CN115815345A - Mechanism collaborative prediction method and system for predicting mechanical property of full-flow hot-rolled strip steel - Google Patents
Mechanism collaborative prediction method and system for predicting mechanical property of full-flow hot-rolled strip steel Download PDFInfo
- Publication number
- CN115815345A CN115815345A CN202211434900.8A CN202211434900A CN115815345A CN 115815345 A CN115815345 A CN 115815345A CN 202211434900 A CN202211434900 A CN 202211434900A CN 115815345 A CN115815345 A CN 115815345A
- Authority
- CN
- China
- Prior art keywords
- rolling
- mechanical property
- parameters
- cooling
- strip steel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention discloses a mechanism collaborative prediction method and a mechanism collaborative prediction system for predicting mechanical property of a full-process hot-rolled strip steel, which belong to the technical field of metallurgical production and comprise the following steps: s1, acquiring process parameters of all working conditions in real time, online forecasting grain sizes and phase changes of all racks, and simultaneously performing big data training of a neural network; s2, based on comparison of actual working conditions and a sample data set, correcting the corresponding relation between rolling force, rolling speed, strain rate, rolling temperature and deformation resistance by using a rolling mechanism model, and improving the predicted value of mechanical property; s3, forecasting the mechanical property of the cooling section after rolling according to the cooling condition after rolling by taking the rolling mechanical property forecast value as an initial condition; and S4, identifying and clustering the rolling parameters, the cooling parameters and the corresponding mechanical properties of the whole process by combining big data intelligent training of field measured data and real-time forecasting of a mechanism model. The invention can carry out intelligent collaborative prediction on the mechanical property mechanism of the whole process of rolling control and cooling control.
Description
Technical Field
The invention belongs to the technical field of metallurgical production, and particularly relates to a mechanism collaborative prediction method and system for predicting mechanical property of a full-flow hot-rolled strip steel.
Background
The mechanical property is one of key technical indexes of the hot rolled strip steel and depends on the deformation state of finish rolling and the change of a cooling structure after rolling, so that the online prediction of the mechanical property is a core technology for improving the hot rolling production efficiency and the quality of a final product. In the prior production process, the mechanical properties of the hot-rolled strip steel are usually obtained by adopting a mode of off-line detection of stress strain or hardness and elongation, and then the mechanical properties are fed back to a production line for further optimization and adjustment of a subsequent rolling process. However, the traditional detection method has low efficiency and poor precision, cannot accurately predict the performance change of the strip steel of each stand in the rolling process, belongs to an indirect manual feedback regulation and control mode, and leads the rolling process to depend on more operation experience. Although high-quality strip steel meeting the quality requirement can be produced through long-term process grope, the online prediction of the mechanical performance is imperative for the automatic adjustment or the high-efficiency fine control of a production line. Based on the purposes, the mechanical property of the hot-rolled strip steel in the whole process is required to be forecasted by combining a deformation mechanism model and a big data neural network intelligent model around a rolling process and a cooling system after rolling, so that the rolling control and cooling control precision and the production efficiency of a hot-rolling production line are improved.
Disclosure of Invention
The invention provides a mechanism collaborative prediction method and a mechanism collaborative prediction system for predicting the mechanical property of a full-flow hot-rolled strip steel, aiming at synchronously considering the tissue deformation process of a finish rolling section and the temperature drop process of a cooling section after rolling, carrying out intelligent collaborative prediction on the mechanical property mechanism of the full-flow controlled rolling and controlled cooling process, providing high-precision deformation resistance for a rolling model and providing an online prediction model for the mechanical property of cooling after rolling, thereby realizing the fine control of the full flow of the hot-rolled strip steel.
The invention aims to provide a mechanism collaborative prediction method for predicting the mechanical property of a full-flow hot-rolled strip steel, which comprises the following steps:
s1, acquiring process parameters of all working conditions in real time, online forecasting grain sizes and phase changes of all racks, and simultaneously performing big data training of a neural network;
s2, based on comparison of actual working conditions and a sample data set, correcting the corresponding relation between rolling force, rolling speed, strain rate, rolling temperature and deformation resistance by using a rolling mechanism model, and improving the predicted value of mechanical property;
s3, taking the rolling mechanical property prediction value as an initial condition, predicting the mechanical property of a cooling section after rolling according to the cooling condition after rolling, observing the change of the structure property in the cooling control process in real time, and realizing the online prediction of the mechanical property of the hot-rolled strip steel after final cooling;
and S4, carrying out real-time identification clustering on the rolling parameters, the cooling parameters and the corresponding mechanical properties of the whole process by combining the big data intelligent training result of the field measured data and the high-precision online forecasting of the mechanism model, and simultaneously forming standard sample data so as to meet the subsequent high-precision online forecasting requirement.
