historical state data of the rolling mill and the historical temperature data of the rolling area of ​​the rolling mill, a parametric gravity network is constructed. According to different rolling stages, the heat transfer mechanism and interference factors of the rolling process are analyzed in segments. Based on the segmented analysis results, dynamic thermodynamic models of different rolling stages are constructed to achieve temperature adjustment in different rolling stages. S3. Obtain the real-time status data of the rolling mill and the real-time temperature data of the rolling area of ​​the rolling mill, and preprocess the real-time status data and real-time temperature data to obtain the preprocessed real-time status data and real-time temperature data. S4. Based on the preprocessed real-time status data, determine the rolling stage of cold heading steel wire rod and extract the temperature control parameters of the rolling stage. Compare the extracted temperature control parameters with the preprocessed real-time temperature data to obtain the real-time temperature deviation. S5. Based on the dynamic thermodynamic model and real-time temperature deviation, the control quantity is calculated through the PID algorithm, and the control command is output to the actuator of the rolling mill. S6. The rolling mill's actuator receives instructions and performs temperature adjustment actions to achieve closed-loop stable control of the rolling temperature of cold heading steel wire rod. The parametric gravity analysis method, combined with pre-acquired historical state data of the rolling mill and historical temperature data of the rolling zone, constructs a parametric gravity network. Based on different rolling stages, it performs segmented analysis of the heat transfer mechanism and interference factors in the rolling process. Based on the segmented analysis results, it constructs dynamic thermodynamic models for different rolling stages, thereby achieving the following steps for different rolling stages: S21. Perform timestamp alignment processing on the historical state data of the rolling mill and the historical temperature data of the rolling area of ​​the rolling mill, and construct a parametric gravity network based on the alignment results; S22. Based on different rolling stages, the parametric gravitational network is segmented to obtain several network communities, and the heat transfer mechanism and interference factors of the network community corresponding to each rolling stage are analyzed. S23. Based on the heat transfer mechanism and interference factors of each rolling stage, construct a dynamic thermodynamic model for each rolling stage.

2. The closed-loop control method for temperature control in cold heading steel wire rod rolling based on PID according to claim 1, characterized in that, The process of aligning the historical state data of the rolling mill and the historical temperature data of the rolling area of ​​the rolling mill with timestamps, and constructing a parametric gravity network based on the alignment results and in conjunction with niche theory, includes the following steps: S211. Using a time synchronization algorithm, the historical state data of the rolling mill and the historical temperature data of the rolling area of ​​the rolling mill are timestamped and a time-aligned data matrix containing temperature sequence and state parameter sequence is constructed. S212. Based on the time-aligned data matrix, calculate the niche width of each state parameter, and filter out key state parameters that meet the preset width threshold to obtain a set of key state parameters. S213. Based on the set of key state parameters, calculate the niche overlap between any two sets of state parameters, and select strongly correlated parameters that meet the preset overlap threshold to obtain a set of strongly correlated parameters. S214. Based on the set of strongly correlated parameters and the niche width of each state parameter, construct a parameter gravity network with time points as nodes and the gravitational strength between parameters as edge weights.

3. The closed-loop control method for temperature control in cold heading steel wire rod rolling based on PID according to claim 2, characterized in that, The construction of a parametric gravity network based on a set of strongly correlated parameters and the niche width of each state parameter, with time points as nodes and the gravitational strength between parameters as edge weights, includes the following steps: S2141. Take each sampling time point as a network node. For any two network nodes, calculate the comprehensive state characterization value of any two network nodes based on the niche width of the state parameters. S2142. Based on the comprehensive state characterization values ​​of any two network nodes, and combined with the weighted average of the niche overlap between their corresponding strongly correlated parameter pairs, calculate the basic gravitational strength between the two network nodes. S2143. Based on the preset time-distance decay factor and the time interval between two network nodes, calculate the final gravitational strength. S2144. Based on each network node, construct a parametric gravity network using the calculated gravitational strength as the weight of the directed edge.

4. The closed-loop control method for temperature control in cold heading steel wire rod rolling based on PID according to claim 1, characterized in that, The process of segmenting the parametric gravitational network based on different rolling stages to obtain several network communities, and analyzing the heat transfer mechanism and interference factors of the network community corresponding to each rolling stage, includes the following steps: S221. Determine the time interval for each rolling stage, and divide the parametric gravity network according to the time interval to obtain several network communities. S222. Classify the network nodes in each network community and determine the heat transfer mechanism of each rolling stage through centrality analysis based on Shapley value. S223. By combining the dynamic evolution characteristics of key state parameters and niche indicators in the network community, identify the interference factors in each rolling stage.

