Temperature control system for high-grade oriented silicon steel

Abstract

The invention discloses a temperature control system for high-grade oriented silicon steel, relates to the technical field of automatic control, and solves the problem that the strip temperature is difficult to analyze according to a temperature model. Secondly, the surface heat transfer coefficient is difficult to analyze; thirdly, the final rolled piece temperature is difficult to comprehensively analyze, and the flow of cooling emulsion is automatically adjusted to control the strip temperature; and operation abnormity in the video image is difficult to analyze and real-time early warning is difficult to carry out. The system comprises the following modules: a data acquisition module, a data processing module, a data analysis module, a comprehensive analysis module and a real-time early warning module. And through temperature control in the cold rolling process, various temperature models are adopted for automatically controlling the temperature of the strip. Through feedforward control and feedback control over the temperature, accurate control over the temperature of the strip is achieved, and therefore qualified HiB steel is produced. In addition, various operation abnormalities are analyzed by monitoring video images in real time, and real-time early warning is carried out.

Description

Technical Field

[0001] The invention belongs to the technical field of automatic control, in particular to a temperature control system for high-grade oriented silicon steel. Background Art

[0002] HiB (High Induction Grain Oriented Electrical Steel) is high magnetic induction oriented silicon steel, which refers to oriented silicon steel with magnetic polarization intensity J≥1.88T. Traditional HiB rolling generally uses manual flow adjustment to control temperature. Generally, according to the different raw material temperatures, there is a temperature target value for each pass, and the outlet strip temperature is manually controlled to reach this target value. When just starting, the speed is very low, and the required flow rate is very low. As the speed increases, the flow rate needs to be increased. As the speed changes, the flow rate needs to be adjusted. Finally, the speed is reduced, and the flow rate also needs to be reduced. Manual control has a large lag and instability, and there will be a long period of time when the target temperature cannot be reached.

[0003] The existing temperature control system for high-grade oriented silicon steel has the following problems: first, it is difficult to analyze the strip temperature based on the temperature model; second, it is difficult to analyze the surface heat transfer coefficient; then, it is difficult to comprehensively analyze the final rolled piece temperature and automatically adjust the cooling emulsion flow rate to control the strip temperature; finally, it is difficult to analyze operational abnormalities in video images and provide real-time warnings. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a temperature control system for high-grade oriented silicon steel, which is used to solve the following technical problems: First, it is difficult to analyze the strip temperature based on the temperature model. Second, it is difficult to analyze the surface heat transfer coefficient. Third, it is difficult to comprehensively analyze the final rolled piece temperature and automatically adjust the cooling emulsion flow rate to control the strip temperature. Finally, it is difficult to analyze operational anomalies in video images and issue real-time warnings.

[0005] To solve the above problems, the first aspect of the present invention provides a temperature control system for high-grade oriented silicon steel, comprising the following modules: Data acquisition module: real-time acquisition and measurement of strip temperature and video image data at the inlet and outlet sides of the rolling mill; Data processing module: pre-processing the collected strip temperature, including filtering, normalization, removal of outliers and erroneous data; pre-processing the collected video image data, including denoising and enhancement; Data analysis module: Real-time calculation of the strip temperature through the temperature model; analysis of the surface heat transfer coefficient based on the neural network model; Comprehensive analysis module: The final rolled piece temperature is calculated based on the temperature of the strip calculated by the temperature model; the flow rate of the cooling emulsion is automatically adjusted based on the calculation results to control the temperature of the strip; Real-time warning module: Analyzes operation abnormalities based on the collected and processed video images and calculates the operation abnormality coefficient based on the analysis results to issue real-time warnings.

[0006] As a further solution of the present invention: the temperature of the strip is calculated in real time by a temperature model, comprising the following steps: The temperature model includes: a deformation heat model, a friction heat transfer model, a convection heat transfer model and a heat conduction model; The real-time calculation of the temperature of the strip includes: calculating the thermal deformation temperature rise according to the deformation heat model; calculating the friction temperature rise according to the friction heat transfer model; calculating the temperature of the strip after cooling according to the convection heat transfer model; and calculating the heat conduction temperature drop according to the heat conduction model.

[0007] As a further solution of the present invention: calculating the thermal deformation temperature rise according to the deformation thermal model includes the following steps: By analyzing the formula: The rolling force per unit area is obtained ; By analyzing the formula: Get thermal deformation temperature rise , represents the deformation heat, represents density, represents specific heat capacity, H represents inlet thickness, h represents outlet thickness, Indicates efficiency, Indicates the rolling force per unit area.

