Mechanism synergy prediction method and system for predicting mechanical properties of full-process hot-rolled strip steel

Abstract

The application discloses a mechanism cooperative prediction method and system for predicting mechanical properties of full-process hot-rolled strip steel, and belongs to the technical field of metallurgical production, which comprises the following steps: S1, real-time acquisition of process parameters under full working conditions, online prediction of grain size and phase change of each rack, and big data training of a neural network; S2, comparison based on actual working conditions and sample data sets, correction of the corresponding relationship between rolling force, rolling speed, strain rate, rolling temperature and deformation resistance by using a rolling mechanism model, and improvement of the predicted value of mechanical properties; S3, taking the rolling mechanical property prediction value as an initial condition, and then performing mechanical property prediction of the post-rolling cooling section according to post-rolling cooling conditions; and S4, identification and clustering of rolling parameters, cooling parameters and corresponding mechanical properties by combining big data intelligent training of field measured data and real-time prediction of the mechanism model. The application can perform intelligent cooperative prediction of the mechanical properties mechanism in the whole process of controlled rolling and controlled cooling.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of metallurgical production, and particularly relates to a mechanism cooperative prediction method and system for predicting mechanical properties of hot-rolled strip steel in a whole process. BACKGROUND

[0002] Mechanical properties are one of the key technical indexes of hot-rolled strip steel, which depend on the deformation state of finish rolling and the microstructure change after rolling, so online prediction of the mechanical properties is a core technology for improving the production efficiency of hot rolling and the quality of final products. In the past production process, the mechanical properties of hot-rolled strip steel are usually obtained by offline detection of stress, strain or hardness, elongation, and then fed back to the production line to further optimize and adjust the subsequent rolling process. However, this traditional detection method is low in efficiency and poor in accuracy, and cannot accurately predict the performance change of the strip steel in each stand during rolling, which is an indirect manual feedback control method, which will lead to more reliance on operating experience in the rolling process. Although high-quality strip steel meeting the quality requirements can be produced through long-term process exploration, online prediction of mechanical properties is imperative for automatic adjustment or efficient and refined control of the production line. Based on the above purpose, it is necessary to predict the mechanical properties of hot-rolled strip steel in the whole process in combination with the deformation mechanism model and the intelligent model of big data neural network around the rolling process and the cooling system after rolling, so as to improve the precision and production efficiency of the controlled rolling and controlled cooling of the hot rolling production line. SUMMARY

[0003] The application provides a mechanism cooperative prediction method and system for predicting mechanical properties of hot-rolled strip steel in a whole process, which aims to simultaneously consider the microstructure deformation process of the finish rolling section and the temperature drop process of the cooling section after rolling, and to intelligently cooperatively predict the mechanical properties in the whole process of controlled rolling and controlled cooling, so as to provide high-precision deformation resistance for the rolling model and to provide an online prediction model for the mechanical properties after rolling, thereby realizing fine control of the whole process of hot-rolled strip steel.

[0004] The first object of the application is to provide a mechanism cooperative prediction method for predicting mechanical properties of hot-rolled strip steel in a whole process, which comprises the following steps:

[0005] S1, real-time acquisition of process parameters in a whole working condition, online prediction of grain size and phase change of each stand, and big data training of a neural network;

[0006] S2, based on comparison of actual working conditions and sample data sets, using a rolling mechanism model to calibrate the corresponding relationship between rolling force, rolling speed, strain rate, rolling temperature and deformation resistance, and improving the predicted value of mechanical properties;

[0007] S3, taking the rolling mechanical property prediction value as an initial condition, and then performing mechanical property prediction of the post-rolling cooling section according to the post-rolling cooling condition, observing the microstructure and property changes in the controlled cooling process in real time, and realizing online prediction of the mechanical property after the hot-rolled strip is finally cooled;

[0008] S4, combining the intelligent training results of the field measured data and the high-precision online prediction of the mechanism model, the rolling parameters, the cooling parameters and the corresponding mechanical properties of the whole process are identified and clustered in real time, and standard sample data are formed to meet the requirements of subsequent high-precision online prediction.