Preferably, S1 comprises:
s101, collecting current working condition parameters in real time, comparing sample data according to strip steel components, rolling force, rolling speed, rolling temperature and strain rate, and forecasting grain size and phase change;
s102, calculating transient mechanical property change on line according to the grain size and the predicted value of phase change, so as to obtain deformation resistance corresponding to the mechanical property of the strip steel from frame to frame;
s103, training a neural network model of the current data to form a sample data set.
Preferably, S2 comprises:
s201, obtaining a corresponding relation between rolling parameters and deformation resistance of corresponding working conditions by using clustering training data of a neural network, and providing accurate boundary conditions and initial conditions for a mechanism model;
s202, based on measured data after neural network training, high-precision online prediction of deformation resistance is carried out by utilizing a rolling mechanism model, and a deformation resistance prediction value which accords with the actual situation is obtained through iteration.
Preferably, S3 comprises:
s301, tracking rolling process parameters and mechanical property forecast values in real time;
s301, calculating a heat exchange coefficient and a transient temperature on line according to the actually measured cooling rate, cooling time and cooling water quantity;
s301, based on the parameters, the organizational performance is rapidly forecasted, and the mechanical performance change rule of the whole process is obtained.
Preferably, the process parameters of the full operating condition include:
acquiring data of original chemical components in a smelting part of a steel billet, wherein the data comprises content data of C, si, P and S components in steel;
the production process parameters of the hot-rolled strip steel comprise: the outlet temperature of the billet heating furnace, the inlet temperature and the outlet temperature of the roughing mill, the reduction of each pass, the rolling speed and the outlet thickness, the inlet temperature and the outlet temperature of the finishing mill, the rolling speed of each pass, the reduction and the outlet thickness, the coiling temperature, the coiling speed, the coiling thickness and the like;
and combining the grain size and deformation resistance parameters obtained by product analysis to form a sample data set.
Preferably, screening the data in the sample data set to remove the measurement error caused by the measurement problem and the abnormal data caused by the production problem; the screening comprises the following steps: and setting fluctuation ranges for various parameters, and removing the data exceeding the fluctuation ranges integrally.
Preferably, the rolling mechanism model includes: the recrystallization model and the rheological stress model are used for calculating the microscopic change of the hot-rolled strip steel, and the mechanical property is forecasted by taking the grain size, the recrystallization rate and the rheological stress parameters as input conditions for mechanical property forecast; each rolling stage comprises: a furnace discharging stage of a heating furnace, each pass of rough rolling, each pass of finish rolling and a layer cooling stage.
The second purpose of the invention is to provide a mechanism collaborative prediction system for predicting the mechanical property of the full-flow hot-rolled strip steel, which comprises the following steps:
a training module: acquiring process parameters of all working conditions in real time, online forecasting grain size and phase change of each rack, and simultaneously performing big data training of a neural network;
the relation establishment module: based on the comparison between the actual working condition and the sample data set, the corresponding relation between the rolling force, the rolling speed, the strain rate, the rolling temperature and the deformation resistance is corrected by utilizing a rolling mechanism model, and the predicted value of the mechanical property is improved;
an analysis module: taking the rolling mechanical property prediction value as an initial condition, predicting the mechanical property of a cooling section after rolling according to the cooling condition after rolling, observing the change of the structure property in the cooling control process in real time, and realizing the online prediction of the mechanical property of the hot-rolled strip steel after final cooling;
a standard sample generation module: and (3) combining big data intelligent training of field measured data and real-time prediction of a mechanism model, identifying and clustering rolling parameters, cooling parameters and corresponding mechanical properties of the whole process to form standard sample data, and meeting the subsequent on-line prediction requirement.
The third purpose of the invention is to provide an information data processing terminal for realizing the mechanism collaborative prediction method for predicting the mechanical property of the full-flow hot-rolled strip steel.