5. The closed-loop control method for temperature control in cold heading steel wire rod rolling based on PID according to claim 4, characterized in that, The process of classifying network nodes in each network community and determining the heat transfer mechanism for each rolling stage through Shapley value-based centrality analysis includes the following steps: S2221. Divide the network nodes in each network community into heat transfer drive nodes and heat transfer response nodes. S2222. A connectivity test algorithm is used to analyze the connectivity of network nodes in the network community, and network nodes that meet the connectivity threshold and contain heat transfer driving nodes and heat transfer response nodes are selected to obtain a feasible subset of heat transfer nodes. S2223. Calculate the heat transfer utility of each feasible subset of heat transfer based on the preset utility function; S2224. Calculate the total heat transfer utility of the current network community based on the heat transfer utility of each feasible subset of heat transfer. S2225. Based on the total heat transfer efficiency of the current network community, calculate the heat transfer contribution centrality of each network node based on the Shapley value, and determine the heat transfer mechanism of the rolling stage based on the heat transfer contribution centrality ranking results.

6. The closed-loop control method for temperature control in cold heading steel wire rod rolling based on PID according to claim 5, characterized in that, The method of identifying interference factors in each rolling stage by combining the dynamic evolution characteristics of key state parameters in the network community with niche indicators includes the following steps: S2231. For key state parameters in the network community, calculate the mean, variance and mutation rate of each key state parameter respectively. S2232. Based on the mean, variance and mutation rate, combined with the historical normal fluctuation range and niche width of key state parameters, identify and mark key state parameters that show abnormal fluctuations, and preliminarily determine them as potential disturbance factors. S2233. Analyze the niche overlap of potential interference sources within the network community, identify other key state parameters that cause changes in niche overlap due to their abnormal fluctuations, in order to determine the set of parameters affected by interference and the propagation path of interference. S2234. By comprehensively considering potential interference factors, the set of parameters affected by interference, and the propagation path of interference, assess the intensity and scope of each interference factor in order to identify interference factors in the current rolling stage.

7. A PID-based closed-loop control system for cold heading steel wire rod rolling temperature, used to implement the PID-based closed-loop control method for cold heading steel wire rod rolling temperature as described in any one of claims 1-6, characterized in that, The system includes: The temperature control parameter determination module is used to determine the temperature control parameters of the rolling zone of the rolling mill at different rolling stages of cold heading steel wire rod according to the process requirements of cold heading steel wire rod. The model building module is used to construct a parametric gravity network based on parametric gravity analysis, combined with pre-acquired historical state data of the rolling mill and historical temperature data of the rolling area of ​​the rolling mill. According to different rolling stages, it performs segmented analysis on the heat transfer mechanism and interference factors of the rolling process. Based on the segmented analysis results, it constructs dynamic thermodynamic models for different rolling stages to achieve temperature adjustment at different rolling stages. The real-time data acquisition and processing module is used to acquire the real-time status data of the rolling mill and the real-time temperature data of the rolling area of ​​the rolling mill, and to preprocess the real-time status data and real-time temperature data to obtain the preprocessed real-time status data and real-time temperature data. The real-time temperature deviation calculation module is used to determine the rolling stage of cold heading steel wire rod based on the pre-processed real-time status data and extract the temperature control parameters of the rolling stage. The extracted temperature control parameters are compared with the pre-processed real-time temperature data to obtain the real-time temperature deviation. The control quantity calculation module is used to calculate the control quantity based on the dynamic thermodynamic model and real-time temperature deviation using the PID algorithm, and output the control command to the actuator of the rolling mill. The temperature dynamic adjustment module is used by the rolling mill's actuator to receive instructions and perform temperature adjustment actions to achieve closed-loop stable control of the rolling temperature of cold heading steel wire rod.

8. An electronic device, characterized in that, The electronic device includes: one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to perform the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein the steps of the method according to any one of claims 1 to 6 are implemented when the computer program controls the device where the computer-readable storage medium is located to execute during runtime.

Images