[0009] Preferably, S1 comprises:

[0010] S101, real-time acquisition of current working condition parameters, prediction of grain size and phase change according to strip composition, rolling force, rolling speed, rolling temperature and strain rate, and comparison with sample data;

[0011] S102, online calculation of the instantaneous mechanical property change according to the predicted value of the grain size and phase change, so as to obtain the corresponding deformation resistance of the strip mechanical property of each rack;

[0012] S103, neural network model training of the current data is performed to form a sample data set.

[0013] Preferably, S2 comprises:

[0014] S201, using the clustering training data of the neural network to obtain the corresponding relationship between the rolling parameters and the deformation resistance corresponding to the working condition, and providing accurate boundary conditions and initial conditions for the mechanism model;

[0015] S202, based on the measured data after the neural network training, the rolling mechanism model is used to perform high-precision online prediction of the deformation resistance, and the predicted value of the deformation resistance conforming to the actual situation is obtained through iteration.

[0016] Preferably, S3 comprises:

[0017] S301, real-time tracking of rolling process parameters and mechanical property prediction values;

[0018] S301, online calculation of heat transfer coefficient and instantaneous temperature according to measured cooling rate, cooling time and cooling water volume;

[0019] S301, based on the above parameters, rapid microstructure and property prediction is performed to obtain the mechanical property change rule of the whole process.

[0020] Preferably, the process parameters of the whole working condition comprise:

[0021] The original chemical composition is collected in the smelting part of the billet, including the content data of C, Si, P and S in the steel;

[0022] The hot-rolled strip production process parameters include: the billet heating furnace discharge temperature, the rough rolling mill inlet temperature, the outlet temperature, the pass reduction, the rolling speed, the outlet thickness, the finish rolling mill inlet temperature, the outlet temperature, the pass rolling speed, the reduction, the outlet thickness, the coiling temperature, the speed, the thickness, and the like;

[0023] The grain size and deformation resistance parameters obtained by combining the finished product analysis form a sample data set.

[0024] Preferably, the data in the sample data set is screened to remove measurement errors caused by measurement problems and abnormal data caused by production problems; the screening includes setting a fluctuation range for each type of parameter, and removing the data that exceeds the fluctuation range as a whole.

[0025] Preferably, the rolling mechanism model includes a recrystallization model and a flow stress model, which are used to calculate the micro changes of the hot-rolled strip and to perform mechanical property prediction with the grain size, recrystallization rate and flow stress parameters as input conditions for the mechanical property prediction; each rolling stage includes a heating furnace discharge stage, rough rolling passes, finish rolling passes and a layer cooling stage.

[0026] The second object of the present application is to provide a mechanism collaborative prediction system for predicting the mechanical properties of hot-rolled strips in the whole process, which comprises:

[0027] The training module: real-time acquisition of process parameters in the whole working condition, online prediction of grain size and phase change of each rack, and at the same time, big data training of neural network;

[0028] The relationship establishment module: based on the comparison of actual working conditions and sample data set, the corresponding relationship between rolling force, rolling speed, strain rate, rolling temperature and deformation resistance is calibrated by using the rolling mechanism model, so as to improve the predicted value of mechanical properties;

[0029] The analysis module: taking the rolling mechanical property prediction value as the initial condition, and then according to the cooling conditions after rolling, the mechanical property prediction of the cooling section after rolling is carried out, the microstructure performance change in the cooling process is observed in real time, and the online prediction of the mechanical properties after the final cooling of the hot-rolled strip is realized;

[0030] The standard sample generation module: combining the big data intelligent training of the field measurement data and the real-time prediction of the mechanism model, the rolling parameters, cooling parameters and corresponding mechanical properties in the whole process are identified and clustered to form standard sample data, which meets the subsequent online prediction requirements.

[0031] The third object of the present application is to provide an information data processing terminal for realizing the mechanism collaborative prediction method for predicting the mechanical properties of hot-rolled strips in the whole process.

[0032] The fourth object of the present application is to provide a computer-readable storage medium comprising instructions which, when executed on a computer, cause the computer to perform the above-mentioned mechanism collaborative prediction method for predicting mechanical properties of hot-rolled strip steel in the whole process.

[0033] The present application has the advantages and positive effects that:

[0034] The present application uses mechanism models to predict grain size and phase change in real time, and uses intelligent algorithms to train the corresponding relationship between process parameters and microstructure change based on big data, so as to obtain online mechanical property prediction values that meet the actual needs of production, so that they better match the actual values. On this basis, a real-time prediction system is constructed based on the whole mechanism, which can track and control the mechanical properties of hot-rolled strip steel in the whole process, and feedback to the rolling mill system in real time, which helps to improve the prediction accuracy of rolling force, online optimization of rolling process and accurate control of mechanical properties of rolled strip steel. Therefore, the collaborative prediction method for mechanical properties of hot-rolled strip steel in the whole process is very beneficial to improve the rolling efficiency, cooling efficiency and product performance of the strip steel.

[0035] The present application considers the continuity characteristics of the whole process of the production line, the parameters of each process segment are coupled and inherited in real time, and the neural network is used for big data training, and the mechanical properties are collaboratively predicted by using theoretical mechanism models; this collaborative mode can not only use the optimal topological structure of the neural network for cluster analysis to improve the sample training of complex parameters, but also use the mechanism model to accurately predict the grain size and phase change of the current production condition, so as to meet the real-time prediction of the mechanical properties of the whole process of the production line, and provide control targets for rolling process and cooling process, which helps to realize closed-loop control of mechanical properties, thereby improving the intelligent level and prediction accuracy of the production line. BRIEF DESCRIPTION OF DRAWINGS

[0036] Fig. 1 is a basic collaborative relationship diagram of the finishing process and the post-rolling cooling process in the preferred embodiment of the present application;

[0037] Fig. 2 is a microstructure sample extraction flowchart in the preferred embodiment of the present application;

[0038] Fig. 3 is a full-condition prediction flowchart of mechanical properties of hot-rolled strip steel;

[0039] Fig. 4 is a neural network basic training flowchart of controlled rolling and controlled cooling mechanical properties. DETAILED DESCRIPTION

[0040] In order to further understand the inventive content, characteristics and effects of the present application, the following embodiments are exemplified and described in detail as follows with reference to the accompanying drawings:

[0041] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the technical solutions in the present application, all the other embodiments obtained by those of ordinary skill in the art without creative work, fall within the protection scope of the present application.

[0042] Please refer to Figs. 1 to 4 .

[0043] A mechanism cooperative prediction method for predicting mechanical properties of full-process hot-rolled strip steel is used for synchronous mechanical property prediction of hot rolling, controlled rolling and controlled cooling to obtain high-precision online control of hot-rolled strip steel finishing rolling and post-rolling cooling; comprising the following steps:

[0044] S1, real-time acquisition of process parameters in full working condition, online prediction of grain size and phase change of each stand, and simultaneous big data training of neural network; S1 comprises:

[0045] S101, real-time acquisition of current working condition parameters, comparison of sample data according to strip composition, rolling force, rolling speed, rolling temperature and strain rate, prediction of grain size and phase change;

[0046] S102, online calculation of transient mechanical property change according to the predicted value of grain size and phase change, so as to obtain the corresponding deformation resistance of the mechanical property of each stand;

[0047] S103, neural network model training of current data, forming a sample data set.

[0048] S1 is sample data preparation, based on the current rolling condition, collecting full-process rolling process parameters, and measuring product mechanical properties and grain size parameters, predicting grain size and corresponding deformation resistance of each stand, determining model basic parameters through neural network model training, and forming a sample data set;

[0049] The full-process rolling process parameters include: original chemical composition of the steel and content thereof, hot-rolled strip production process parameters, and mechanical property parameters. The original chemical composition is collected at the steel billet smelting part, mainly including content data of important components such as C, Si, P, S, etc. in the steel; the hot-rolled strip production process parameters are collected at the strip production operation platform, mainly including: billet heating furnace discharge temperature, rough rolling mill inlet temperature, rough rolling mill outlet temperature, rough rolling pass reduction amount, rough rolling pass rolling speed, finishing mill inlet temperature, finishing mill outlet temperature, finishing mill pass rolling speed, finishing mill pass reduction amount, coiling temperature, coiling speed, billet original thickness, rough rolling mill set pass outlet thickness, finishing mill set pass outlet thickness, etc.; the mechanical property parameters are collected through sampling and experiments on the product, mainly including: tensile strength, yield strength, and elongation, etc. After all the data collection is completed, the data is sorted according to the steel coil number of the sampling product to form a data document for data screening.

[0050] The main purpose of the data screening work is to remove measurement errors caused by measurement problems and abnormal data caused by production problems; the main work includes: setting a fluctuation range for each type of parameter, and removing the data that exceeds the fluctuation range as a whole, facilitating subsequent model training and theoretical model use, improving the calculation accuracy of the model, and improving the prediction accuracy of the model.

[0051] The prediction of the grain size of each stand and the deformation resistance is calculated through a basic theoretical model, and the theoretical model mainly includes: recrystallization model and flow stress model, etc. It is mainly used for calculating the micro changes of the hot-rolled strip, and taking the grain size, recrystallization rate, and flow stress as the input conditions for the mechanical property prediction to predict the mechanical properties.

[0052] The slab undergoes recrystallization processes under different deformation conditions during rough rolling and finishing rolling. According to the critical strain of dynamic recrystallization, the recrystallization process can be divided into dynamic recrystallization and static recrystallization processes, and the calculation formulas of the recrystallization rate and the grain size of dynamic recrystallization and static recrystallization are shown in the following formulas (1) and (2):

[0053]

[0054]

[0055] wherein m is a variable related to temperature, ε c is the critical strain of dynamic recrystallization, ε s is related to deformation temperature, strain rate, and initial grain size, D is a constant related to dynamic recrystallization activation energy, Z is the Zener-Hollomen parameter, n is a constant related to the steel grade, and t 0.5is the time required for the recrystallization rate to reach 50%, and d0is the original grain size.

[0056] In addition to calculating the grain size, the rheological stress in the recrystallization process is also calculated, and the calculation formula is shown in the following formula (3):

[0057] σ = σ ss - (σ ss - σ ds ) · {1 - exp[-k (ε - ε p ) n ]} (3)

[0058] 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 the strain rate, the gas constant, and the absolute temperature.

[0059] In the hot rolling process, the dislocation density calculation formula in the deformation and recrystallization process is shown in the following formula (4) and (5):

[0060]

[0061] ρ S = (1 - X S ) ρ + X S ρ0 (5)

[0062] wherein σ and σ X are the rheological stress and the yield stress at the deformation temperature, respectively, M is the Taylor factor, α is a constant, μ and b are the shear modulus and the Burgers modulus, respectively, f ρ is the dislocation density factor, and ρ0is the initial dislocation density.

[0063] In the rolling process, there may be a certain participating strain, and when calculating the grain size of each pass, the single-pass strain ε needs to be added to the residual strain Δε of the previous pass, and the calculation formula is as follows:

[0064]

[0065] The strip rolling process parameters collected in the early data collection work can be calculated to calculate the recrystallization type and recrystallization grain size and recrystallization rate of austenite grains in each pass in the finishing and roughing processes through theoretical calculation. The recrystallized grains will undergo phase transformation in the laminar cooling process, and the austenite grains will be transformed into ferrite, pearlite, etc. under the influence of cooling rate and cooling conditions. Since the process of phase transformation is not easy to calculate using a theoretical model, in order to ensure the accuracy of the calculation of mechanical properties, a neural network intelligent algorithm is used to further predict the mechanical properties.

[0066] S2, based on the comparison of the actual working condition and the sample data set, using the rolling mechanism model to correct the corresponding relationship of rolling force, rolling speed, strain rate, rolling temperature and deformation resistance, and to improve the predicted value of mechanical properties; S2 includes:

[0067] S201, using the clustering training data of the neural network, obtaining the corresponding relationship of rolling parameters and deformation resistance corresponding to the working condition, providing boundary conditions and initial conditions for the mechanism model;

[0068] S202, based on the measured data after the neural network training, using the rolling mechanism model to perform online prediction of deformation resistance, and obtaining high-precision deformation resistance prediction value through iteration.

[0069] S2 is to use the rolling mechanism model to calculate the deformation resistance, and to use the neural network to train the characteristic parameters to obtain the cooperative topology structure of the mechanism model prediction parameters and the measured parameters.

[0070] The training and use of the neural network model mainly include: sample data sorting and classification, neural network model training and verification, and actual use of the neural network model.

[0071] In the data collected, the data of the laminar cooling section (finish rolling outlet temperature, coiling temperature, coiling speed, etc.) and the mechanical property data (tensile strength, yield strength, elongation, etc.) are extracted, and they are corresponded with the recrystallization rate and grain size calculated by the theoretical model of each rolling pass, combined into a new data set, and the obtained data is randomly divided into three parts of training set, test set and verification set for the training and verification of the neural network model.

[0072] In the original data, including the composition of the strip steel raw material, the rolling data of the finish rolling and the rough rolling, the coiling and laminar cooling data and the mechanical property parameters, the number of data is large, and the data is further processed by the mechanism model, the finish rolling and rough rolling data are processed into the grain size, recrystallization rate and flow stress of each pass, for the training and verification of the neural network model.

[0073] Before the neural network model is trained, the data needs to be screened first, and the Laplacian score is used to determine the correlation in the present application, and the calculation formula is:

[0074]

[0075] In the formula, cov is the covariance, and var is the variance.

[0076] The re-integrated data are calculated for correlation with mechanical properties, and according to different required predicted mechanical property parameters (tensile strength, yield strength and elongation), parameters with higher correlation are selected as input parameters of the neural network model.

[0077] After the training is completed using the training set, the prediction effect of the model is first tested and further corrected using the test set, and the actual prediction effect of the model is tested using the verification set, and finally a set of models for predicting the mechanical properties of hot-rolled strip steel is formed for use in actual production processes.

[0078] S3, using the rolling mechanical property prediction value as the initial condition, and according to the post-rolling cooling condition, the mechanical property prediction of the post-rolling cooling section is performed, the microstructure and property changes in the controlled cooling process are observed in real time, and the mechanical property online prediction after the final cooling of the hot-rolled strip steel is realized; S3 comprises:

[0079] S301, real-time tracking of rolling process parameters and mechanical property prediction values;

[0080] S301, online calculation of heat transfer coefficient and transient temperature according to measured cooling rate, cooling time and cooling water volume;

[0081] S301, based on the above parameters, rapid microstructure and property prediction is performed to obtain the mechanical properties after the final cooling.

[0082] S3 is a combination of a neural network model and a theoretical calculation model for high-precision prediction of each stand, which is used as the initial condition of post-rolling cooling, and based on the simultaneous consideration of post-rolling temperature and cooling rate, the mechanical property prediction of the post-rolling controlled cooling process is performed;

[0083] The neural network model, after being trained, can realize the mechanical property prediction and grain size calculation of the strip steel in each process section of the rolling process, and can optimize the rolling process by combining the actual production process changes, and can perform the controlled cooling process after rolling based on the comprehensive consideration of the whole process, and can perform the mechanical property prediction.

[0084] S4, combined with the intelligent training of large data of field measured data and the real-time prediction of mechanism model, the rolling parameters, cooling parameters and corresponding mechanical properties of the whole process are identified and clustered to form standard sample data, which meets the subsequent online prediction requirements. That is, compared with the offline measured data, the whole working condition is trained, the finishing rolling and post-rolling cooling process parameters are optimized as a whole, and the online prediction requirements of the mechanical properties of the hot-rolled strip steel under the whole working condition are met.

[0085] In the above embodiments:

[0086] Step one, sample data preparation, based on the current rolling situation, collect the whole process rolling process parameters, and measure the product mechanical properties and grain size parameters, predict the grain size and its corresponding deformation resistance of each rack, determine the model parameters through neural network model training, form the sample data set;

[0087] Step two, use the rolling mechanism model to calculate the deformation resistance, and use the neural network to train the characteristic parameters, obtain the collaborative topology structure of the mechanism model prediction parameters and the measured parameters;

[0088] Step three, the neural network model and the theoretical calculation model are combined to carry out high-precision prediction of each rack, which is used as the initial condition of the post-rolling cooling. Based on the synchronous consideration of post-rolling temperature and cooling rate, the mechanical property prediction of the post-rolling controlled cooling process is carried out.

[0089] Step four, compare and verify with offline measured data, train the whole working condition, optimize the finishing rolling and post-rolling cooling process parameters, and meet the online prediction requirements of the mechanical properties of hot rolled strip under all working conditions.

[0090] A mechanism collaborative prediction system for predicting the mechanical properties of hot rolled strip throughout the process, including the changes of strip grain size and deformation resistance in each rolling stage calculated by the basic theoretical model, and the product deformation resistance and grain size predicted by the neural network model, which facilitates real-time monitoring of product quality changes during the rolling production process and dynamic adjustment of rolling process.

[0091] Training module: real-time acquisition of process parameters under all working conditions, online prediction of grain size and phase change of each rack, and big data training of neural network;

[0092] Relationship establishment module: based on the comparison of actual working conditions and sample data set, the corresponding relationship between rolling force, rolling speed, strain rate, rolling temperature and deformation resistance is corrected using the rolling mechanism model, and the predicted value of mechanical properties is improved;

[0093] Prediction module: taking the rolling mechanical property prediction value as the initial condition, and then predicting the mechanical properties of the post-rolling cooling section according to the post-rolling cooling conditions, real-time observation of the microstructure and performance changes during the controlled cooling process, and realization of the online prediction of the mechanical properties of hot rolled strip after final cooling;

[0094] Standard sample generation module: combining the big data intelligent training of field measured data and the real-time prediction of mechanism model, identifying and clustering the rolling parameters, cooling parameters and corresponding mechanical properties throughout the process to form standard sample data, meeting the subsequent online prediction requirements.

[0095] An information data processing terminal for implementing the above-mentioned mechanism collaborative prediction method for predicting the mechanical properties of hot rolled strip throughout the process.

[0096] A computer readable storage medium comprising instructions which, when executed on a computer, cause the computer to carry out the mechanism collaborative prediction method for predicting mechanical properties of hot-rolled strip steel in whole process.

[0097] In the above embodiments, all or part of the embodiments can be realized by software, hardware, firmware or any combination thereof. When all or part of the embodiments are realized in the form of a computer program product, the computer program product comprises one or more computer instructions. When the computer program instructions are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instructions can be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL) or wireless (such as infrared, wireless, microwave, etc.)) mode. 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, data center, etc. integrated with one or more available media sets. The available medium can be a magnetic medium (for example, floppy disk, hard disk, magnetic tape), an optical medium (for example, DVD), or a semiconductor medium (for example, solid state disk (SSD)) and the like.

[0098] The above only describes the preferred embodiments of the present application, and does not limit the present application in any form. Any simple modification, equivalent change and modification made on the basis of the technical essence of the present application to the above embodiments are within the scope of the technical solution of the present application.

Claims

1. A mechanism synergistic prediction method for predicting mechanical properties of hot-rolled strip steel in the whole process, characterized in that, Comprise: S1, real-time acquisition of process parameters under all working conditions, online prediction of grain size and phase change of each stand, and large data training of neural network; Specifically comprising: S101, real-time acquisition of current working condition parameters, comparison with sample data according to strip composition, rolling force, rolling speed, rolling temperature and strain rate to predict grain size and phase change; S102, online calculation of transient mechanical property change according to the predicted value of grain size and phase change, so as to obtain the deformation resistance corresponding to the mechanical property of the strip of each stand; S103, neural network model training of current data to form sample data set; S2, based on the comparison between actual working conditions and sample data set, using rolling mechanism model to correct the corresponding relationship between rolling force, rolling speed, strain rate, rolling temperature and deformation resistance, and improve the prediction value of mechanical property; S3, taking the rolling mechanical property prediction value as the initial condition, and then according to the cooling condition after rolling, the mechanical property prediction of the cooling section after rolling is carried out, the microstructure and performance change in the cooling process are observed in real time, and the online prediction of the mechanical property of the hot-rolled strip after final cooling is realized; Specifically comprising: S301, real-time tracking of rolling process parameters and mechanical property prediction value; S301, online calculation of heat transfer coefficient and transient temperature according to measured cooling rate, cooling time and cooling water volume; S301, based on the above parameters, quickly predict the microstructure and performance to obtain the mechanical property change rule of the whole process; S4, combining the online prediction of the rolling mechanism model with the intelligent training results of the measured data and the online prediction of the rolling mechanism model, the rolling parameters, cooling parameters and corresponding mechanical properties of the whole process are identified and clustered in real time, and standard sample data is formed to meet the subsequent online prediction requirements.