A fourth object of the present invention is to provide a computer-readable storage medium, which includes instructions that, when executed on a computer, cause the computer to execute the above mechanistic co-prediction method for predicting the mechanical property of a full-flow hot-rolled strip.
The invention has the advantages and positive effects that:
the invention uses a mechanism model to forecast the grain size and the phase change in real time, and simultaneously uses an intelligent algorithm to carry out the corresponding relation between big data training process parameters and tissue changes, thereby obtaining an online mechanical property forecast value meeting the actual production needs and better conforming to the actual value. On the basis, a real-time forecasting system with full-frame mechanism cooperation is constructed, so that the whole-process tracking and control of the mechanical property of the hot-rolled strip steel can be carried out, the real-time forecasting system can feed back the mechanical property of the hot-rolled strip steel to a rolling mill system in real time, and the improvement of the forecasting precision of the rolling force, the online optimization of the rolling process and the accurate control of the mechanical property of the rolled strip steel are facilitated. Therefore, the full-process hot-rolled strip steel mechanical property collaborative prediction method is very beneficial to improving the rolling efficiency, the cooling efficiency and the strip steel product performance.
According to the method, the continuity characteristic of the whole process of a production line is considered, parameters of all process sections are coupled and inherited in real time, a neural network is used for big data training, and a theoretical mechanism model is used for collaborative prediction of mechanical property; the cooperative mode can not only utilize the optimal topological structure of the neural network to perform cluster analysis and improve the sample training of complex parameters, but also utilize a mechanism model to accurately forecast the grain size and the phase change condition of the current production working condition, thereby meeting the real-time forecasting of the mechanical property of the whole production line, providing regulation and control targets for a rolling process and a cooling process, being beneficial to realizing the closed-loop regulation and control of the mechanical property, and improving the intelligent level and the forecasting precision of the production line.
Drawings
FIG. 1 is a diagram showing the essential synergy between the finish rolling process and the post-rolling cooling process in the preferred embodiment of the present invention;
FIG. 2 is a flow chart of the extraction of a microtissue sample in a preferred embodiment of the present invention;
FIG. 3 is a flow chart of the full condition prediction of the mechanical properties of hot rolled strip steel;
fig. 4 is a basic training flow chart of the neural network for controlling rolling and cooling mechanical properties.
Detailed Description
In order to further understand the contents, features and effects of the present invention, the following embodiments are illustrated and described in detail with reference to the accompanying drawings:
the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments obtained by those skilled in the art without creative efforts based on the technical solutions of the present invention belong to the protection scope of the present invention.
Please refer to fig. 1 to 4.
A mechanism collaborative prediction method for predicting the mechanical property of a full-flow hot-rolled strip steel is used for the synchronous mechanical property prediction of hot-rolling controlled cooling to obtain the high-precision on-line control of the finish rolling and cooling after rolling of the hot-rolled strip steel; the method comprises the following steps:
s1, acquiring process parameters of all working conditions in real time, online forecasting grain sizes and phase changes of all racks, and simultaneously performing big data training of a neural network; s1 comprises the following steps:
s101, collecting current working condition parameters in real time, comparing sample data according to strip steel components, rolling force, rolling speed, rolling temperature and strain rate, and forecasting grain size and phase change;
s102, calculating transient mechanical property change on line according to the grain size and the predicted value of phase change, and thus obtaining deformation resistance corresponding to the mechanical property of each rack;
s103, training a neural network model of the current data to form a sample data set.
S1, sample data preparation, namely acquiring parameters of a full-process rolling process based on the current rolling condition, actually measuring parameters of mechanical properties and grain sizes of products, predicting the grain sizes and corresponding deformation resistance of the grain sizes on a rack-by-rack basis, and determining basic parameters of a model through neural network model training to form a sample data set;
the full-process rolling process parameters comprise: original chemical components and content of steel, production process parameters of hot rolled strip steel and mechanical property parameters. The method comprises the following steps of collecting data of original chemical components in a smelting part of a steel billet, wherein the data mainly comprises content data of important components such as C, si, P, S and the like in the steel; the production process parameters of the hot-rolled strip steel are acquired from a production operation table of the strip steel, and the data acquisition method mainly comprises the following steps: the method comprises the following steps of (1) discharging temperature of a billet heating furnace, inlet temperature of a roughing mill, outlet temperature of the roughing mill, reduction of each pass of the roughing mill, rolling speed of each pass of the roughing mill, inlet temperature of a finishing mill, outlet temperature of the finishing mill, rolling speed of each pass of the finishing mill, reduction of each pass of the finishing mill, curling temperature, curling speed, original thickness of a billet, outlet thickness of each pass of a roughing mill group, outlet thickness of each pass of a finishing mill group and the like; the mechanical property parameters are acquired through sampling and experiment of the product, and mainly comprise: tensile strength, yield strength, section elongation and the like. And after all data are acquired, sorting according to the steel coil numbers of the sampled products to form data files for data screening.
