Method for setting temperature of stainless steel strip annealing furnace by zone according to steel type

Abstract

The invention belongs to the technical field of temperature setting of stainless steel strip annealing furnaces, and discloses a method for zoning temperature setting of stainless steel strip annealing furnaces according to steel types. The specific technical scheme is as follows: a model system automatically obtains information of incoming steel coils, and at the same time collects production process data, matches and tracks information according to steel type and specification of the steel coils, realizes production and process data collection, data matching and storage of stainless steel strip annealing furnace lines and pickling parts, and generates real-time and historical trend curves, and at the same time builds a secondary model system platform for annealing furnaces and pickling, realizes production line steel coil tracking in each room and each zone of the production line, statistically analyzes production snapshot information of each steel coil according to a 50-meter step report, realizes optimal solution calculation of temperature setting value of each furnace zone, and automatically sends it to a primary PLC system, so as to provide temperature setting values ​​under steady-state production and transition state for the basic automation level, save energy consumption, and help enterprises to control costs more finely and optimize production decisions.

Description

Technical Field

[0001] The invention belongs to the technical field of temperature setting of stainless steel strip annealing furnaces, and in particular relates to a method for setting the temperature of stainless steel strip annealing furnaces by zones according to steel types. Background Art

[0002] Annealing is a metal heat treatment process, which is a heat treatment method for metal according to process temperature, process time, and cooling rate. Its purpose is to reduce hardness, soften workpiece, improve machinability; improve or eliminate various structural defects and residual stress caused by steel in the casting, forging, rolling and welding process, reduce deformation, cracking or crack tendency of workpiece; refine grains, improve structure to improve mechanical properties of workpiece, eliminate structural defects; uniform material structure and composition, improve material properties. Annealing temperature is the main control target of the whole process and the most important factor affecting product quality. Annealing furnaces are divided into three different types according to different heating methods: electric heating, fuel heating, and gas (natural gas) heating. They are electric heating annealing furnace, fuel annealing furnace, and gas (natural gas) annealing furnace. In the annealing furnace process requirements, temperature is the key indicator, and the temperature control accuracy and stability of the annealing furnace become the decisive factors of product quality.

[0003] Annealing is an important process in the production of hot-rolled and cold-rolled stainless steel strips. The cleaned and cooled strips are subjected to different heat treatments one by one, and finally rolled out of the furnace to become standard products. According to different forms, continuous annealing furnaces can be divided into vertical and horizontal types; according to the process, they can be divided into continuous annealing, galvanizing, aluminum plating and other annealing furnaces. The difference lies in the setting of open flame heating section, over-aging section, etc. The temperature accuracy of the annealing furnace directly affects the final quality of stainless steel products. Summary of the invention

[0004] In order to solve the technical problems existing in the prior art, the present invention provides a method for setting the zone temperature of a stainless strip annealing furnace according to the type of steel, which is used to calculate the zone temperature of the stainless strip annealing furnace. By integrating big data collection, cleaning, management, and modeling, the dynamic model calculation of the zone temperature of the annealing furnace is realized and the furnace control system is automatically issued, thereby saving energy and effectively reducing the energy consumption cost of the enterprise.

[0005] To achieve the above object, the technical solution adopted by the present invention is: a method for setting the temperature of a stainless steel strip annealing furnace according to the type of steel, and the specific steps are as follows:

[0006] Step 1: Build an automatic real-time accounting model, which is connected to the production system to automatically obtain the PDI information of the incoming steel coils, realize the production and process data collection, data matching and storage of the stainless steel strip annealing furnace line and pickling part, and generate real-time and historical trend curves;

[0007] Step 2: Build a secondary model system platform for annealing furnace and pickling, realize the tracking of production line steel coils in each room and each zone of the production line, statistically analyze the production snapshot information of each steel coil according to the 50-meter step report, perform big data cleaning, governance, and data analysis on the collected data, and use a multi-dimensional least squares linear regression algorithm to model the furnace zone temperature model and perform model simulation operations to achieve the optimal solution calculation of the temperature setting value of each furnace zone, and automatically send it to the first-level PLC system to provide the basic automation level with temperature setting values ​​under steady-state production and transition conditions.

[0008] In step one, the automatic real-time accounting model is matched with the production line model and the data of the MES system. Through the L1 or L2 system data interface, the product specifications, current speed and tension of the annealing furnace in each zone are obtained, and the actual production data is provided.

[0009] Summarize the data of annealing, pickling and other models or internal functional modules and write them into the database to display and query the data of each process and generate reports;

[0010] The automatic real-time accounting model reserves a data interface for connecting with other models or internal functional modules.