2. The mechanism synergic prediction method of predicting mechanical properties of hot-rolled steel strips throughout the whole process according to claim 1, characterized in that S2 Comprise: S201, using the clustering training data of neural network to obtain the corresponding relationship between rolling parameters and deformation resistance under corresponding working conditions, and providing boundary conditions and initial conditions for mechanism model; S202, based on the measured data after neural network training, using rolling mechanism model to carry out online prediction of deformation resistance, and obtaining deformation resistance prediction value conforming to actual situation through iteration.

3. The mechanism synergic prediction method of predicting mechanical properties of hot-rolled steel strips throughout the whole process according to claim 1, characterized in that, The process parameters under all working conditions comprise: The original chemical composition is collected in the smelting part of the billet, including the content data of C, Si, P and S in the steel; The hot-rolled strip production process parameters include: billet heating furnace discharge temperature, rough rolling mill inlet temperature, outlet temperature, pass reduction, rolling speed, outlet thickness, finishing mill inlet temperature, outlet temperature, pass rolling speed, reduction, outlet thickness, coiling temperature, speed and thickness; The sample data set is formed by combining the grain size and deformation resistance parameters obtained by analyzing the finished product.

4. The mechanism synergic prediction method of predicting mechanical properties of hot rolling strip steel in whole process according to claim 1, characterized in that, The data in the sample data set is screened to remove measurement errors caused by measurement problems and abnormal data caused by production problems; The screening includes: setting the fluctuation range of each type of parameter, and removing the data beyond the fluctuation range as a whole.

5. The mechanism synergic prediction method of predicting mechanical properties of hot rolling strip steel in whole process according to claim 1, characterized in that, The rolling mechanism model comprises a recrystallization model and a rheological stress model, is used for calculating micro changes of hot-rolled strip steel, and takes grain size, recrystallization rate and rheological stress parameters as input conditions of mechanical property prediction to perform mechanical property prediction; each rolling stage comprises a heating furnace discharge stage, rough rolling passes, finish rolling passes and a layer cooling stage.

6. A mechanism synergic prediction system for predicting mechanical properties of hot-rolled strip steel throughout the whole process, characterized in that, Comprise: The training module: real-time acquisition of process parameters in all working conditions, online prediction of grain size and phase change of each rack, and at the same time, big data training of neural network; Specifically comprising: Real-time acquisition of current working condition parameters, comparison of sample data according to strip composition, rolling force, rolling speed, rolling temperature and strain rate, prediction of grain size and phase change; According to the predicted value of grain size and phase change, the instantaneous mechanical property change is calculated online, so as to obtain the corresponding deformation resistance of the strip steel mechanical property of each rack; Carrying out neural network model training of current data to form a sample data set; Analysis module: based on the comparison of actual working conditions and sample data set, the corresponding relationship between rolling force, rolling speed, strain rate, rolling temperature and deformation resistance is corrected by using the rolling mechanism model, and the prediction value of mechanical property is improved; The prediction module: taking the rolling mechanical property prediction value as the initial condition, and then according to the cooling condition after rolling, the mechanical property prediction of the cooling section after rolling is carried out, the microstructure and performance change in the cooling process is observed in real time, and the online prediction of the mechanical property of the hot-rolled strip after final cooling is realized; comprising: Real-time tracking of rolling process parameters and mechanical property prediction value; According to the measured cooling rate, cooling time and cooling water quantity, the heat exchange coefficient and instantaneous temperature are calculated online; Based on the above parameters, the microstructure and performance are quickly predicted, and the mechanical property change rule of the whole process is obtained; The standard sample generation module: combining the big data intelligent training of the field measured data and the real-time prediction of the mechanism model, the rolling parameters, cooling parameters and corresponding mechanical properties of the whole process are identified and clustered to form standard sample data, which meets the subsequent online prediction requirements.

Images