The screening work of the data mainly aims to remove measurement errors caused by measurement problems and abnormal data caused by production problems; the main work includes: the fluctuation range is set for various parameters, and data exceeding the fluctuation range are removed integrally, so that subsequent model training and theoretical model use are facilitated, the calculation precision of the model is improved, and the prediction accuracy of the model is improved.
The grain size and deformation resistance prediction of each frame are calculated through a basic theoretical model, and the theoretical model mainly comprises the following components: recrystallization model and rheological stress model, etc. The method is mainly used for calculating the microscopic change of the hot-rolled strip steel, and the mechanical property forecast is carried out by taking parameters such as grain size, recrystallization rate and rheological stress as input conditions of the mechanical property forecast.
In the process of rolling the plate blank, the recrystallization process can occur under different deformation conditions in the processes of rough rolling, finish rolling and the like. The dynamic recrystallization is classified according to the critical strain at which dynamic recrystallization occurs, the recrystallization process can be classified into dynamic recrystallization and static recrystallization, and the calculation formula for calculating the recrystallization rate and the crystal grain size of the dynamic recrystallization and the static recrystallization is shown in the following formulas (1) and (2):
where m is a temperature-dependent variable,. Epsilon c Is the critical strain, ε, at which dynamic recrystallization occurs s Relating to deformation temperature, strain rate and initial grain size, D is a constant relating to dynamic recrystallization activation energy, Z is a Zener-Hollomen parameter, n is a constant relating to steel type, t 0.5 The time required for the recrystallization rate to reach 50%, d 0 Is the original grain size.
In addition to the calculation of the grain size, the calculation of the rheological stress during recrystallization was also carried out, and the calculation formula is shown in the following formula (3):
σ＝σ ss -(σ ss -σ ds )·{1-exp[-k(ε-ε p ) n ]} (3)
wherein σ ss Is the saturation stress, σ ds Is the steady state stress, ε p Is the peak strain, and n and k are constants related to strain rate, gas constant and absolute temperature.
In the hot rolling process, the calculation formula of dislocation density in the deformation and recrystallization processes is shown as formulas (4) and (5):
ρ S ＝(1-X S )ρ+X S ρ 0 (5)
where, σ and σ X Yield stress, M is the Taylor factor, alpha is a constant, mu and b are the shear and Berth moduli, respectively, f ρ Is the dislocation density factor, p 0 The initial dislocation density.
During rolling, certain strains may exist, and when the grain size of each pass is calculated, the residual strain delta epsilon of the previous pass is added to the strain epsilon of a single pass, and the calculation formula is as follows:
the recrystallization type, the recrystallization grain size and the recrystallization rate of austenite grains in each pass in the finish rolling and rough rolling processes can be respectively calculated through theoretical calculation according to strip steel rolling technological parameters acquired in the previous data acquisition work, the recrystallized grains can generate phase change in the laminar cooling process, and the austenite grains are converted into ferrite, pearlite and the like under the influence of the cooling rate and the cooling condition. The process of the phase change is not easy to calculate by using a theoretical model, but in order to ensure the accuracy of calculation of the mechanical performance, the mechanical performance is further forecasted by using a neural network intelligent algorithm.
S2, based on comparison of actual working conditions and a sample data set, correcting the corresponding relation between rolling force, rolling speed, strain rate, rolling temperature and deformation resistance by using a rolling mechanism model, and improving the predicted value of mechanical property; s2 comprises the following steps:
s201, obtaining a corresponding relation between rolling parameters and deformation resistance of corresponding working conditions by using clustering training data of a neural network, and providing boundary conditions and initial conditions for a mechanism model;
s202, based on measured data after neural network training, online prediction of deformation resistance is carried out by utilizing a rolling mechanism model, and a high-precision deformation resistance prediction value is obtained through iteration.