[0008] As a further solution of the present invention: calculating the friction temperature rise according to the friction heat transfer model includes the following steps: By analyzing the formula: Friction temperature rise ;in, represents friction heat, represents density, H represents inlet thickness, h represents outlet thickness, represents the average linear velocity, Indicates efficiency, is the rolling force per unit area, Represents the contact time constant.

[0009] As a further solution of the present invention: calculating the temperature of the strip after cooling according to the convection heat transfer model includes the following steps: By Newton's cooling formula: Where: represents the heat flux density, Indicates the strip temperature, Indicates the coolant temperature, represents the surface heat transfer coefficient; By differentiating the convection equation analytically: By analyzing the formula: The solution to the initial value problem is: Get the strip temperature ;in, Indicates the strip temperature before cooling, represents the time constant, Indicates the cooldown time.

[0010] As a further solution of the present invention: analyzing the surface heat transfer coefficient according to the neural network model includes the following steps: The surface heat transfer coefficient is analyzed by using a neural network model; The surface heat transfer coefficients at different flow rates and temperatures are used as training data for the neural network model; The surface heat transfer coefficient is input as the target variable into the neural network model for training; The temperature and flow data collected and processed in real time by the temperature control system of high-grade oriented silicon steel are input into the trained neural network model in real time to predict the surface heat transfer coefficient; According to the prediction results of the neural network model, the parameters are adjusted to make the surface heat transfer coefficient close to the actual value.

[0011] As a further solution of the present invention: calculating the heat conduction temperature drop according to the heat conduction model includes the following steps: By analyzing the formula: Get the heat conduction temperature drop , represents the heat flow, represents density, represents the specific heat capacity, represents the thermal conductivity of the rolled piece, is the contact time constant, represents the equivalent roll temperature, Indicates the initial temperature of the strip.

[0012] As a further solution of the present invention: comprehensively calculating the final rolled piece temperature according to the temperature of the strip calculated by the temperature model comprises the following steps: By analyzing the formula: Get the final rolling temperature ,in, express Weight of thermal deformation temperature rise, express The weight of the friction temperature rise, express The weight of the strip temperature after cooling, express Weight of the temperature drop due to heat conduction.

[0013] As a further solution of the present invention, the flow rate of the cooling emulsion is automatically adjusted according to the calculation results to control the temperature of the strip, including the following steps: S1: According to the calculated final rolled product temperature, by changing the cooling flow rate of each part, the cooling flow rate is calculated using Newton's method, and the temperature of the outlet strip is controlled to reach the target value; S2: By changing the speed of the steel strip, the flow rate of the steel strip at different speeds is obtained, and the flow rate of the steel strip is pre-controlled using the speed-flow curve; S3: By changing the inlet temperature, different outlet temperatures and outlet temperature / inlet temperature influence coefficients are obtained, and feedforward control of the temperature is performed through the coefficients under different conditions; S4: Temperature feedback control is performed using the obtained flow rate and outlet temperature sensitivity coefficient.

[0014] As a further solution of the present invention, the operation abnormality is analyzed according to the collected and processed video images, and the operation abnormality coefficient is calculated according to the analysis result to perform real-time warning, including the following steps: Based on the pre-processed video image data, the operation anomalies in the video image are analyzed through the deep learning model; Six types of operation abnormalities are marked, including: abnormal vibration of the rolling mill rolls, deviation of the rolled pieces, blockage of the rolling mill rolls, uneven force on the rolling mill rolls, low rolling speed, and overheating of the rolling mill; The six annotated operational anomalies were input into the convolutional neural network of the deep learning model for model training; The temperature control system of high-grade oriented silicon steel collects processed video images in real time to identify six types of operational anomalies; By analyzing the formula: Get the abnormal operation coefficient ,in represents the number of times the i-th operational abnormality occurs, Indicates the total number of abnormal operation occurrences. represents the weight of the i-th abnormal operation, where i ; Setting thresholds based on the value of the operational abnormality coefficient for real-time warning includes: when When an abnormality occurs, an early warning signal is issued to notify management personnel to handle the abnormal operation; when When the real-time video image recognition is running abnormally.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention uses the production process of high magnetic induction oriented silicon steel (HiB), especially the temperature control during the cold rolling process. Various temperature models are used to automatically control the temperature of the strip, including: deformation heat, friction heat transfer, heat convection and heat conduction. Based on the temperature model, the temperature of the strip is more accurately controlled, thereby improving the quality of HiB steel; The present invention learns the heat transfer coefficient through a neural network, making it closer to the actual value, thereby improving the accuracy and stability of temperature control; The present invention obtains coefficients under different conditions by changing different strip speeds, inlet temperatures and other conditions, which are used for temperature feedforward control and feedback control. This achieves precise control of strip temperature, thereby producing qualified HiB steel. The present invention improves the accuracy and safety and stability of temperature control by real-time monitoring of video images and analyzing operational abnormalities, thereby optimizing the production efficiency of HiB. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 This is a schematic diagram of the system module framework of the present invention; Figure 2 The present invention is a flow chart of the method. DETAILED DESCRIPTION