[0011] In step 1, the PDI information of the incoming steel coil includes the coil number, steel type, width, length, thickness, weight, time of entering each section and zone, and furnace temperature of the incoming steel coil.

[0012] Acquisition and assembly of key data:

[0013] According to the key data set in the model, it is obtained from the production PLC system and assembled according to pre-defined standards.

[0014] Acquisition and transmission of furnace charging information:

[0015] From the MES production system, the coil number is used as the keyword to accurately obtain key information such as steel coil information and planned production time, and the data is passed to the model system through the message interface processing program.

[0016] Collection and collation of process data:

[0017] The process data of the unit collected through the opc ua protocol include current speed, current TV value, current furnace temperature in each zone, oxygen content, total air flow, air-fuel ratio, weld tracking signal, furnace pressure, etc.

[0018] Calculation of model temperature settings:

[0019] The core algorithm of the model uses the least squares linear regression method to depict two variables (x, y) and a series of data pairs [(x1, y1), (x2, y2)...(xm, ym)] in xy coordinates. If it is found that these points are all near a straight line, then the initial expression of the equation of this line is:

[0020] y=mx+c

[0021] In this model system, y represents the furnace zone temperature after linear regression; x is the feature; m is the slope of the regression linear function, also called the gradient; c is the intercept of the temperature prediction line on the y-axis, that is, the error; in mathematics, m is the gradient and c is the intercept.

[0022] The optimal solution is to minimize the sum of squares of the errors between the obtained data and the actual data. Linear regression assumes that there is a linear relationship between the features and results in the data set.

[0023] It is necessary to calculate the m and c values ​​through model calculation so that the error between mx+c (predicted temperature result) and the actual y (actual furnace temperature) is minimized. The square error is used to measure the error between the estimated value and the actual value, and the square loss function L is used to calculate the error between the actual value and the predicted value:

[0024] L n =(y n *(mx n +c)) 2

[0025] The average loss over the entire dataset is:

[0026]

[0027] The best matching m and c are required to minimize L. The specific expression is:

[0028]

[0029] Among them, x represents the TV value, y represents the actual furnace temperature value, n represents the natural count of the sample 1, 2, 3..., m is the slope of the regression linear function, also called the gradient, c is the intercept of the temperature prediction line on the y-axis, mx+c represents the predicted temperature result value, ∑ is the summation formula, n represents the natural count of the sample 1, 2, 3..., N represents the upper limit value of the summation ∑, f is the predicted temperature result value, and the argmin function is often used to find the optimal parameter. This formula represents finding the parameter value that minimizes the sum of squared errors;

[0030] The least squares method is used to find the optimal value of the objective function, and the matching items are found by minimizing the sum of squares of errors. The least squares method is used to find the optimal solution of linear regression;

[0031] In the actual production process, the steel grades are divided into sections according to their actual TV values ​​(speed*thickness) in the unit, and the least squares linear regression is performed separately.

[0032] The calculation results are then referred to the real-time furnace conditions of each partition of the furnace section to perform model compensation and obtain the optimal solution.

[0033] Statistical analysis of production results:

[0034] A secondary system platform for annealing furnaces and pickling was built to enable tracking of steel coils in four rooms and eight zones of the production line. Production snapshot information of each steel coil was statistically analyzed using 50-meter step reports.

[0035] At the same time, through logical operations on the above data, the model can display intuitive and clear trend charts in real time, enabling relevant personnel to observe and analyze production change trends more conveniently and intuitively, providing a strong basis for decision-making.

[0036] Compared with the prior art, the specific beneficial effects of the present invention are as follows: the present invention is connected to the upstream MES system. When the steel coil enters the preheating section, the model system will automatically obtain important information such as the steel type, thickness, width, and length of the incoming steel coil, and collect production process data at the same time, match and track the information according to the steel type and specifications of the steel coil, and realize the production and process data collection, data matching and storage of the stainless steel strip annealing furnace line and the pickling part, and generate real-time and historical trend curves. At the same time, a secondary model system platform for annealing furnaces and pickling is built to realize the production line steel coil tracking in each room and each area of ​​the production line and the statistical analysis of the production snapshot information of each steel coil according to the 50-meter step report. The optimal solution calculation of the temperature setting value of each furnace area is realized, and it is automatically sent to the first-level PLC system to provide the temperature setting value for steady-state production and transition state for the basic automation level, saving energy consumption, helping enterprises to control costs more finely, and optimize production decisions.