S2, reversely calculating deformation resistance by using a rolling mechanism model, and simultaneously performing big data training on the characteristic parameters by using a neural network to obtain a cooperative topological structure of forecast parameters and actual measurement parameters of the mechanism model;
the training and using work of the neural network model mainly comprises the following steps: sample data sorting and classifying, training and verifying of a neural network model and actual use of the neural network model.
Extracting data (outlet temperature of finish rolling, curling temperature, curling rate and the like) and mechanical property data (tensile strength, yield strength, elongation and the like) of a laminar cooling section from the acquired data, corresponding the data to the recrystallization rate and the grain size of each rolling pass calculated by a theoretical model, combining the data into a new data set, and randomly dividing the obtained data into a training set, a testing set and a verification set for training and verifying a neural network model.
The method comprises the steps of obtaining the raw material components of the strip steel, the finish rolling and rough rolling data, the curling and laminar cooling data and the mechanical property parameters, wherein the quantity of the data is large, the data is further processed through the data calculation of a mechanism model, and the finish rolling and rough rolling data are processed into the grain size, the recrystallization rate and the rheological stress of each pass for the training and verification of a neural network model.
Before training a neural network model, data is required to be screened, in the invention, laplacian scores are used for judging the correlation, and the calculation formula is as follows:
where cov is covariance and var is variance.
And respectively calculating the correlation between the reintegrated data and the mechanical property, and respectively selecting parameters with higher correlation as input parameters of the neural network model according to different required and forecasted mechanical property parameters (tensile strength, yield strength and elongation).
After the training set is used for completing training, the testing set is used for testing and further correcting the forecasting effect of the model, the verification set is used for testing the actual forecasting effect of the model, and finally a set of model capable of forecasting the mechanical property of the hot-rolled strip steel is formed and is used in the field actual production process.
S3, taking the rolling mechanical property forecast value as an initial condition, forecasting the mechanical property of a cooling section after rolling according to the cooling condition after rolling, observing the change of the structure property in the cooling control process in real time, and realizing the online forecasting of the mechanical property of the hot-rolled strip steel after final cooling; s3 comprises the following steps:
s301, tracking rolling process parameters and mechanical property forecast values in real time;
s301, calculating a heat exchange coefficient and a transient temperature on line according to the actually measured cooling rate, cooling time and cooling water quantity;
s301, based on the parameters, the organizational performance is rapidly forecasted, and the mechanical performance after final cooling is obtained.
S3, combining a neural network model and a theoretical calculation model to carry out high-precision prediction on each rack, taking the prediction as an initial condition of cooling after rolling, and carrying out mechanical property prediction in a controlled cooling process after rolling on the basis of synchronously considering the temperature and the cooling rate after rolling;
after the training of the neural network model is completed, the mechanical property prediction and the grain size calculation of the strip steel of each process section of the whole rolling process can be realized by matching with a theoretical calculation model, the process optimization can be carried out on the rolling process by combining the actual production process change, and meanwhile, the cooling control process after the rolling is completed is carried out on the basis of comprehensively considering the whole process, and the mechanical property prediction is carried out.
And S4, combining big data intelligent training of the field measured data and real-time forecasting of the mechanism model, identifying and clustering the rolling parameters, the cooling parameters and the corresponding mechanical properties of the whole process to form standard sample data, and meeting the subsequent on-line forecasting requirement. The method is compared and verified with offline measured data, big data training of all working conditions is carried out, finish rolling and cooling process parameters after rolling are optimized integrally, and the requirement of online prediction of all-working-condition mechanical properties of the hot-rolled strip steel is met.
In the above embodiment:
step one, sample data preparation, namely acquiring parameters of a full-process rolling process based on the current rolling condition, actually measuring parameters of mechanical properties and grain sizes of products, predicting the grain sizes and corresponding deformation resistance of the grain sizes on a rack-by-rack basis, and determining basic parameters of a model through neural network model training to form a sample data set;
step two, reversely calculating deformation resistance by using a rolling mechanism model, and simultaneously performing big data training on the characteristic parameters by using a neural network to obtain a cooperative topological structure of forecast parameters and actual measurement parameters of the mechanism model;
thirdly, combining the neural network model and the theoretical calculation model to carry out high-precision prediction on each rack, taking the prediction as an initial condition of cooling after rolling, and carrying out mechanical property prediction in the cooling control process after rolling on the basis of synchronously considering the temperature and the cooling rate after rolling;
and step four, comparing and verifying the data with offline measured data, performing big data training under all working conditions, integrally optimizing finish rolling and cooling process parameters after rolling, and meeting the requirement of online prediction of all-working-condition mechanical properties of the hot-rolled strip steel.