[0018] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] See also Figure 1-2 The first embodiment of the present invention provides a temperature control system for high-grade oriented silicon steel, comprising the following modules: Data acquisition module: real-time acquisition and measurement of strip temperature and video image data at the inlet and outlet sides of the rolling mill; Data processing module: pre-processing the collected strip temperature, including filtering, normalization, removal of outliers and erroneous data; pre-processing the collected video image data, including denoising and enhancement; Data analysis module: Real-time calculation of the strip temperature through the temperature model; analysis of the surface heat transfer coefficient based on the neural network model; Comprehensive analysis module: The final rolled piece temperature is calculated based on the temperature of the strip calculated by the temperature model; the flow rate of the cooling emulsion is automatically adjusted based on the calculation results to control the temperature of the strip; Real-time warning module: Analyzes operation abnormalities based on the collected and processed video images and calculates the operation abnormality coefficient based on the analysis results to issue real-time warnings.

[0020] Specifically, the strip temperature is mainly determined by the rolling speed, inlet temperature, emulsion flow rate, and emulsion temperature. There is a thermometer on the inlet and outlet sides of the rolling mill to measure the strip temperature in real time. The emulsion responsible for cooling the strip is controlled in sections, generally divided into the outside of the inlet mill, the inside of the inlet mill, the inlet roll gap, the outlet roll gap, the inside of the outlet mill, the outside of the outlet mill, and the backing bearing cooling. Each cooling section has a proportional servo valve, and the flow of each section can be controlled by controlling the opening degree. Oriented silicon steel is mainly used in the fields of transformer and reactor core manufacturing. The collected video images are de-noised, and the contrast, brightness, smoothness and edge information of the video images are adjusted and enhanced to make the details of the video images clearer. High magnetic induction oriented silicon steel has a high magnetic induction intensity, which can effectively reduce the core loss of the transformer, improve the efficiency and performance of the transformer, and reduce energy loss. Its production process is complex, and strict requirements are placed on parameters such as composition control, temperature, and pressure during the manufacturing process. Cold rolling is an important process in the production of HiB. HiB rolling generally uses silicon steel raw materials of about 2.3mm, uses the Sendzimir rolling mill, and is cold rolled once, and rolled to the finished thickness in five or six passes. In order to produce qualified HiB, unlike the cold rolling of other steels, the strip temperature needs to be controlled, first heated up, and then cooled down. Specifically, the first three passes are heated to a target temperature, and then the next pass is cooled to about 50°C. This requires real-time control of the strip temperature.

[0021] In one embodiment of the present invention, the temperature of the strip is calculated in real time by a temperature model, comprising the following steps:

[0022] The temperature model includes: a deformation heat model, a friction heat transfer model, a convection heat transfer model and a heat conduction model; The real-time calculation of the temperature of the strip includes: calculating the thermal deformation temperature rise according to the deformation heat model; calculating the friction temperature rise according to the friction heat transfer model; calculating the temperature of the strip after cooling according to the convection heat transfer model; and calculating the heat conduction temperature drop according to the heat conduction model.

