JPWO2016038705A1 - Rolling simulation device - Google Patents

Abstract

An object of this invention is to provide the rolling simulation apparatus which can implement | achieve a highly accurate rolling simulation. A rolling simulation apparatus connected to a rolling system including a rolling line and a setting computer is provided. The rolling simulation apparatus includes a simulation condition setting unit that sets simulation conditions related to product quality and operating conditions in a virtual operation in which a virtual metal material is heated, rolled, cooled, and conveyed in a rolling line. Moreover, the control target value of the actuator group provided in the rolling line and the state prediction value of the virtual metal material are set so as to achieve the simulation condition by using the second model expression similar to the first model expression that the setting computer has. It has a virtual rolling line setting calculation part to calculate. Furthermore, when the model parameter group of the first model formula is updated, a parameter updating unit is provided that updates the model parameter group of the second model formula based on the model parameter group of the first model formula.

Description

  The present invention simulates a rolling line and a rolling operation for producing a metal product, and changes the dimensions and alloy composition of the metal material, the target value and process of heating, rolling and cooling, the operation stability, the process intermediate material, The present invention relates to a rolling simulation apparatus for predicting product quality.

  In metal materials such as steel, materials such as mechanical properties (strength, formability, toughness, etc.) and electromagnetic properties (permeability, etc.) depend on the alloy composition, heating conditions, processing conditions, and cooling conditions. Change. The alloy composition is adjusted by controlling the addition amount of the component elements, but one lot unit is large at the time of component adjustment, such as using a component adjustment furnace capable of holding molten steel of about 100 tons. For this reason, it is impossible to change the addition amount for each product of about 15 tons. Therefore, in order to manufacture a product of a desired material, it is important to make the material by making heating conditions, processing conditions, and cooling conditions appropriate. Furthermore, these process conditions are important not only for materials but also for product quality such as product dimensions and shapes, and for the realization of stable operations.

  In the hot rolling process, products are created by changing target values of various process parameters, which are process conditions relating to product quality and operating conditions. Process parameters include, for example, target temperatures at each point on the rolling line represented by finish entry temperature, finish delivery temperature, coiling temperature, plate thickness schedule for each pass, and rolling mills. Necessary use of descaler for each pass, use necessity of interstand cooling placed between stands of continuous rolling mill and initial flow rate, amount of lubricant used in finishing mill, cooling pattern used in runout table, etc. There is.

  The process control by the setting computer is performed so as to achieve the target product quality, that is, to achieve the target values of the various process parameters.

  The setting computer uses a model expression expressing physical phenomena of each process such as heating, rolling, cooling, and conveyance, and performs setting calculation so as to achieve the target values of the various process parameters. In the setting calculation, calculation of control target values for various actuators and prediction calculation of the state of the rolled material (metal material) at each stage of the process are repeatedly performed.

  Model formulas used in setting calculations to calculate physical quantities such as rolling load, deformation resistance, roll gap, temperature, and grain size are expressed as functions with input variables, machine constants, adjustment terms, and learning terms as inputs. .

  In Patent Document 1 and Patent Document 2, the setting computer compares the model predicted value with actual values such as temperature, shape, sheet thickness, sheet width, and rolling load obtained from sensors provided in the rolling line. Thus, a method for automatically learning a learning term of a model formula in advance and improving the accuracy of the model formula and the control accuracy using the model term is disclosed.

  Further, Patent Document 3 discloses a mechanical property such as a tensile test and a structure observation performed on some product coils, regarding a material prediction model for predicting a change in microstructure of a rolled material and a mechanical property of a final product. A model learning method has been proposed using actual values of mechanical properties obtained from the measurement test results. In general, model parameters (mechanical constants, adjustment terms, learning terms) are set using a stratified table divided by factors that are prone to model errors, such as steel grade, target plate thickness, target plate width, target temperature, etc. It is managed in the database belonging to the computer.

  The applicant has recognized the following documents including the above-mentioned documents as related to the present invention.

Japanese Patent No. 4119684 Japanese Patent No. 4402502 Japanese Unexamined Patent Application Publication No. 2010-172962 Japanese Unexamined Patent Publication No. 2001-25805

Plastic processing technology series 7 plate rolling (Corona) 198-229 pages The 173rd and 174th Nishiyama Memorial Technology Lecture “Prediction of structural changes and material properties of hot-rolled steel” (published by Japan Iron and Steel Institute), page 125

  Conventionally, the process parameters of the rolling operation are determined on the basis of many years of experience for each product specification, and a method of performing temperature control and dimensional control to achieve this has been common. However, in recent years, the demand for product specifications has become increasingly sophisticated and diversified, and the method based on experience cannot always determine these target values properly and achieve the final quality of the target such as desired dimensions and mechanical properties. There are cases where it is not possible. In addition, it may be difficult to determine whether the target value of the process parameter can be achieved with existing equipment.

  For this reason, a process model that models heating, processing, and cooling manufacturing processes is used to examine in advance whether a product manufactured under a certain alloy composition and process parameters will achieve the desired product quality. An apparatus for simulating a manufacturing process offline has been proposed (for example, Patent Document 4). The simulation device predicts the state of the metal material from time to time, such as dimensions, temperature, and position on the production line, and provides information on the alloy composition of the metal material and information on the processing history and temperature history obtained from the simulation of the production process. As input values, the microstructure prediction model predicts changes in the microstructure of the rolled material and the mechanical properties of the final product. The simulation device is also used to find alloy composition and process parameter target values that will achieve the desired quality.

  In the simulation, the same model as the model used in the actual operation setting calculation, a simplified model, or a highly accurate model in which a part is modeled more faithfully to a physical phenomenon is used. For the simulation, a setting computer used for actual operation and a database for managing the process parameters and model parameters are not used, and a computer and a database dedicated to the simulation are prepared separately. The same applies to the case where a simulation is performed for the purpose of simulating an actual rolling operation and the same function as the model formula used in the actual operation setting calculation is used. This is because the load for calculation and reading / writing to the database for simulation should not affect the actual operation.