A mechanism collaborative prediction system for predicting the mechanical property of a full-flow hot-rolled strip steel comprises: the grain size and deformation resistance change of the strip steel in each rolling stage calculated by the basic theoretical model and the deformation resistance and grain size condition of the product predicted by the neural network model are convenient for monitoring the quality change of the product in real time in the rolling production process and dynamically adjusting the rolling process; the method comprises the following steps:
a training module: acquiring process parameters of all working conditions in real time, online forecasting grain size and phase change of each rack, and simultaneously performing big data training of a neural network;
a relationship establishing module: based on the comparison between the actual working condition and the sample data set, the corresponding relation between the rolling force, the rolling speed, the strain rate, the rolling temperature and the deformation resistance is corrected by utilizing a rolling mechanism model, and the predicted value of the mechanical property is improved;
a forecasting module: taking the rolling mechanical property prediction value as an initial condition, predicting the mechanical property of a cooling section after rolling according to the cooling condition after rolling, observing the change of the structure property in the cooling control process in real time, and realizing the online prediction of the mechanical property of the hot-rolled strip steel after final cooling;
a standard sample generation module: and (3) combining big data intelligent training of field measured data and real-time prediction of a mechanism model, identifying and clustering rolling parameters, cooling parameters and corresponding mechanical properties of the whole process to form standard sample data, and meeting the subsequent on-line prediction requirement.
An information data processing terminal is used for realizing the mechanism collaborative forecasting method for forecasting the mechanical property of the full-flow hot-rolled strip steel.
A computer-readable storage medium comprising instructions which, when executed on a computer, cause the computer to perform the above mechanistic co-prediction method for predicting mechanical properties of a full-run hot rolled strip.
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When used in whole or in part, can be implemented in a computer program product that includes one or more computer instructions. When loaded or executed on a computer, cause the flow or functions according to embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, the computer instructions may be transmitted from one website site, computer, server, or data center to another website site, computer, server, or data center via wire (e.g., coaxial cable, fiber optic, digital Subscriber Line (DSL), or wireless (e.g., infrared, wireless, microwave, etc.)). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that includes one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid State Disk (SSD)), among others.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the present invention in any way, and all simple modifications, equivalent changes and modifications made to the above embodiment according to the technical spirit of the present invention are within the scope of the technical solution of the present invention.
Claims (8)
1. A mechanism collaborative prediction method for predicting the mechanical property of a full-flow hot-rolled strip steel is characterized by comprising the following steps:
s1, acquiring process parameters of all working conditions in real time, online forecasting grain sizes and phase changes of all racks, and simultaneously performing big data training of a neural network;
s2, based on comparison of actual working conditions and a sample data set, correcting the corresponding relation between rolling force, rolling speed, strain rate, rolling temperature and deformation resistance by using a rolling mechanism model, and improving the predicted value of mechanical property;
s3, taking the rolling mechanical property prediction value as an initial condition, predicting the mechanical property of a cooling section after rolling according to the cooling condition after rolling, observing the change of the structure property in the cooling control process in real time, and realizing the online prediction of the mechanical property of the hot-rolled strip steel after final cooling;
and S4, combining the big data intelligent training result of the field measured data and the online forecast of the mechanism model, carrying out real-time identification clustering on the rolling parameters, the cooling parameters and the corresponding mechanical properties of the whole process, and simultaneously forming standard sample data so as to meet the subsequent online forecast requirement.
2. The mechanistic collaborative prediction method for predicting mechanical properties of full-flow hot-rolled strip steel according to claim 1, wherein S1 comprises:
s101, collecting current working condition parameters in real time, comparing sample data according to strip steel components, rolling force, rolling speed, rolling temperature and strain rate, and forecasting grain size and phase change;
s102, calculating transient mechanical property change on line according to the grain size and the predicted value of phase change, so as to obtain deformation resistance corresponding to the mechanical property of the strip steel from frame to frame;
s103, training a neural network model of the current data to form a sample data set.