[0023] Specifically, the temperature of the strip is automatically controlled by using various temperature models. The temperature models are divided into deformation heat, friction heat transfer, heat convection, and heat conduction. The deformation heat model is used to calculate the temperature rise of the strip due to plastic deformation during the rolling process. This model can predict the temperature rise under different reductions and rolling speeds. The friction heat transfer model is used to estimate the effect of the heat generated by the friction between the rolls and the strip on the temperature of the strip during the rolling process. The convection heat transfer model is used to calculate the temperature drop of the strip after the heat is dissipated by convection under the action of the coolant. This model can be used to optimize the cooling strategy to achieve ideal cooling effect and temperature control. The heat conduction model describes the process of heat inside the strip propagating inside the material through the conduction mechanism. By real-time monitoring of the rolling process and dynamically adjusting parameters, the temperature of the strip is ensured to be within an appropriate range, thereby ensuring product quality and process stability.

[0024] In one embodiment of the present invention, the temperature of the strip is calculated in real time by a temperature model, comprising the following steps: By analyzing the formula: The rolling force per unit area is obtained ; By analyzing the formula: Get thermal deformation temperature rise , represents the deformation heat, represents density, represents specific heat capacity, H represents inlet thickness, h represents outlet thickness, Indicates efficiency, Indicates the rolling force per unit area.

[0025] Specifically, the deformation heat of the rolled piece is the main source of temperature rise. During the rolling process, to evaluate the impact of deformation heat on temperature change, it is necessary to calculate the rolling force per unit area and the thermal deformation temperature rise, and comprehensively apply multiple formulas and parameters. The calculation of thermal deformation temperature rise depends on factors such as the rolling force per unit area, material density, specific heat capacity, and efficiency. Accurate measurement and estimation of these parameters are crucial. For example, poor lubrication conditions may cause the actual rolling force to be greater than the theoretical value, resulting in a higher temperature rise. At the same time, phenomena such as flow stress changes, dynamic recovery, and recrystallization of materials at high temperatures need to be considered. Therefore, when using the above formula for calculations, it is necessary to continuously adjust the model in combination with experimental data and on-site observations to ensure the accuracy and reliability of the prediction. By calculating the thermal deformation temperature rise, the temperature changes during the rolling process can be effectively evaluated, among which The efficiency is 0.9.

[0026] In one embodiment of the present invention, the friction temperature rise is calculated according to the friction heat transfer model, comprising the following steps: By analyzing the formula: Friction temperature rise ;in, represents friction heat, represents density, H represents inlet thickness, h represents outlet thickness, represents the average linear velocity, Indicates efficiency, is the rolling force per unit area, Represents the contact time constant.

[0027] Specifically, the friction temperature rise is obtained by comprehensive calculation of the contact time constant and friction heat. ,in The efficiency is 0.5. Frictional heat is a minor source of temperature rise. In the production of hot rolled strip, frictional heat is a minor source of heat. This frictional heat comes from the contact friction between the roll and the strip surface, and the friction when the strip passes through the guide roll and the tension roll. Although this heat is relatively small, its cumulative effect may have a significant impact on the temperature distribution of the strip under high-speed rolling and high pressure.

[0028] In one embodiment of the present invention, calculating the temperature of the strip after cooling according to the convection heat transfer model includes the following steps: By Newton's cooling formula: Where: represents the heat flux density, Indicates the strip temperature, Indicates the coolant temperature, represents the surface heat transfer coefficient; By differentiating the convection equation analytically: By analyzing the formula: The solution to the initial value problem is: Get the strip temperature ;in, Indicates the strip temperature before cooling, represents the time constant, Indicates the cooldown time.

[0029] Specifically, the cooling effect of the emulsion on the strip is calculated using the Newton cooling formula and the surface heat transfer coefficient. In the production of hot-rolled strip, convection heat transfer is the main way to cause the temperature drop of the strip. The strip is mainly lubricated and cooled by the emulsion, and the appropriate emulsion flow rate is the key to controlling the temperature. Convection heat transfer refers to the process in which the high-temperature strip contacts the surrounding lower-temperature emulsion during the hot rolling process, and the heat is taken away by the flow of the emulsion, thereby cooling the strip. In the production process of hot-rolled strip, the convection heat transfer model achieves efficient cooling through the contact between the high-temperature strip and the low-temperature emulsion. This process involves the flow and large heat capacity characteristics of the emulsion, which can quickly absorb the heat on the surface of the strip.

[0030] In one embodiment of the present invention, analyzing the surface heat transfer coefficient according to the neural network model includes the following steps: The surface heat transfer coefficient is analyzed by using a neural network model; The surface heat transfer coefficients at different flow rates and temperatures are used as training data for the neural network model; The surface heat transfer coefficient is input as the target variable into the neural network model for training; The temperature and flow data collected and processed in real time by the temperature control system of high-grade oriented silicon steel are input into the trained neural network model in real time to predict the surface heat transfer coefficient; According to the prediction results of the neural network model, the parameters are adjusted to make the surface heat transfer coefficient close to the actual value.