  In actual operation, for a single rolled material, several times based on trial calculation of setting calculation during heating furnace in-furnace, setting calculation during extraction of heating furnace, and actual values collected at any time during rolling after extraction The setting calculation is repeated several tens of times. Since the parameters used, the actual values collected, and the output of setting calculations are enormous and data is frequently exchanged, the read / write load on the database is heavy. Further, the same setting calculation and reading / writing to the database are performed for several rolled materials before and after the rolled material. If a problem arises in the actual operation setting calculation due to a sudden increase in calculation load or frequent reading and writing to the database, the rolling operation must be stopped, resulting in a large loss. In actual operation setting calculation, calculation timing, database access method and timing are carefully designed based on the rolling pitch of the equipment and the number of collected data.

  By the way, in the case of simulation, since the simulated process is calculated offline, unlike the actual rolling operation process control, the actual values of the load, temperature, dimensions obtained by sensors installed in various parts of the process, and the product coil machine The actual value of physical properties cannot be obtained. Therefore, when actual rolling is performed under the same conditions as the simulation, it cannot be completely confirmed from the simulation whether the target value of the process parameter, and thus the target product quality, is achieved. In addition, since various controls such as feedforward, feedback, and dynamic during rolling based on the actual values are not performed, if model prediction errors occur, they are accumulated, and the target values of the process parameters are The difference from the actual value becomes large, and the product quality is not predicted correctly. Therefore, the accuracy of the model expression of the process used in the simulation (particularly, the accuracy of the model parameters) itself is significantly related to the accuracy of the simulation, that is, how much the actual rolling operation can be simulated.

  However, since the production line and metal material are simulated by a model on the simulator, and the setting calculation for actual operation does not share the computer or the database belonging to it, adjustment terms and learning in the model formula Term, the machine constant representing the mechanical characteristics of each actuator is not updated. In patent document 4, the device which can correct the model parameter which concerns on an installation easily is done so that an actual rolling mill line can be simulated even if an installation update is carried out. However, the learning terms and adjustment terms of the model have enormous types of parameters and are frequently updated every time rolling or adjustment is performed, so that it is difficult to appropriately correct the parameters every change. In that case, unlike the model inherent in the computer used for control and prediction of actual operation, the actual rolling process cannot be simulated with high accuracy, so even if the alloy composition and process parameters used in the simulation were applied to the actual operation, it was expected. There was a problem that product quality could not be obtained.

  The present invention has been made to solve the above-described problems, and is a rolling simulation capable of accurately simulating the rolling process of a virtual metal material using a rolling line in actual operation on a computer different from the actual operation setting computer. The object is to provide a device.

In order to achieve the above object, the first invention includes an actuator group that heats, rolls, cools, and conveys a metal material, and a sensor group that detects a control result value of the actuator group and a state result value of the metal material. A rolling simulation apparatus connected to a rolling system and a rolling system including a setting computer for calculating a control target value of the actuator group and a state prediction value of the metal material,
The setting calculator is
A model formula expressing physical phenomena of each process of heating, rolling, cooling, and conveyance in the rolling line, and having a first model formula expressed by a function having an input variable and a model parameter group as inputs,
Using the first model formula, in order to achieve process conditions related to product quality and operating conditions in actual operation, calculate the control target value of the actuator group and the state prediction value of the metal material,
The model parameter group of the first model formula is updated as needed based on a comparison value obtained by comparing the control target value and the predicted state value with the actual control value and the actual state value detected by the sensor group. ,
The rolling simulation apparatus is
A simulation condition setting unit that sets simulation conditions related to product quality and operation conditions in virtual operation in which the virtual metal material is heated, rolled, cooled, and transported in the rolling line;
The second model formula is the same as the first model formula, and the control target value of the actuator group and the predicted state value of the virtual metal material are used to achieve the simulation condition using the second model formula. A virtual rolling line setting calculation unit for calculating
A parameter updating unit that updates the model parameter group of the second model formula based on the model parameter group of the first model formula when the model parameter group of the first model formula is updated; It is characterized by providing.

The second invention is the first invention, wherein
The parameter update unit
An update timing designating unit for designating a timing at which calculation is not executed by the setting computer in actual operation;
And a parameter copy unit that copies the model parameter group of the first model formula to the model parameter group of the second model formula at the timing.

The third invention is the first invention, wherein
The parameter update unit
Of the model parameter group of the second model formula, an update parameter selection unit that selects a part of the model parameter group necessary for model calculation using the simulation condition in the virtual rolling line setting calculation unit,
A parameter copy unit that copies only the partial model parameter group selected by the update parameter selection unit from the model parameter group of the first model formula.

  According to the first aspect, when the model parameter group of the first model formula is updated, the model parameter group of the second model formula is updated based on the model parameter group of the first model formula. Thereby, the model parameter of a rolling simulation apparatus can be updated to the newest data in the setting computer of actual operation. For this reason, according to 1st invention, on the computer different from the setting computer of actual operation, the rolling process of the virtual metal material using the rolling line of actual operation can be simulated with high precision.

  According to the 2nd or 3rd invention, the model parameter of a rolling simulation apparatus can be updated to the newest data in the setting computer of actual operation, suppressing the load increase given to the calculation in the setting computer of actual operation.

The figure which shows an example of the hot sheet rolling line in Embodiment 1 of this invention. The block diagram which shows the rolling system in Embodiment 1 of this invention. It is a block diagram which shows the function of the rolling simulation apparatus 24 in Embodiment 1 of this invention. It is an input screen which inputs the chemical component of a virtual metal material. It is a figure which shows an example of the slab temperature rising pattern in the heating furnace. It is a figure for demonstrating the cooling pattern in the runout table. It is a figure for demonstrating the cooling pattern in the runout table. It is a figure which shows the model group and model parameter table group which the virtual rolling line setting calculation part 32 has. It is a figure for demonstrating the process which the parameter update part 33 performs. 3 is a block diagram showing a configuration of a parameter update unit 33. FIG. It is a figure which shows the update timing suitable for updating the model parameter used with the rolling simulation apparatus 24 in the newest state. It is a figure which shows one specific example of the process which updates the adjustment term and learning term which are required for simulation to the newest state, when a simulation execution command is received. It is a flowchart of the process routine which updates the parameter of each model inherent in the virtual rolling line setting calculation part 32 to the same value as the newest parameter used by actual operation. It is a flowchart which shows one procedure which examines an alloy composition and manufacturing conditions using the rolling simulation apparatus.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In addition, the same code | symbol is attached | subjected to the element which is common in each figure, and the overlapping description is abbreviate | omitted.

Embodiment 1 FIG.
[System Configuration of Embodiment 1]
(Rolling line)
FIG. 1 is a diagram showing an example of a hot sheet rolling line in Embodiment 1 of the present invention. The object of the subsequent description is a simulator simulating the hot sheet rolling line shown in FIG. This simulator can also be applied to other rolling lines.

  The rolling line includes a heating device, a rolling mill, a cooling device, a winding device, and a conveyance table that connects them. These devices are driven by actuators such as electric motors and hydraulic devices. Specifically, the rolling line 1 shown in FIG. 1 includes a heating furnace 11, a rough rolling mill 12, a bar heater 13, a finishing mill entry-side thermometer 14, a finishing mill 15, and a finish in order from the upstream side of the conveyance table 10. A rolling mill outlet side thermometer 16, a runout table 17, a winder inlet side thermometer 18, and a winder 19 are provided.

  The heating furnace 11 is a furnace for heating the slab. The heating furnace 11 is controlled to obtain a desired slab temperature rising pattern and heating furnace extraction temperature. The rough rolling mill 12 includes a single or a plurality of stands. In the example illustrated in FIG. 1, the rough rolling mill 12 is a reversible rough rolling mill including a single stand. The bar heater 13 is a device that raises the temperature of the rolled product by electromagnetic induction heating or the like in order to control the temperature of the rolled product (including the state in the middle of the product from the slab to the finished product). The finishing mill 15 is composed of one or a plurality of stands, and in the example shown in FIG. 1, is a tandem type finishing mill composed of seven stands. The runout table 17 is a cooling device that cools the rolled product with cooling water in order to control the temperature of the rolled product. In addition, the rolling line 1 may be provided with a cooling table, a forced cooling device, etc. as a cooling device. The winder 19 is an apparatus for winding a rolled product into a shape that can be easily conveyed. The conveyance table 10 is an apparatus for conveying the rolled product in each process to the next process. These devices are driven by actuators such as electric motors and hydraulic devices.

  FIG. 2 is a block diagram showing the rolling system in Embodiment 1 of the present invention. The rolling system 20 shown in FIG. 2 has a hierarchical structure from level 0 to level 3. Level 0 includes a drive control device that controls an electric motor that drives each device of the rolling line 1 and a hydraulic device that drives each device of the rolling line 1. Level 1 includes a controller 21 for control. Level 2 is composed of a setting computer 23. In addition, it is good also as a structure which replaces with the setting computer 23 and uses a process controller. Level 3 is composed of a host computer 25 for production management. The rolling simulation device 24 does not affect rolling in actual operation, but is connected to the setting computer 23 for parameter update.

(Setting calculator)
In the hot rolling process in actual operation, products are made differently by changing process conditions relating to product quality and operating conditions, that is, target values of the various process parameters described above. The process control by the setting computer 23 is performed so as to achieve the target product quality, that is, to achieve the target values of the various process parameters.

  The target value of the process parameter may be specified from the upper computer 25 of the level 3 that is higher than the setting computer 23 of the level 2. In addition, the target value of the process parameter may have a table in a database belonging to the setting computer 23, and may be specified using steel type, plate thickness, plate width, etc. as keys. Further, the target value of the process parameter may be changed during rolling by an operator.

  The setting computer 23 represents a model formula expressing physical phenomena of each process such as heating, rolling, cooling, and conveyance in the rolling line 1 (hereinafter, the model formula of the setting computer 23 is also referred to as “first model formula”). Have. The setting calculator 23 uses the first model formula to perform setting calculation so as to achieve the target values (process conditions) of the various process parameters described above in actual operation. In the setting calculation, calculation of control target values of various actuators and calculation of the state of the rolled material (state prediction value of the metal material) at each stage of the process are repeatedly performed.

  The control target value of the actuator includes a roll gap of the rolling mill, a rolling speed, a conveying speed, a flow rate of a descaler and various sprays, and ON / OFF of a valve of the runout table. The state of the rolled material (predicted value of the state of the metal material) at each stage of the process includes dimensions, shape, temperature, microstructure, and the like.

  The controller 21 for control receives the setting calculation result from the setting computer 23, and controls various actuators so as to follow the control target value. In the actual hot rolling process, various sensors are installed throughout the rolling line to monitor and collect actual values of parameters that affect process control, such as temperature, shape, plate thickness, plate width, and rolling load.

  These actual values are used for process control, accuracy improvement of the model formula (first model formula), and quality control. Compare the target value of the process parameter with the actual value obtained by various sensors, or the actual value and the actual value recalculated from the calculated value. If the target value of the process parameter is not achieved, calculate the setting again. . Based on the result, various controls such as feedforward control, feedback control, and dynamic control are performed.

  The model formula (first model formula) that calculates physical quantities such as rolling load, deformation resistance, roll gap, temperature, and grain size used in the setting calculation is input variables, model parameter groups (machine constants, adjustment terms, learning This is expressed as a function with ( The input variable is a physical quantity correlated with the model output. For example, when the model output is a rolling load, the deformation resistance, the width of the rolled material, the amount of reduction, and the like are input variables. The mechanical constant is a physical quantity that represents the mechanical characteristics of the actuator, such as the roll diameter, mill curve, and spray flow rate of the rolling roll. The machine constant is updated at any time because it changes due to roll change, equipment repair and adjustment, and aging. The adjustment term and the learning term are terms for improving the prediction accuracy of the model formula.

  No matter how well the process model equation simulates a physical phenomenon, a model prediction error actually occurs. Therefore, the engineer finely adjusts the coefficient and constant applied to each term in the model formula to improve the prediction accuracy of the model formula. The adjustment term is a coefficient or constant of each term in the model formula, and a factor that is likely to cause a model error, for example, a stratified table divided by steel type, target plate thickness, target plate width, target temperature, etc. Each layer is managed in a database belonging to the setting computer 23. The adjustment term is adjusted not only at the time of starting the operation, but mainly when rolling a new steel type or rolling with a combination of new process parameters. The adjustment term may be adjusted by an engineer based on experience or numerical analysis results, or in recent years may be semi-automatically adjusted using a statistical technique such as a neural network. The learning term is a term that is multiplied and added to the model expression in order to fill an error between the model output and the actual process output.

(Rolling simulation equipment)
FIG. 3 is a block diagram showing functions of the rolling simulation device 24 according to Embodiment 1 of the present invention. The rolling simulation apparatus 24 simulates each process in the hot sheet rolling line shown in FIG. 1, and the operational stability when the metal material dimensions, alloy composition, heating, rolling, and cooling target values and processes are changed. Predict the state of rolled material and product quality during the process. The rolling simulation device 24 includes a simulation condition setting unit 31, a virtual rolling line setting calculation unit 32, and a parameter update unit 33. The rolling simulation device 24 is a computer including an arithmetic processing device, a storage device, and an input / output device. The storage device stores a program describing the processing contents of the above-described units. Each of the above sections is realized by executing a program loaded from the storage device on the arithmetic processing unit.

((Simulation condition setting part))
The simulation condition setting unit 31 sets simulation conditions related to product quality and operation conditions in a virtual operation in which a virtual metal material is heated, rolled, cooled, and conveyed in the rolling line 1. Details will be described below.

  The simulation condition setting unit 31 sets the parameters of the rolling operation process in the rolling simulation device 24 as simulation conditions. Here, the parameters of the rolling operation process are, for example, the alloy composition and dimensions of the rolled material, the target plate thickness, the target plate width, the slab temperature rising pattern in the heating furnace, the heating furnace discharge, which are given from the host computer 25 in the actual operation. Side temperature, finish delivery target temperature, finish entry target temperature, cooling pattern, winding target temperature, and the like. In addition, for example, in the actual operation, the parameters of the rolling operation process are set in advance in the setting computer 23 for each steel type and each target sheet thickness category, or are given from the HMI by the operator from the HMI and the distribution of the reduction rate of each pass. , Plate speed and acceleration rate.

  For each simulation condition, the operating conditions of the metal material rolled or planned to be rolled in the actual operation stored in the upper computer 25 or the setting computer 23 of the actual operation are copied via a communication LAN or a storage medium. Can be used. In addition, all or some of the conditions can be set manually. Further, it is possible to reuse the simulation conditions used in the past by the rolling simulation apparatus or to change a part of the simulation conditions.

  FIG. 4 is an input screen for inputting chemical components of a virtual metal product (virtual metal material). In actual operation, since one lot unit is large and is about 15 tons, it is impossible to change the addition amount of the alloy component for each product. Therefore, the simulation condition setting unit 31 includes, for example, the content (wt%) of each chemical component in each virtual metal product so that the change in product quality when the alloy composition is changed can be easily calculated. Enter to set. It is also possible to call the alloy composition of the metal product rolled in actual operation or the alloy composition used in the past simulation as a reference value, and to simulate by changing a part of it.

  FIG. 5 is a diagram illustrating an example of a slab temperature rising pattern in the heating furnace 11. The slab temperature rising pattern in the heating furnace 11 and the heating furnace extraction temperature also affect the product material and quality. For example, when the slab is not heated sufficiently, the solid solution amount of the microalloy is not sufficiently obtained, the Solute drag effect due to the solid solution microalloy is reduced, and the amount of precipitation during rolling and cooling after extraction is reduced, There is a concern that the pinning effect due to precipitates is reduced. Furthermore, rolling a low-temperature rolled material causes a hard material to be rolled, and there is a concern that the rolling operation may become unstable due to an increase in the rolling load in the rolling mill and the power consumption of the rolling motor may increase. The simulation condition setting unit 31 sets a slab temperature rising pattern in the heating furnace 11 as shown in FIG. When the temperature pattern does not affect the quality or when a simple calculation is desired, only the target value of the heating furnace extraction temperature is set.

  In actual operation, in order to manage the temperature of the rolled material during rolling, target values of process parameters such as finishing target temperature, finishing input target temperature, and winding target temperature are set via the host computer 25 or HMI. The rolling speed, the temperature rising pattern of the heating device in the middle of the rolling line, various sprays, and the cooling pattern on the run-out table 17 are controlled so as to follow the target value given by the operator. The cooling pattern may be specified by the host computer 25. In the simulation condition setting unit 31, the finish side target temperature and the finish side target temperature are used as simulation conditions so that the effect on product quality and rolling operation when the temperature history during rolling is changed can be confirmed by simulation. The winding target temperature and the cooling pattern on the run-out table 17 are set. 6 and 7 are diagrams for explaining a setting example of the cooling pattern in the run-out table 17. The cooling pattern in the run-out table 17 includes three patterns of pre-cooling that preferentially uses the upstream cooling equipment, post-cooling that preferentially uses the downstream cooling equipment, and gentle cooling that uses all the cooling equipment. There is a method of selecting either of them and setting the cooling rate of the zone for water cooling and the time for air cooling as target values. Further, as the cooling pattern in the run-out table 17, a pattern in which water cooling is performed on the upstream side and downstream side of the cooling facility shown in FIG. 7 and air cooling is selected in the middle stream is selected. For example, the water cooling rate on the upstream side and the air cooling time are selected. And a method of setting the temperature at the midpoint of the runout table as a target value.

  In the setting computer 23 used in actual operation, in order to achieve stable rolling and sheet passing, target values of process parameters such as the rolling amount of each pass, the rolling rate distribution, the plate passing speed and the acceleration rate are set in advance for each steel type and target plate. Store in the database for each thickness category. Alternatively, the operator inputs a target value for the process parameter. On the other hand, in the rolling simulation device 24 used in the virtual operation, the simulation condition setting unit 31 changes the product quality and the rolling operation when the rolling amount and rolling rate distribution of each pass, the sheet feeding speed and the acceleration rate are changed. Set simulation conditions so that the impact can be calculated easily.

((Virtual rolling line setting calculation part))
The virtual rolling line setting calculation unit 32 has a model formula similar to the first model formula (referred to as a second model formula) and uses the second model formula to achieve the simulation conditions of the actuator group. A control target value and a predicted state value of the virtual metal material are calculated. Details will be described below.

  The virtual rolling line setting calculation unit 32 sets the setting values of each process for rolling the virtual metal material on the virtual rolling line so as to follow each target value given by the simulation condition setting unit 31, and the metal every moment. Calculate material dimensions, location, and temperature.

  FIG. 8 is a diagram illustrating a model group and a model parameter table group included in the virtual rolling line setting calculation unit 32. The virtual rolling line setting calculation unit 32 includes a process model, a transfer model, a temperature model, and a material model as a model group. The process model calculates set values for each rolling process such as a heating device, a rolling device, and a cooling device. The transport model calculates the position of the virtual metal material at each time. The temperature model calculates the temperature of the virtual metal material at each time at each location. The material model predicts the microstructure of the metal material and the final product material at each time at each location on the virtual rolling line based on the alloy composition, processing history, and temperature. In addition, the virtual rolling line setting calculation unit 32 includes a storage device such as a database that stores a model parameter table group for storing the parameters of each model described above, and performs simultaneous calculation of each model formula.

  The process model uses information on the position of the virtual metal material at each time at each location given by the transfer model and the temperature of the virtual metal material at each time at each location given by the temperature model, and a target value given by the simulation condition setting unit 31 Set temperature pattern of heating furnace 11, rolling pass schedule, roll gap, machining history of virtual metal material from time to time, dimensions and shape, rolling speed, ON / OFF setting of various sprays and flow rate setting, The cooling setting of the runout table 17 is calculated.

  The transfer model calculates the position of the virtual metal material at each time at each location using the distance between the processes and the path schedule given by the process model. In addition, the transport model calculates the transport speed that follows each target temperature using the temperature information of the virtual metal material given by the temperature model.

  The temperature model includes virtual metal material dimension information in each process, machine specification information, a path schedule given from the simulation condition setting unit 31 and the process model, a roll gap, a rolling speed, a conveying speed, and a heating device in the middle of the rolling line. The temperature of the virtual metal material at each time at each location on the virtual rolling line is calculated from information on the command value such as the temperature rising pattern.

  The material model predicts the microstructure of the virtual metal material during and after the rolling process by using the processing history of the virtual metal given by the process model and the temperature history information given by the temperature model. The predicted microstructure is, for example, the fraction of each structure such as particle size, dislocation density, austenite, ferrite, pearlite and the like. Furthermore, parameters related to mechanical properties such as yield stress and tensile strength are calculated based on the microstructure prediction results. A variety of microstructure prediction models that formulate metallurgical phenomena have been proposed and consist of a group of mathematical expressions that represent static recovery, static recrystallization, dynamic recovery, dynamic recrystallization, grain growth, etc. Is widely known. An example is described in pages 198 to 229 of the plastic working technology series 7 plate rolling (Corona). It is widely known that materials represented by mechanical properties such as yield stress and tensile strength can be predicted from metallographic information and alloy composition. An example is published on page 125 of the 173rd and 174th Nishiyama Memorial Technology Course “Structural Change and Material Prediction of Hot Rolled Steel” (published by Japan Iron and Steel Institute).

  The process model, the transfer model, the temperature model, and the material model are represented by the same function as the model formula (first model formula) inherent in the setting computer 23 used in the actual operation of hot rolling. For example, in recent years, a method has been widely used in which all computer architectures are copied in a virtual machine environment (Virtual Machine) and virtually realized on different computers. According to this, the model formula (first model formula) of the setting computer 23 used for the actual operation and the database structure for managing the model parameters can be ported to the rolling simulation device 24. Therefore, the rolling simulation device 24 has a model formula (second model formula) similar to the first model formula. The model formula is a function with input variables, machine constants, and adjustment terms as inputs, and is expressed by the following formula.

here,
f: Model expression that does not include learning term Y: Output of model expression that does not include learning term X _i : Input variable m _i related to model expression f: Machine constant a _j : Adjustment term

  The input variable is a physical quantity correlated with the model output. For example, when the model output is a rolling load, the deformation resistance, the width of the rolled material, the amount of reduction, and the like are input variables. The mechanical constant is a physical quantity representing mechanical characteristics such as a roll diameter of a rolling roll, a mill curve, and a spray flow rate. The machine constant changes due to roll change, periodic repair, equipment replacement, aging, and the like. In actual operation, the machine constant is managed by a database table belonging to the setting computer 23 used for actual operation, and is corrected as needed in accordance with the above change. The adjustment term is a term for improving the prediction accuracy of the model formula. The adjustment term is a coefficient or a constant that is provided to reduce the model error and is allowed to be corrected. The adjustment term belongs to the setting computer 23 used for actual operation for each stratification using a stratification table divided by factors that are likely to cause model errors, for example, steel type, target plate thickness, target plate width, target temperature, etc. Managed in the database. In the actual operation, the adjustment term is adjusted mainly when a new steel type is rolled or when a new process parameter combination is used, in addition to when the operation is started. An engineer may make adjustments based on experience or numerical analysis results of actual operations, or in recent years, semi-automatic adjustments may also be made using statistical methods such as neural networks.

  In the actual hot rolling process, various sensors are installed in the rolling line 1 to monitor and collect the actual values of parameters that affect the process control, such as temperature, shape, plate thickness, plate width, and rolling load. These actual values are used for process control, accuracy improvement of the model formula (first model formula), and quality control. Compare the model predicted value of the setting calculation with the actual value acquired by various sensors, the actual value and the actual calculated value recalculated from the calculated value, and learn the model formula. A method for improving the used control accuracy is used. The learning term is multiplied or added to the model expression to fill in the error between the model output and the actual process output. Each of the multiplication type and the addition type is expressed as follows.

Multiplication type:
Y _L = Z _p · Y (2)
Addition type:
Y _L = Y + Z _A (3)

here,
Y _L : Prediction result of learned model expression Y: Output of model expression not including learning term Z _p : Multiplicative learning term Z _A : Additive learning term

  The learning term is updated by obtaining the actual value of the parameter corresponding to the output of the model formula with a sensor or the like. For example, in the multiplication type, the learning term is updated as follows.

here,
Z _P ^ACT : Multiplying learning term Y ^ACT calculated based on the actual value Y ^ACT : Actual value of parameter according to the model equation output Y: Model equation output not including the learning term Z _P ^NEW : Multiplicative learning term Z _P after update ^OLD : multiplication learning term before update α: smoothing gain

  The learning term is automatically updated for each stratification using a stratification table divided by factors that are likely to cause model errors, such as steel type, target plate thickness, target plate width, target temperature, and the like. Regarding the microstructure prediction model for predicting changes in the microstructure of the rolled material and mechanical properties of the final product, the results of mechanical properties measurement tests such as tensile tests and structural observations performed on some product coils The model is learned using the actual values of mechanical properties obtained in step (1). Model parameters of the model formula (first model formula) used in the actual operation setting calculation, that is, machine constants, adjustment terms, and learning terms are managed in a database belonging to the actual operation setting computer 23.

  The model model (second model formula) of the process model, the transfer model, the temperature model, and the material model in which the virtual rolling line setting calculation unit 32 of the rolling simulation apparatus 24 in FIG. 8 is used is used for the actual operation of hot rolling. A function having the same definition as the model formula (first model formula) inherent in the setting computer 23 is used. The model parameter table group in which the parameters of the machine constant, the adjustment term, and the learning term are stored for each layer is managed by a database belonging to the virtual rolling line setting calculation unit 32. The database table belonging to the virtual rolling line setting calculation unit 32 has the same structure as the table storing machine constants, adjustment terms, and learning terms in the database belonging to the setting computer 23 used for actual operation.

((Parameter update part))
FIG. 9 is a diagram for explaining processing executed by the parameter update unit 33. The parameter updating unit 33 updates the model parameter group of the second model formula based on the model parameter group of the first model formula when the model parameter group of the first model formula is updated. Details will be described below.

  In the parameter update unit 33, as shown in FIG. 9, the model parameters of the model formula of the virtual rolling line setting calculation unit 32, that is, machine parameters, adjustment terms, and learning terms are parameters of the database belonging to the actual operation setting computer 23. Update based on parameters stored in tables. In simulation, unlike actual process control of rolling operations, it is impossible to obtain actual values of loads, temperatures, dimensions, and actual values of mechanical properties of product coils obtained by sensors installed in various parts of the process. It is necessary to prevent the calculation by simulation and the load by reading and writing to the database from affecting the setting calculation of actual operation. Therefore, in the simulation, the setting computer 23 and its database used for actual operation are not used, and a computer and database dedicated to the simulation are used. Therefore, the machine constant, the adjustment term, and the learning term in the second model formula are not updated at the same timing as the first model formula of the setting computer 23 used for actual operation. The parameter updating unit 33 does not affect the actual operation setting calculation, and the second model formula of the virtual rolling line setting calculation unit 32 of the simulation is ensured so as to ensure the same model accuracy as the actual operation setting calculation. Model parameters, that is, machine constants, adjustment terms, and learning terms are updated.

  FIG. 10 is a block diagram illustrating a configuration of the parameter update unit 33. As illustrated in FIG. 10, the parameter update unit 33 includes an update timing designation unit 41, an update parameter selection unit 42, and a parameter copy unit 43. The update timing designation unit 41 automatically designates the timing for updating the simulator parameters. For example, the timing at which the calculation is not executed by the setting computer 23 in the actual operation is designated. The update parameter selection unit 42 selects a parameter to be updated. For example, from the model parameter group of the second model formula, a part of the model parameter group necessary for the model calculation using the simulation condition in the virtual rolling line setting calculation unit 32 is selected. The parameter copy unit 43 stores only a part of the model parameter group selected by the update parameter selection unit 42 at the update timing obtained from the update timing specification unit 41 in the database belonging to the setting computer 23 for actual operation. Copy from the model parameter group of the first model formula.

  By the way, in setting calculation of actual operation, during rolling operation, a trial calculation of setting calculation is executed for one rolled material in the furnace, and setting for setting up the actuator at the time of heating furnace extraction is performed. The calculation is executed, and after the extraction, the actual value is collected and the setting calculation is repeated several to several tens of times based on the value. Since the parameters used, the actual values collected, and the output of setting calculations are enormous and frequent, the load of reading and writing to the database is large. In addition, since the same setting calculation and reading / writing to the database are performed for several rolled materials before and after the rolled material, the calculation timing and access to the database are based on the rolling pitch of the equipment and the number of collected data points. Methods and timing are carefully designed and controlled.

  However, when the rolling line stops, for example, a roll change period in which the line is stopped for several tens of minutes at a frequency of several hours, or a line is stopped for several hours to several tens of hours at a frequency of several days or weeks. During the period of the regular repair, the calculation load of the setting computer 23 and reading / writing to the database are lost, or the load and the frequency of reading / writing are significantly reduced. The update timing designating unit 41 inherent in the parameter updating unit 33 of the rolling simulation device 24 selects the roll change, regular repair, and facility update periods in the actual operation as the model parameter update timing of the rolling simulation device 24. For example, it does not affect the setting calculation of actual operation.

  The model parameters (machine constants, adjustment terms, learning terms) used for the actual operation setting calculation are updated at different timings. For example, among the mechanical constants, the initial roll diameter of the rolling roll is changed every time the roll is changed every several hours. Even with the machine constant, the mill curve, which is an indicator of the elongation of the rolling mill, is updated over a relatively long span of several months or years. The flow rates of various sprays change over time, but the amount of change is gradual, and it is difficult to measure the flow rate if a flow meter is not installed in advance. Therefore, the flow rates of various sprays are not measured unless there are special circumstances, such as when a malfunction occurs or when equipment is updated. The adjustment term is adjusted not only at the time of starting the operation, but mainly when rolling a new steel type or rolling with a combination of new process parameters. Of these, the learning term has the highest update frequency. The learning term is managed for each rolling by a stratified table divided by factors that are likely to cause model errors, such as steel type, target plate thickness, target plate width, target temperature, etc. The learning terms in the corresponding layer of the model formula are updated. If the model parameters used in the rolling simulation device 24 are updated at a frequency equivalent to the update frequency of the model parameters used in actual operation, the rolling in actual operation can be accurately simulated.

  FIG. 11 is a diagram showing an update timing suitable for updating the model parameters used in the rolling simulation device 24 to the latest state. For example, at the timing shown in FIG. 11, the same value as the actual operation parameter is copied to each parameter used in the rolling simulation device 24. In actual operation, the adjustment term and the learning term may be updated at a frequency that is higher than that at the time of regular repair or roll change, or at a different timing. However, since the number of these parameters is enormous, the automatic update of all parameters should be performed at the timing when the rolling operation is stopped. On the other hand, in the rolling simulation device 24, a simulation including the dimensions and alloy composition of the rolling material to be simulated given by the simulation condition setting unit 31, target product dimensions and materials, and parameters of the rolling operation process such as rolling reduction distribution and cooling pattern, etc. From the conditions, the model parameters necessary for the simulation to be executed are clear. Therefore, if the adjustment term and the learning term necessary for the simulation are updated to the latest state at the time of executing the simulation, it is possible to accurately simulate the rolling in the actual operation under the conditions.

  FIG. 12 is a diagram illustrating one specific example of the process of updating the adjustment term and the learning term necessary for the simulation to the latest state when a simulation execution command is received. First, the steel type and the plate thickness classification of the virtual rolled material used for the simulation are identified from the alloy composition, the target plate thickness, etc. among the simulation conditions. In the example shown in FIG. 12, the simulation condition setting unit 31 sets steel type = C and 0.1 ≦ thickness <0.3 as simulation conditions.

  The database table group belonging to the actual operation setting computer 23 and the database table group belonging to the rolling simulation device 24 have the same table structure, and data can be copied between the databases. The database includes an adjustment term table group, a learning term table group, and a machine constant table group. In the example illustrated in FIG. 12, the update parameter selection unit 42 selects, as an update parameter, a parameter of steel type = C and 0.1 ≦ sheet thickness <0.3 from both table groups. The update parameter is notified to the update timing designation unit 41 and the parameter copy unit 43. The update timing specifying unit 41 confirms that the update of the selected parameter does not affect the actual operation setting calculation. When there is no influence, the update timing is designated and notified to the parameter copy unit 43. Then, the parameter copy unit 43 copies the selected update parameter from the database belonging to the actual operation setting computer 23 to the database belonging to the rolling simulation device 24 at the designated update timing.

(flowchart)
FIG. 13 is a flowchart of a processing routine for updating the parameters of each model inherent in the virtual rolling line setting calculation unit 32 to the same values as the latest parameters used in actual operation. This processing routine is repeatedly executed by the parameter update unit 33.

  In step S131, the parameter update unit 33 determines whether or not the rolling line 1 is under periodic repair. Specifically, the update timing designation unit 41 inquires of the controller 21 for control or the setting computer 23 whether the rolling line 1 is undergoing periodic repair. The update timing designation unit 41 determines whether or not the rolling line 1 is undergoing periodic repair based on the inquiry result. If it is under periodic repair, the process of step S132 is executed. If the regular repair is not in progress, the process of step S133 is executed.

  In step S132, the parameter update unit 33 selects all model parameters as update parameters. At the time of periodic repair, it is preferable to update all parameters inherent in the virtual rolling line setting calculation unit 32 to the same values as the latest parameters used in the actual operation setting calculator 23. Therefore, the update parameter selection unit 42 selects all parameters at the time of periodic repair.

  In step S133, the parameter update unit 33 determines whether or not a role change has been performed. Specifically, the update timing designation unit 41 inquires of the controller 21 for control or the setting computer 23 whether or not the operation mode of the rolling line 1 is the roll change mode. The update timing designation unit 41 determines whether or not the roll change mode is set based on the inquiry result. If it is the roll change mode, the process of step S134 is executed. If it is not the roll change mode, the process of step S135 is executed.

  In step S134, the parameter update unit 33 selects machine constants related to the roll and all learning terms as update parameters. Specifically, the update parameter selection unit 42 selects machine constants related to the roll and all learning terms as model parameters to be updated.

  In step S135, the parameter update unit 33 determines whether a simulation execution command has been issued. Specifically, the update timing command unit 41 of the parameter update unit 33 inquires of the simulation condition setting unit 31 whether the simulation condition is input and a simulation execution command is issued. The parameter update unit 33 determines whether or not a simulation execution command has been issued based on the inquiry result. If a simulation execution command is given, the process of step S136 is executed.

  In step S136, the parameter update unit 33 selects a learning term and an adjustment term related to the simulation condition as an update parameter before executing the simulation calculation. Specifically, the update parameter selection unit 42 selects learning terms and adjustment terms related to simulation conditions as model parameters to be updated.

  In step S137, the update timing designation unit 41 of the parameter update unit 33 determines whether the update of the parameter selected by the update parameter selection unit 42 does not affect the actual operation setting calculation. Specifically, the update timing designating unit 41 calculates the load applied to the setting computer 23 when acquiring the parameter selected by the update parameter selecting unit 42 from the setting computer 23 for actual operation. In addition, the update timing designation unit 41 confirms the load status of the setting computer 23. Based on these, the update timing designating unit 41 confirms that even if a load due to parameter acquisition occurs, it does not affect the actual operation setting calculation. If it is determined that the parameter update does not affect the actual operation setting calculation, the process of step S139 is executed. If it is determined that the parameter update affects the actual operation setting calculation, the process of step S138 is executed.

  In step S138, it is determined whether the number of executions of the determination process in step S137 is less than the upper limit number. If the determination condition is satisfied, the process of step S137 is re-executed after the specified time has elapsed. If the determination condition is not satisfied, the process of this routine is terminated.

  In step S139, the parameter copy unit 43 updates the model parameter selected by the update parameter selection unit 42 to the same value as the latest parameter used in the setting calculator 23 for actual operation.

  FIG. 14 is a flowchart showing one procedure for examining the alloy composition and manufacturing conditions using the rolling simulation device 24.

  First, in step S141, the parameter updating unit 33 updates all model parameters at the time of periodic repair. The processing in step S141 is the same as the processing in steps S131 and S132 in FIG.

  In step S142, the parameter update unit 33 updates the machine constant and all the learning terms related to the roll. The processing in step S142 is the same as the processing in steps S133 and S134 of FIG. 13 described above.

  In step S143, the user inputs simulation conditions using the input / output device of the rolling simulation device 24. Specifically, the user can obtain initial slab information (dimensions, alloy composition) of virtual metal used for the simulation, various target values (heating furnace extraction target temperature, finishing input target temperature, finishing output target temperature, winding target temperature). , Rough rolling delivery target plate thickness, rough rolling delivery target plate width, target plate width, target plate thickness, target crown ratio, target flatness, etc.) and more detailed conditions (for example, heating furnace slab temperature rising pattern, Enter the roughing-side target temperature, the rolling amount and rolling rate distribution of each pass, the plate feed speed and acceleration rate, the cooling pattern at the run-out table, various spray settings, finish rolling bender, work roll shift, etc., as necessary. . The input simulation condition is set in the simulation condition setting unit 31.

  In step S144, the parameter updating unit 33 updates learning terms and adjustment terms related to simulation conditions before executing simulation calculation. Specifically, the update parameter selection unit 42 selects learning terms and adjustment terms related to simulation conditions as model parameters to be updated. The parameter copy unit 43 updates the selected model parameter to the same value as the latest parameter used for actual operation.

  In step S145, the virtual rolling line setting calculation unit 32 sets the set value of each process and the size and position of the metal material at each time when the virtual metal material is rolled on the virtual rolling line based on the set simulation conditions. The temperature is calculated using a process model, a transfer model, and a temperature model. Further, the virtual rolling line setting calculation unit 32 predicts the final product material using the material model as input values of the processing history and temperature history information of the given virtual metal. Moreover, the metal structure change during virtual rolling of the virtual metal is predicted using the metal structure prediction model. Further, by using the mechanical property prediction model, the material such as the mechanical properties such as yield stress and tensile strength is predicted using the finally calculated structure and alloy composition of the virtual metal product as input values.

  In step S146, the user confirms the quality of the virtual metal product based on the calculation result in step S145. In step S147, the user confirms the setting value of each process. In step S148, the simulation conditions are changed as necessary, and the processes in steps S144 to S147 are repeated. In step S149, the user considers application of the simulation result to actual operation.

  According to such a procedure, an accurate simulation simulating actual operation rolling is possible without affecting actual operation rolling. In addition, the simulation conditions are changed, the above simulation is repeatedly performed, and the results are analyzed to obtain guidelines for improving the heating, rolling and cooling conditions and slab alloy composition in actual operation.

DESCRIPTION OF SYMBOLS 1 Rolling line 10 Conveying table 11 Heating furnace 12 Rough rolling mill 13 Bar heater 14 Finishing mill entry side thermometer 15 Finishing rolling mill 16 Finishing rolling mill exit side thermometer 17 Runout table 18 Winding machine entrance side thermometer 19 Winding machine DESCRIPTION OF SYMBOLS 20 Rolling system 21 Control controller 23 Setting computer 24 Rolling simulation apparatus 25 Host computer 31 Simulation condition setting part 32 Virtual rolling line setting calculation part 33 Parameter update part 41 Update timing designation part 42 Update parameter selection part 43 Parameter copy part

Claims (3)

A rolling line having an actuator group that heats, rolls, cools, and conveys the metal material, a control record value of the actuator group, and a sensor group that detects a record value of the state of the metal material; A rolling simulation device connected to a rolling system comprising a setting computer for calculating a predicted state value of the metal material,
The setting calculator is
A model formula expressing physical phenomena of each process of heating, rolling, cooling, and conveyance in the rolling line, and having a first model formula expressed by a function having an input variable and a model parameter group as inputs,
Using the first model formula, in order to achieve process conditions related to product quality and operating conditions in actual operation, calculate the control target value of the actuator group and the state prediction value of the metal material,
The model parameter group of the first model formula is updated as needed based on a comparison value obtained by comparing the control target value and the predicted state value with the actual control value and the actual state value detected by the sensor group. ,
The rolling simulation apparatus is
A simulation condition setting unit that sets simulation conditions related to product quality and operation conditions in virtual operation in which the virtual metal material is heated, rolled, cooled, and transported in the rolling line;
The second model formula is the same as the first model formula, and the control target value of the actuator group and the predicted state value of the virtual metal material are used to achieve the simulation condition using the second model formula. A virtual rolling line setting calculation unit for calculating
A parameter updating unit that updates the model parameter group of the second model formula based on the model parameter group of the first model formula when the model parameter group of the first model formula is updated;
A rolling simulation apparatus comprising: The parameter update unit
An update timing designating unit for designating a timing at which calculation is not executed by the setting computer in actual operation;
A parameter copy unit for copying the model parameter group of the first model formula to the model parameter group of the second model formula at the timing;
The rolling simulation apparatus according to claim 1, comprising: The parameter update unit
Of the model parameter group of the second model formula, an update parameter selection unit that selects a part of the model parameter group necessary for model calculation using the simulation condition in the virtual rolling line setting calculation unit,
A parameter copy unit that copies only the partial model parameter group selected by the update parameter selection unit from the model parameter group of the first model formula;
The rolling simulation apparatus according to claim 1, comprising:

Images