[0037] A mathematical model of annealing furnace temperature based on digital technology was constructed, providing strong support and guarantee for the production management of stainless steel annealing furnaces. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is the overall flow chart of the present invention.

[0039] Figure 2 In order to perform least squares machine learning on the furnace temperature sample set produced by different steel types under different TV in the furnace zone, a linear regression function diagram of the optimal solution is fitted. In the figure, the large dots represent the actual values ​​of qualified furnace temperatures produced by different steel types under different TV in the furnace zone, and the dotted line represents the optimal solution regression function fitted after machine learning. DETAILED DESCRIPTION

[0040] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0041] The method for setting the temperature of stainless steel strip annealing furnace by zone according to steel type is as follows:

[0042] Step 1: Research and clarification of key data:

[0043] First, for the annealing furnace production process, in order to accurately determine the key data that the model needs to automatically calculate and present in real time, after comprehensive and detailed discussions, the following key data were identified:

[0044] 11) Detailed information of the incoming steel coils, including the coil number, steel type, width, length, thickness, weight, time of entering each section and zone, and furnace temperature. These data are the basic material information required to accurately calculate the furnace temperature.

[0045] 12) Various process data related to furnace temperature setting, such as current process speed, gas flow, fan speed, inlet and outlet plate temperature, and their impact on model temperature calculation.

[0046] Step 2: Acquisition and assembly of key data:

[0047] According to the key data set in the model, it is obtained from the relevant systems and assembled according to the pre-defined standards.

[0048] 21) Acquisition and transmission of furnace charging information:

[0049] From the MES production system, the coil number is used as the keyword to accurately obtain key information such as steel coil information and planned production time. Through the message interface processing program, these data are transferred to the model system.

[0050] 22) Collection and collation of process data:

[0051] Through the opc ua protocol, the process data of the unit is collected, such as current speed, current TV, current furnace temperature in each zone, oxygen content, total air flow, air-fuel ratio, weld tracking signal, furnace pressure, etc.

[0052] Step 3: Calculation of model temperature setting:

[0053] The core algorithm of the model uses the least squares linear regression method. Two variables (x, y) and a series of data pairs [(x1, y1), (x2, y2)...(xm, ym)] are plotted in xy coordinates. If it is found that these points are all near a straight line, then the initial expression of the equation of this line is:

[0054] y=mx+c

[0055] y is the result, x is the feature, m is the coefficient, c is the error, and in mathematics, m is the gradient and c is the intercept.

[0056] The optimal solution is to minimize the sum of squares of the errors between the obtained data and the actual data. Linear regression assumes that there is a linear relationship between the features and results in the data set.

[0057] We need to find m and c so that the result obtained by mx+c has the smallest error with the actual y. Here, we use the squared difference to measure the error between the estimated value and the actual value (if only the difference is used, there may be a negative number). The squared loss function L is used to calculate the error between the actual value and the predicted value.

[0058] L n =(y n -(mx n +c)) 2

[0059] The average loss over the entire dataset is:

[0060]

[0061] The best matching m and c are required to minimize L. The specific expression is:

[0062]

[0063] The least squares method is used to find the optimal value of the objective function. The least squares method is used to minimize the sum of squares of errors to find matching items. The least squares method is used to find the optimal solution for linear regression.

[0064] Step 4: Piecewise least squares linear regression:

[0065] In the actual production process, the steel grades are divided into sections according to their actual TV values ​​(speed*thickness) in the unit, and the least squares linear regression is performed separately.

[0066] Combination Figure 2 As shown in Table 1, from the actual production process, the steel type is segmented according to the actual TV value (speed * thickness) in the unit, and the least squares linear regression machine learning is performed respectively, and the m and c values ​​are fitted in segments, so as to fit the regression function in segments, and then the furnace temperature setting value is calculated in real time.

[0067] The calculation results are then referred to the real-time furnace conditions of each partition of the furnace section to perform model compensation and finally obtain the optimal solution.

[0068] Table 1 Segmented fitting regression function table of different steel grades produced under different TV

[0069] Steel Type Segment TV TV Range Segmented fitting m value Segmented fitting c value Regression function for piecewise fitting Su304 TV5 200 m5 c5 y5=m5*x+c5 Su304 TV4 140--200 m4 c4 y4=m4*x+c4 Su304 TV3 120--140 m3 c3 y3=m3*x+c3 Su304 TV2 90--120 m2 c2 y2=m2*x+c2 Su304 TV1 60--90 m1 c1 y1=m1*x+c1 Su304 TV0 30--60 m0 c0 y0=m0*x+c0

[0070] Step 5: Statistical analysis of production results:

[0071] A secondary system platform for annealing furnaces and pickling was built to achieve production line coil tracking in each room and each area of ​​the production line, and statistical analysis of production snapshot information of each steel coil was carried out according to 50-meter step reports.