3. The mechanism collaborative prediction method for predicting the mechanical property of the full-flow hot-rolled strip steel according to claim 1, wherein S2 includes:
s201, obtaining a corresponding relation between rolling parameters and deformation resistance of corresponding working conditions by using clustering training data of a neural network, and providing boundary conditions and initial conditions for a mechanism model;
s202, based on the measured data after the neural network training, the rolling mechanism model is utilized to carry out online prediction on the deformation resistance, and a deformation resistance prediction value which accords with the actual situation is obtained through iteration.
4. The mechanistic collaborative prediction method for predicting mechanical properties of full-flow hot-rolled strip steel according to claim 1, wherein S3 comprises:
s301, tracking rolling process parameters and mechanical property forecast values in real time;
s301, calculating a heat exchange coefficient and a transient temperature on line according to the actually measured cooling rate, cooling time and cooling water quantity;
s301, based on the parameters, the organizational performance is rapidly forecasted, and the mechanical performance change rule of the whole process is obtained.
5. The mechanistic collaborative forecasting method for predicting mechanical property of full-flow hot-rolled strip steel according to claim 1, wherein the process parameters of the full-working condition comprise:
acquiring data of original chemical components in a smelting part of a steel billet, wherein the data comprises content data of C, si, P and S components in steel;
the production process parameters of the hot-rolled strip steel comprise: the tapping temperature of the billet heating furnace, the inlet temperature and the outlet temperature of the roughing mill, the rolling reduction, the rolling speed and the outlet thickness of each pass, the inlet temperature and the outlet temperature of the finishing mill, the rolling speed, the rolling reduction and the outlet thickness of each pass, and the coiling temperature, speed and thickness of each pass;
and forming a sample data set by combining the grain size and deformation resistance parameters obtained by product analysis.
6. The mechanistic collaborative forecasting method for predicting the mechanical property of the full-flow hot-rolled strip steel according to claim 2, characterized in that data in the sample data set are screened to remove measurement errors caused by measurement problems and abnormal data caused by production problems; the screening comprises the following steps: and setting fluctuation ranges for various parameters, and removing the data exceeding the fluctuation ranges integrally.
7. The mechanistic collaborative forecasting method for predicting mechanical property of full-flow hot-rolled strip steel according to claim 1, wherein the rolling mechanism model comprises: the recrystallization model and the rheological stress model are used for calculating the microscopic change of the hot-rolled strip steel, and the mechanical property is forecasted by taking the grain size, the recrystallization rate and the rheological stress parameters as input conditions for mechanical property forecast; each rolling stage comprises: a furnace discharging stage of the heating furnace, each pass of rough rolling, each pass of finish rolling and a layer cooling stage.
8. A mechanism collaborative prediction system for predicting the mechanical property of a full-flow hot-rolled strip steel is characterized by comprising the following steps:
a training module: acquiring process parameters of all working conditions in real time, online forecasting grain size and phase change of each rack, and simultaneously performing big data training of a neural network;
an analysis module: based on the comparison between the actual working condition and the sample data set, the corresponding relation between the rolling force, the rolling speed, the strain rate, the rolling temperature and the deformation resistance is corrected by utilizing a rolling mechanism model, and the predicted value of the mechanical property is improved;
a forecasting module: taking the rolling mechanical property prediction value as an initial condition, predicting the mechanical property of a cooling section after rolling according to the cooling condition after rolling, observing the change of the structure property in the cooling control process in real time, and realizing the online prediction of the mechanical property of the hot-rolled strip steel after final cooling;
a standard sample generation module: and (3) combining big data intelligent training of field measured data and real-time prediction of a mechanism model, identifying and clustering rolling parameters, cooling parameters and corresponding mechanical properties of the whole process to form standard sample data, and meeting the subsequent on-line prediction requirement.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211434900.8A CN115815345A (en) | 2022-11-16 | 2022-11-16 | Mechanism collaborative prediction method and system for predicting mechanical property of full-flow hot-rolled strip steel |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211434900.8A CN115815345A (en) | 2022-11-16 | 2022-11-16 | Mechanism collaborative prediction method and system for predicting mechanical property of full-flow hot-rolled strip steel |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115815345A true CN115815345A (en) | 2023-03-21 |