[0031] Specifically, the surface heat transfer coefficient is a variable. The surface heat transfer coefficient is different under different flow rates and temperatures. Here, a neural network is used to learn this heat transfer coefficient, making it closer to the actual value and improving the accuracy of temperature control. Collect surface heat transfer coefficient data under different flow and temperature conditions. Usually, an algorithm-optimized neural network model is used to train the network using the collected training data, and the network parameters are adjusted through an iterative optimization algorithm so that the model can accurately predict the surface heat transfer coefficient. This process may require a lot of training time and computing resources. Deploy temperature and flow sensors on the production line to collect data in real time. Use the trained neural network model to analyze the incoming data in real time, predict the surface heat transfer coefficient, and dynamically adjust the parameters of the temperature control system accordingly to ensure that the production process achieves the optimal heat exchange effect. In practical applications, the model needs to continuously verify its prediction accuracy, and make necessary feedback adjustments by comparing the actual results with the model predictions.

[0032] In one embodiment of the present invention, calculating the heat conduction temperature drop according to the heat conduction model includes the following steps: By analyzing the formula: Get the heat conduction temperature drop , represents the heat flow, represents density, represents the specific heat capacity, represents the thermal conductivity of the rolled piece, is the contact time constant, represents the equivalent roll temperature, Indicates the initial temperature of the strip.

[0033] Specifically, heat conduction is a secondary way of temperature drop. In the production process of hot rolled strip, heat conduction is induced by the contact between the roll and other rolls, thus affecting the temperature distribution of the strip. The contact between the strip and the roll and other rolls causes heat to be transferred from the strip to these rolls, resulting in a temperature drop of the strip. Calculate the energy loss caused by heat conduction when the strip is in contact with the roll.

[0034] In one embodiment of the present invention, the final rolled piece temperature is calculated based on the temperature of the strip calculated by the temperature model, comprising the following steps: By analyzing the formula: Get the final rolling temperature ,in, express Weight of thermal deformation temperature rise, express The weight of the friction temperature rise, express The weight of the strip temperature after cooling, express Weight of the temperature drop due to heat conduction.

[0035] Specifically, the comprehensive formula is applied to the actual production process and adjusted and controlled by the automation system. The production process is continuously monitored, data feedback is collected, and model parameters are continuously optimized to improve the accuracy and efficiency of temperature control. The calculation results of all the above models are combined to form a comprehensive formula for predicting and controlling the temperature of the strip. This may include weighted merging of the temperature changes calculated by each model to obtain the final temperature prediction value. Among them, is 0.4, is 0.3, is 0.2, is 0.1.

[0036] In one embodiment of the present invention, the flow rate of the cooling emulsion is automatically adjusted according to the calculation result to control the temperature of the strip, including the following steps: S1: According to the calculated final rolled product temperature, by changing the cooling flow rate of each part, the cooling flow rate is calculated using Newton's method, and the temperature of the outlet strip is controlled to reach the target value; S2: By changing the speed of the steel strip, the flow rate of the steel strip at different speeds is obtained, and the flow rate of the steel strip is pre-controlled using the speed-flow curve; S3: By changing the inlet temperature, different outlet temperatures and outlet temperature / inlet temperature influence coefficients are obtained, and feedforward control of the temperature is performed through the coefficients under different conditions; S4: Temperature feedback control is performed using the obtained flow rate and outlet temperature sensitivity coefficient.

[0037] Specifically, the final rolled piece temperature can be calculated by calculating the temperature changes of each part according to known conditions. In this way, the cooling flow of each part can be changed, and the Newton method can be used to obtain the appropriate cooling flow to control the temperature of the outlet strip to reach the target value. Then, by changing different strip speeds, the appropriate strip flow at different speeds can be obtained. This speed-flow curve can be used to pre-control the flow of the strip. It is also possible to obtain different outlet temperatures by changing the inlet temperature, thereby obtaining the outlet temperature / inlet temperature influence coefficient, and the coefficient under different conditions can be used for temperature feedforward control. A similar method can also be used to obtain the sensitivity coefficient of flow and outlet temperature for temperature feedback control. Through the above-mentioned pre-control, temperature feedforward, and feedback control, the strip temperature of the first few passes can be controlled to reach the target value, thereby rolling out qualified HiB steel.

[0038] In one embodiment of the present invention, the operation abnormality is analyzed based on the collected and processed video images and the operation abnormality coefficient is calculated based on the analysis result to issue a real-time warning, including the following steps: Based on the pre-processed video image data, the operation anomalies in the video image are analyzed through the deep learning model; Six types of operation abnormalities are marked, including: abnormal vibration of the rolling mill rolls, deviation of the rolled pieces, blockage of the rolling mill rolls, uneven force on the rolling mill rolls, low rolling speed, and overheating of the rolling mill; The six annotated operational anomalies were input into the convolutional neural network of the deep learning model for model training; The temperature control system of high-grade oriented silicon steel collects processed video images in real time to identify six types of operational anomalies; By analyzing the formula: Get the abnormal operation coefficient ,in represents the number of times the i-th operational abnormality occurs, Indicates the total number of abnormal operation occurrences. represents the weight of the i-th abnormal operation, where i ; Setting thresholds based on the value of the operational abnormality coefficient for real-time warning includes: when When an abnormality occurs, an early warning signal is issued to notify management personnel to handle the abnormal operation; when When the real-time video image recognition is running abnormally.

[0039] Specifically, based on the pre-processed video images, the deep learning model is used to analyze the operational anomalies in the video images. The six operational anomalies presented in the video images are carefully annotated and input into the convolutional neural network of the deep learning model for model training; the processed video images are collected in real time through the temperature control system of high-grade oriented silicon steel to identify the six operational anomalies. The operational anomaly coefficient is accurately calculated based on the analysis results. , assign the weight of the first abnormal operation to 0.3, the weight of the second abnormal operation to 0.25, the weight of the third abnormal operation to 0.2, the weight of the fourth abnormal operation to 0.15, the weight of the fifth abnormal operation to 0.05 and the weight of the sixth abnormal operation to 0.05; the square root of the sum of the squares of the weights and the constant in the denominator are used as correction constants to increase the accuracy of the results, and the sum is adjusted accordingly. Set the threshold according to the value of the abnormal operation coefficient for real-time warning. When When an abnormality occurs, an early warning signal is issued to notify management personnel to handle the abnormal operation; When the real-time video image recognition is running abnormally.

[0040] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. Temperature control system for high grade oriented silicon steel, characterized in that: Includes the following modules: Data acquisition module: real-time acquisition and measurement of strip temperature and video image data at the inlet and outlet sides of the rolling mill; Data processing module: pre-processing the collected strip temperature, including filtering, normalization, removal of outliers and erroneous data; pre-processing the collected video image data, including denoising and enhancement; Data analysis module: Real-time calculation of the strip temperature through the temperature model; analysis of the surface heat transfer coefficient based on the neural network model; Comprehensive analysis module: The final rolled piece temperature is calculated based on the temperature of the strip calculated by the temperature model; the flow rate of the cooling emulsion is automatically adjusted based on the calculation results to control the temperature of the strip; Real-time warning module: Analyzes operation abnormalities based on the collected and processed video images and calculates the operation abnormality coefficient based on the analysis results to issue real-time warnings.

2. The temperature control system of high grade grain-oriented silicon steel according to claim 1, characterized in that: The temperature of the strip is calculated in real time by the temperature model, including the following steps: The temperature model includes: a deformation heat model, a friction heat transfer model, a convection heat transfer model and a heat conduction model; The real-time calculation of the temperature of the strip includes: calculating the thermal deformation temperature rise according to the deformation heat model; calculating the friction temperature rise according to the friction heat transfer model; calculating the temperature of the strip after cooling according to the convection heat transfer model; and calculating the heat conduction temperature drop according to the heat conduction model.

3. The temperature control system of high grade grain-oriented silicon steel according to claim 2, characterized in that: The thermal deformation temperature rise is calculated according to the deformation thermal model, including the following steps: By analyzing the formula: The rolling force per unit area is obtained ; By analyzing the formula: Get thermal deformation temperature rise , represents the deformation heat, represents density, represents specific heat capacity, H represents inlet thickness, h represents outlet thickness, Indicates efficiency, Indicates the rolling force per unit area.

4. The temperature control system of high grade grain-oriented silicon steel according to claim 2, characterized in that: The calculation of friction temperature rise according to the friction heat transfer model includes the following steps: By analyzing the formula: Friction temperature rise ;in, represents friction heat, represents density, H represents inlet thickness, h represents outlet thickness, represents the average linear velocity, Indicates efficiency, is the rolling force per unit area, Represents the contact time constant.

5. The temperature control system of high grade grain-oriented silicon steel according to claim 2, characterized in that: The temperature of the strip after cooling is calculated according to the convection heat transfer model, including the following steps: By Newton's cooling formula: Where: represents the heat flux density, Indicates the strip temperature, Indicates the coolant temperature, represents the surface heat transfer coefficient; By differentiating the convection equation analytically: By analyzing the formula: The solution to the initial value problem is: Get the strip temperature ;in, Indicates the strip temperature before cooling, represents the time constant, Indicates the cooldown time.

6. The temperature control system for high-grade oriented silicon steel according to claim 5, characterized in that: The surface heat transfer coefficient is analyzed according to the neural network model, including the following steps: The surface heat transfer coefficient is analyzed by using a neural network model; The surface heat transfer coefficients at different flow rates and temperatures are used as training data for the neural network model; The surface heat transfer coefficient is input as the target variable into the neural network model for training; The temperature and flow data collected and processed in real time by the temperature control system of high-grade oriented silicon steel are input into the trained neural network model in real time to predict the surface heat transfer coefficient; According to the prediction results of the neural network model, the parameters are adjusted to make the surface heat transfer coefficient close to the actual value.

7. The temperature control system of high grade grain-oriented silicon steel according to claim 2, characterized in that: The heat conduction temperature drop is calculated according to the heat conduction model, including the following steps: By analyzing the formula: Get the heat conduction temperature drop , represents the heat flow, represents density, represents the specific heat capacity, represents the thermal conductivity of the rolled piece, is the contact time constant, represents the equivalent roll temperature, Indicates the initial temperature of the strip.

8. The temperature control system for high-grade grain-oriented silicon steel according to claim 1, characterized in that: The final rolled piece temperature is calculated comprehensively based on the temperature of the strip calculated by the temperature model, including the following steps: By analyzing the formula: Get the final rolling temperature ,in, express Weight of thermal deformation temperature rise, express The weight of the friction temperature rise, express The weight of the strip temperature after cooling, express Weight of the temperature drop due to heat conduction.

9. The temperature control system for high-grade oriented silicon steel according to claim 8, characterized in that: The flow rate of the cooling emulsion is automatically adjusted according to the calculation results to control the temperature of the strip, including the following steps: S1: According to the calculated final rolled product temperature, by changing the cooling flow rate of each part, the cooling flow rate is calculated using Newton's method, and the temperature of the outlet strip is controlled to reach the target value; S2: By changing the speed of the strip steel, the flow rate of the strip steel at different speeds is obtained, and the flow rate of the strip steel is pre-controlled using the speed-flow curve; S3: By changing the inlet temperature, different outlet temperatures and outlet temperature / inlet temperature influence coefficients are obtained, and feedforward control of the temperature is performed through the coefficients under different conditions; S4: Temperature feedback control is performed using the obtained flow rate and outlet temperature sensitivity coefficient.

10. The temperature control system of high grade grain-oriented silicon steel according to claim 1, characterized in that: Analyze the operation abnormality according to the collected and processed video images and calculate the operation abnormality coefficient according to the analysis results to issue a real-time warning, including the following steps: Based on the pre-processed video image data, the operation anomalies in the video image are analyzed through the deep learning model; Six types of operation abnormalities are marked, including: abnormal vibration of the rolling mill rolls, deviation of the rolled pieces, blockage of the rolling mill rolls, uneven force on the rolling mill rolls, low rolling speed, and overheating of the rolling mill; The six annotated operational anomalies were input into the convolutional neural network of the deep learning model for model training; The temperature control system of high-grade oriented silicon steel collects processed video images in real time to identify six types of operational anomalies; By analyzing the formula: Get the abnormal operation coefficient ,in represents the number of times the i-th operational abnormality occurs, Indicates the total number of abnormal operation occurrences. represents the weight of the i-th abnormal operation, where ; Setting thresholds based on the value of the operational abnormality coefficient for real-time warning includes: when , send out early warning signals and notify management personnel to deal with operational abnormalities; when , real-time video image acquisition to identify operational anomalies.

Images