[0072] At the same time, through logical operations on the above data, the model can display intuitive and clear trend charts in real time, enabling relevant personnel to observe and analyze production change trends more conveniently and intuitively, providing a strong basis for decision-making.

[0073] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims

1. A method for setting the temperature of a stainless steel strip annealing furnace according to the type of steel, characterized in that: The specific steps are as follows: Step 1: Build an automatic real-time accounting model, which connects with the production system to automatically obtain the PDI information of the incoming steel coils, realize the production and process data collection, data matching and storage of the stainless steel strip annealing furnace line and pickling part, and generate real-time and historical trend curves; Step 2: Build a secondary model system platform for annealing furnace and pickling, realize the tracking of production line steel coils in each room and each area of ​​the production line, statistically analyze the production snapshot information of each steel coil according to the 50-meter step report, perform big data cleaning, governance, and data analysis on the collected data, and use a multi-dimensional least squares linear regression algorithm to model the furnace zone temperature model and perform model simulation operations to realize the optimal solution calculation of the temperature setting value of each furnace zone, and automatically send the optimal solution to the first-level PLC system to provide temperature setting values ​​for steady-state production and transition states for the basic automation level.

2. The method for setting the temperature of a stainless steel strip annealing furnace according to steel type according to claim 1, characterized in that: In step 1, the automatic real-time accounting model matches the production line model and the data of the MES system, obtains the product specifications, current speed and tension in the annealing furnace in each zone through the system data interface, and provides actual production data; Summarize annealing and pickling model data and write them into the database to display and query data of each process and generate reports; The automatic real-time accounting model has reserved data interfaces for connecting with other models.

3. The method for setting the temperature of a stainless steel strip annealing furnace according to steel type according to claim 1, characterized in that: In step 1, the PDI information of the incoming steel coil includes the coil number, steel type, width, length, thickness, weight, time of entering each section and zone, and corresponding furnace temperature of the incoming steel coil.

4. The method for setting the temperature of a stainless steel strip annealing furnace according to steel type according to claim 1, characterized in that: The specific steps are as follows: Acquisition and assembly of key data: Obtain key data from the production PLC system according to the model settings and assemble the data according to pre-defined standards; Acquisition and transmission of furnace charging information: From the MES production system, the coil number is used as the keyword to obtain the steel coil information and the planned production time information, which is processed by the message interface processing program and then transferred to the model system; Collection and collation of process data: The process data of the unit is collected through the opc ua protocol, including the current speed, current TV value, current furnace temperature of each zone, oxygen content, total air flow, air-fuel ratio, weld tracking signal, and furnace pressure; Calculation of model temperature settings: The core algorithm of the model uses the least squares linear regression method to depict two variables (x, y) and data pairs [(x1, y1), (x2, y2)...(xm, ym)] in xy orthogonal coordinates. If it is found that these points are all near a straight line, the initial expression of the straight line equation is: y=mx+c The square error is used to measure the error between the estimated furnace zone temperature setting value and the true value, and the square loss function L is used to calculate the error between the true value and the predicted value: L n =(y n -(mx n +c)) 2 The average loss over the entire dataset is: Find the best matching m and c to minimize L. The specific expression is: Where x represents the TV value, y represents the actual furnace temperature value, n represents the natural count of the sample 1, 2, 3..., m is the slope of the regression linear function, c is the intercept of the temperature prediction line on the y-axis, mx+c represents the predicted temperature result value, ∑ is the summation formula, n represents the natural count of the sample 1, 2, 3..., N represents the upper limit value of the summation ∑, f is the predicted temperature result value, and the argmin function is often used to find the optimal parameters; By calculating the square sum of the errors L between the actual furnace temperature value and the model prediction value, we can find the matching items m and c to minimize L and fit the optimal equation solution of temperature linear regression. The steel grade is divided into sections according to its actual TV value in the unit, and the least squares linear regression is performed respectively. The calculation results are then used to refer to the real-time furnace conditions of each section of the furnace for model compensation, and finally the optimal solution is obtained; Statistical analysis of production results: A secondary system platform for annealing furnaces and pickling was built to achieve production line coil tracking in each room and each area of ​​the production line, and to statistically analyze the production snapshot information of each coil according to the 50-meter step report; through logical operations on the above data, the model displayed the trend chart in real time.

Images