Family
ID=85528508
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211434900.8A Pending CN115815345A (en) | 2022-11-16 | 2022-11-16 | Mechanism collaborative prediction method and system for predicting mechanical property of full-flow hot-rolled strip steel |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115815345A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116000106A (en) * | 2023-03-27 | 2023-04-25 | 东北大学 | Rolling force setting method in cold continuous rolling speed increasing and decreasing stage |
| CN117787507A (en) * | 2024-02-23 | 2024-03-29 | 宝鸡核力材料科技有限公司 | Full chain optimizing method and device for tape rolling process |
| CN117807424A (en) * | 2024-02-29 | 2024-04-02 | 山东钢铁股份有限公司 | Industrial big data driven wide and thick steel plate quality dynamic on-line identification method and device |
| CN117807424B (en) * | 2024-02-29 | 2024-04-30 | 山东钢铁股份有限公司 | Industrial big data driven wide and thick steel plate quality dynamic on-line identification method and device |
-
2022
- 2022-11-16 CN CN202211434900.8A patent/CN115815345A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116000106A (en) * | 2023-03-27 | 2023-04-25 | 东北大学 | Rolling force setting method in cold continuous rolling speed increasing and decreasing stage |
| CN117787507A (en) * | 2024-02-23 | 2024-03-29 | 宝鸡核力材料科技有限公司 | Full chain optimizing method and device for tape rolling process |
| CN117787507B (en) * | 2024-02-23 | 2024-05-03 | 宝鸡核力材料科技有限公司 | Full chain optimizing method and device for tape rolling process |
| CN117807424A (en) * | 2024-02-29 | 2024-04-02 | 山东钢铁股份有限公司 | Industrial big data driven wide and thick steel plate quality dynamic on-line identification method and device |
| CN117807424B (en) * | 2024-02-29 | 2024-04-30 | 山东钢铁股份有限公司 | Industrial big data driven wide and thick steel plate quality dynamic on-line identification method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113333474B (en) | Strip steel hot-rolled strip shape control method and system based on digital twinning | |
| CN100362332C (en) | Method for online test of steel plate mechanic property during rolling process | |
| JP5003483B2 (en) | Material prediction and material control device for rolling line | |
| CN111008477B (en) | Method for adjusting technological parameters based on mechanical properties of cold-rolled galvanized strip steel | |
| CN106345823B (en) | The method of online real-time estimate mechanical performance based on coils of hot-rolled steel production procedure | |
| CN111209967A (en) | Rolling process plate convexity prediction method based on support vector machine | |
| CN112036081B (en) | Method for determining addition amount of silicon-manganese alloy in converter tapping based on yield prediction | |
| CN111563686A (en) | Cold-rolled silicon steel quality judgment method based on full-process data | |
| CN114818456B (en) | Prediction method and optimization method for full-length deformation resistance of cold continuous rolling strip steel | |
| CN110434172B (en) | Load distribution calculation method for continuous rolling of furnace coil and finishing mill group | |
| CN115815345A (en) | Mechanism collaborative prediction method and system for predicting mechanical property of full-flow hot-rolled strip steel | |
| CN112287550B (en) | Strip steel head thickness difference process parameter optimization method based on principal component analysis controller | |
| CN104942019A (en) | Automatic control method for width of steel strips during cold rolling | |
| Ma et al. | Influence of profile indicators of hot-rolled strip on transverse thickness difference of cold-rolled silicon steel | |
| RU2752518C1 (en) | Method for operating the annealing furnace | |
| He et al. | Whole process prediction model of silicon steel strip on transverse thickness difference based on Takagi-Sugeno fuzzy network | |
| CN109593951B (en) | Furnace temperature-based hot dip product dezincification defect dynamic control method | |
| CN111482466B (en) | Method for setting acceleration of rolling mill | |
| CN116108932A (en) | Method for establishing fusion model of steel production process data and mechanism | |
| CN107486587B (en) | Thinning compensation method for improving control precision of shearing setting model | |
| KR100841888B1 (en) | Rolling line material quality prediction and control apparatus | |
| CN115392104A (en) | Method for predicting mechanical property of cold-rolled continuous annealing strip steel based on annealing process | |
| CN117831659B (en) | Method and device for online detection of quality of wide and thick plates, electronic equipment and storage medium | |
| JP2002236119A (en) | Material estimating device for steel product | |
| KR20040057181A (en) | An method for controling the target heating temperature of strip using neural net |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |