CN111191822A - Steel parameter prediction method and terminal equipment - Google Patents

Abstract

Translated fromChinese

本发明适用于轧钢技术领域，提供了一种钢材的参数预测方法及终端设备，包括：获取钢材生产过程中的历史生产数据，并根据历史生产数据构建训练样本集；训练样本集包括携带标签的热轧参数，携带标签的热轧参数所携带的标签为对应的力学性能参数；利用携带标签的热轧参数对力学性能预测模型进行训练，最后将第一目标热轧参数值输入训练完成的力学性能预测模型计算得到第一目标力学性能参数值。本发明通过对以往钢材生产过程中的历史生产数据进行分析建立钢材的力学性能参数预测模型，并根据钢材的热轧参数以及力学性能预测模型得到钢材的力学性能参数，计算速度快，无需切割样本进行测试，节省人工及材料成本，提高了钢材性能参数的确定效率。

The invention is applicable to the technical field of steel rolling, and provides a method for predicting parameters of steel and terminal equipment, including: acquiring historical production data in the steel production process, and constructing a training sample set according to the historical production data; Hot rolling parameters, the labels carried by the hot rolling parameters with labels are the corresponding mechanical properties parameters; the hot rolling parameters with labels are used to train the mechanical properties prediction model, and finally the first target hot rolling parameter value is input into the trained mechanics The performance prediction model calculates the first target mechanical performance parameter value. The present invention establishes the mechanical property parameter prediction model of the steel by analyzing the historical production data in the past steel production process, and obtains the mechanical property parameter of the steel according to the hot rolling parameters of the steel and the mechanical property prediction model, and the calculation speed is fast, and no sample cutting is required. Carry out tests, save labor and material costs, and improve the efficiency of determining steel performance parameters.

Description

Steel parameter prediction method and terminal equipment

Technical Field

The invention belongs to the technical field of steel rolling, and particularly relates to a steel parameter prediction method and terminal equipment.

Background

The hot rolled steel as an economical structural steel is widely applied to various fields such as high-rise buildings, railway bridges, power station construction, petrochemical engineering and the like, and the demand of the hot rolled steel is increasing.

The performance and quality of hot rolled steel products are mostly detected by cutting the hot rolled steel products in batches in the prior art to obtain sample steel, the performance and quality of the hot rolled steel products in the same batch are obtained by testing the sample steel, the operation of the test process is complex, the efficiency is low, labor is wasted, and the waste of the steel is caused.

Disclosure of Invention

In view of this, the embodiment of the invention provides a method for predicting parameters of steel and a terminal device, so as to solve the problems that in the prior art, the performance and quality of a hot rolled steel product need to be detected by cutting a sample steel, the efficiency is low, and manpower and steel are wasted.

The first aspect of the embodiments of the present invention provides a method for predicting parameters of a steel product, including:

obtaining historical production data in the steel production process, and constructing a training sample set according to the historical production data; the training sample set comprises hot rolling parameters carrying labels, and the labels carried by the hot rolling parameters carrying the labels are corresponding mechanical property parameters;

constructing a mechanical property prediction model, and training the mechanical property prediction model by using the hot rolling parameters carrying the labels to obtain a trained mechanical property prediction model;

and acquiring a first target hot rolling parameter value, and inputting the first target hot rolling parameter value into the trained mechanical property prediction model to obtain a first target mechanical property parameter value.

A second aspect of an embodiment of the present invention provides a parameter prediction apparatus for a steel product, including:

the parameter acquisition module is used for acquiring historical production data in the steel production process and constructing a training sample set according to the historical production data; the training sample set comprises hot rolling parameters carrying labels, and the labels carried by the hot rolling parameters carrying the labels are corresponding mechanical property parameters;

the first model training module is used for constructing a mechanical property prediction model, and training the mechanical property prediction model by using the hot rolling parameters carrying the labels to obtain a trained mechanical property prediction model;

and the first result output module is used for acquiring a first target hot rolling parameter value and inputting the first target hot rolling parameter value into the trained mechanical property prediction model to obtain a first target mechanical property parameter value.

A third aspect of the embodiments of the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the method for predicting parameters of a steel product according to the first aspect of the embodiments of the present invention when executing the computer program.

A fourth aspect of the embodiments of the present invention provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the computer program implements the steps of the method for predicting parameters of a steel product according to the first aspect of the embodiments of the present invention.

The method comprises the steps of obtaining historical production data in the steel production process, and constructing a training sample set according to the historical production data; the training sample set comprises hot rolling parameters carrying labels, and the labels carried by the hot rolling parameters carrying the labels are corresponding mechanical property parameters; and training the mechanical property prediction model by using the hot rolling parameters carrying the labels, and finally inputting the first target hot rolling parameter value into the trained mechanical property prediction model to calculate to obtain the first target mechanical property parameter value. According to the embodiment of the invention, the mechanical property parameter prediction model of the steel is established by analyzing the historical production data in the previous steel production process, and the mechanical property parameter of the steel is obtained by automatic calculation according to the hot rolling parameter of the steel and the mechanical property prediction model, so that the calculation speed is high, a sample does not need to be cut for testing, the labor and material costs are saved, and the determination efficiency of the steel property parameter is improved.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.

FIG. 1 is a schematic flow chart illustrating an implementation of a method for predicting parameters of steel according to an embodiment of the present invention;

FIG. 2 is a schematic flow chart illustrating an implementation of a method for predicting parameters of a steel product according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of a historical production data acquisition path provided by an embodiment of the invention;

FIG. 4 is an initial sample set provided by an embodiment of the present inventionHotelling T of each sample²A value diagram;

FIG. 5 shows Hotelling T for each sample in the intermediate sample set according to an embodiment of the present invention²A value diagram;

FIG. 6 is a schematic diagram of coefficients of determination of various initial hot rolling parameters provided by an embodiment of the present invention;

FIG. 7 is a graph illustrating the yield strength prediction and accuracy provided by embodiments of the present invention;

FIG. 8 is a graph illustrating the predicted tensile strength and accuracy provided by an embodiment of the present invention;

FIG. 9 is a diagram illustrating the result of elongation prediction and the accuracy of elongation prediction according to an embodiment of the present invention;

FIG. 10 is a first interface of the online prediction system for the mechanical properties of a hot-rolled strip steel provided by an embodiment of the present invention;

FIG. 11 is a second interface of the online prediction system for the mechanical properties of the hot-rolled strip steel provided by the embodiment of the invention;

FIG. 12 is a schematic view of a steel parameter prediction apparatus according to an embodiment of the present invention;

fig. 13 is a schematic diagram of a terminal device according to an embodiment of the present invention.

Detailed Description

In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular system structures, techniques, etc. in order to provide a thorough understanding of the embodiments of the invention. It will be apparent, however, to one skilled in the art that the present invention may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present invention with unnecessary detail.

In order to explain the technical means of the present invention, the following description will be given by way of specific examples.

Referring to fig. 1, an embodiment of the present invention provides a method for predicting parameters of a steel product, including:

step S101: obtaining historical production data in the steel production process, and constructing a training sample set according to the historical production data; the training sample set comprises hot rolling parameters carrying labels, and the labels carried by the hot rolling parameters carrying the labels are corresponding mechanical property parameters.

In the production process of steel, the hot rolling production line stores production data in the production process, including hot rolling parameters and mechanical property parameters, the hot rolling parameters include process parameters and chemical components, for example, the process parameters may include: rough rolling outlet temperature (RDT), finish rolling outlet temperature (FDT), finish rolling outlet thickness (FDH), Coiling Temperature (CT), coiling temperature set value (CET), reduction rate (epsilon), heating furnace discharge temperature (SRT) and the like; the chemical composition may include: carbon content (C), silicon content (Si), manganese content (Mn), sulfur content (S), vanadium content (V), phosphorus content (P), and the like; the mechanical property parameters may include: yield strength, tensile strength and elongation. Under the conditions of different process parameters and different chemical compositions, the mechanical property parameters of the produced steel are different, and the mechanical property parameters and the hot rolling parameters of the steel have a certain corresponding relation according to the physical metallurgy law.

1780 the production line has large data volume, more data points and perfect data, for example 1780 hot rolled strip steel, referring to fig. 3, historical production data is stored in a secondary server database, a process quality system, a tertiary server database and a materialized view database, and different parameters are respectively stored in each data storage server. And calling data from the secondary server database, the tertiary server database and the materialized view database by using SQL sentences according to the required data and the authority among the data servers at all levels, and calling the dynamic link library to call the data from the process quality system. Because the steel hot rolling parameters and the performance parameters of the same steel coil are the same, the data in the four data storage servers are matched by taking the steel coil number and the furnace number as keywords, and the hot rolling parameters and the corresponding mechanical performance parameters in the steel production process are stored in a one-to-one correspondence manner, so that the historical production data in the conventional steel production process are obtained. And constructing a training sample set according to the historical production data stored in a one-to-one correspondence manner, wherein the training sample set comprises hot rolling parameters carrying labels, and the labels carried by the hot rolling parameters carrying the labels are corresponding mechanical performance parameters. In some embodiments, to ensure comprehensiveness of the historical production data, the historical production data includes data for multiple steel grades.

In some embodiments, constructing the training sample set from historical production data may include:

step S1011: constructing an initial sample set according to historical production data, and preprocessing the initial sample set to obtain an intermediate sample set;

due to the fact that abnormal data may exist in historical production data, in order to guarantee the accuracy of training samples, the initial sample set is preprocessed, and abnormal samples are removed to obtain an intermediate sample set.

In some embodiments, preprocessing the initial sample set to obtain an intermediate sample set includes:

determining a control limit for an initial sample set and a hotelling T for each sample in the initial sample set²A value;

if Hotelling T of each sample²If the values are not greater than the control limit of the initial sample set, the initial sample set is an intermediate sample set;

if Hotelling T exists in the initial sample set²Removing the target samples in the initial sample set to obtain a first sample set if the value of the target samples is larger than the control limit of the initial sample set, taking the first sample set as a new initial sample set, and continuously determining the control limit of the initial sample set and the Hotelling T of each sample in the initial sample set²And (4) value. For example, referring to FIG. 4, before preprocessing the initial sample set, Hotelling T for samples No. 142, No. 177, No. 345, No. 350, and No. 374²If the value exceeds the control limit, it can be determined as an abnormal sample and eliminated. Referring to fig. 5, the initial sample set is processed by the above method to obtain an intermediate sample set, and thus hotelling T of each sample in the intermediate sample set²The values are all less than the control limit of the intermediate sample set, and all can be considered normal samples.

By calculating the Hotelling T of each sample²And abnormal samples are eliminated in the control limit of the sample set, so that the accuracy of the training sample set is ensured, and the steel parameter prediction is improvedThe accuracy of measurement is improved, and the calculation amount of the subsequent calculation process is reduced.

Step S1012: and determining a target hot rolling parameter set influencing the mechanical property parameters according to the intermediate sample set, and extracting data of the target hot rolling parameter set and data of the corresponding mechanical property parameters from the intermediate sample set to construct a training sample set.

Because each sample in the intermediate sample set can comprise a plurality of hot rolling parameters, the influence of each hot rolling parameter on the mechanical property parameter is different, the hot rolling parameter with larger influence on the mechanical property parameter in each hot rolling parameter can be selected to form a target hot rolling parameter set, and the data of the target hot rolling parameter set and the data of the corresponding mechanical property parameter are extracted from the intermediate sample set to construct a training sample set, so that the complexity of subsequent calculation is reduced on the premise of not influencing the parameter prediction accuracy, and the calculation speed is increased.

In some embodiments, the intermediate sample set comprises a correspondence of mechanical property parameters to an initial set of hot rolling parameters, the initial set of hot rolling parameters comprising a plurality of initial hot rolling parameters; determining a target hot rolling parameter set influencing the mechanical property parameters according to the intermediate sample set, wherein the target hot rolling parameter set comprises the following steps:

establishing unary linear regression models respectively corresponding to the mechanical performance parameters and the initial hot rolling parameters according to the corresponding relation between the mechanical performance parameters and the initial hot rolling parameter set, and calculating unary linear regression decision coefficients of the initial hot rolling parameters according to the unary linear regression models and the intermediate sample set;

selecting an initial hot rolling parameter with the maximum unary linear regression decision coefficient, recording as a first hot rolling parameter, and removing the first hot rolling parameter from the initial hot rolling parameter set to obtain an intermediate hot rolling parameter set and a corresponding relation between the mechanical property parameter and the intermediate hot rolling parameter set;

establishing a binary linear regression model corresponding to each hot rolling parameter in the mechanical property parameter, the first hot rolling parameter and the intermediate hot rolling parameter set according to the corresponding relation between the mechanical property parameter and the intermediate hot rolling parameter set, and calculating a binary linear regression decision coefficient of each hot rolling parameter in the intermediate hot rolling parameter set according to each binary linear regression model and the intermediate sample set;

if the binary linear regression decision coefficients of all the hot rolling parameters in the intermediate hot rolling parameter set are the same, all the selected first hot rolling parameters form a target hot rolling parameter set;

and if the binary linear regression decision coefficients of the hot rolling parameters in the intermediate hot rolling parameter set are not identical, selecting the hot rolling parameter with the maximum binary linear regression decision coefficient, recording the hot rolling parameter as a second hot rolling parameter, taking the second hot rolling parameter as a new first hot rolling parameter, taking the intermediate hot rolling parameter set as a new initial hot rolling parameter, and continuously removing the first hot rolling parameter from the initial hot rolling parameter set to obtain the intermediate hot rolling parameter set and the corresponding relation between the mechanical property parameter and the intermediate hot rolling parameter set.

Referring to table 1 and fig. 6, the determination process of the target hot rolling parameter set will be described with reference to table 1, taking the mechanical property parameter as the yield strength as an example.

Firstly, calculating a unary linear regression decision coefficient of each initial hot rolling parameter, wherein the unary linear regression decision coefficient of the finish rolling outlet thickness (FDH) is the largest, recording the unary linear regression decision coefficient as a first hot rolling parameter, removing the first hot rolling parameter from the initial hot rolling parameter set, then establishing a binary linear regression model corresponding to each hot rolling parameter in the first hot rolling parameter and the intermediate hot rolling parameter set respectively, and calculating the binary linear regression decision coefficient of each hot rolling parameter in the intermediate hot rolling parameter set according to each binary linear regression model and the intermediate sample set; obtaining the maximum binary linear regression decision coefficient of the Coiling Temperature (CT) through the second calculation, recording the maximum binary linear regression decision coefficient as a first hot rolling parameter, removing the first hot rolling parameter from the initial hot rolling parameter set, repeatedly executing the steps until the binary linear regression decision coefficients of all the hot rolling parameters in the intermediate hot rolling parameter set are the same, and forming a target hot rolling parameter set by all the selected first hot rolling parameters

TABLE 1 determination coefficient table for each hot rolling parameter corresponding to yield strength

In some embodiments, the binary linear regression coefficients for each hot rolling parameter in the intermediate hot rolling parameter set are the same and may include:

the absolute value of the difference value of each hot rolling parameter in the intermediate hot rolling parameter set is less than or equal to a preset fault tolerance value. That is, if the absolute value of the difference between the respective hot rolling parameters is less than or equal to the predetermined tolerance value, it is considered that the binary linear regression determination coefficients of the respective hot rolling parameters are the same.

In some embodiments, the target hot rolling parameter set may include: a finish rolling outlet thickness (FDH), a finish rolling outlet temperature (FDT), a rough rolling outlet temperature (RDT), a Coiling Temperature (CT), a furnace discharge temperature (SRT), a reduction rate (epsilon), a coiling temperature set value (CET), a vanadium content (V), a silicon content (Si) and a sulfur content (S).

Step S102, constructing a mechanical property prediction model, and training the mechanical property prediction model by using hot rolling parameters carrying labels to obtain a trained mechanical property prediction model;

in some embodiments, the mechanical property prediction model may be a bayesian neural network model.

And obtaining the optimal prediction effect under the condition of fixing the hidden nodes in the model training process, and then comparing the optimal prediction effects of the network under the conditions of different numbers of the hidden nodes to obtain the optimal number of the hidden nodes.

Step S103, obtaining a first target hot rolling parameter value, and inputting the first target hot rolling parameter value into the trained mechanical property prediction model to obtain the first target mechanical property parameter value.

The first target hot rolling parameter value is a hot rolling parameter for producing a steel product to be predicted, and the first target mechanical property parameter value is a mechanical property parameter value of the predicted steel product produced by the hot rolling parameter.

The method comprises the steps of obtaining historical production data in the steel production process, and constructing a training sample set according to the historical production data; the training sample set comprises hot rolling parameters carrying labels, and the labels carried by the hot rolling parameters carrying the labels are corresponding mechanical property parameters; and training the mechanical property prediction model by using the hot rolling parameters carrying the labels, and finally inputting the first target hot rolling parameter value into the trained mechanical property prediction model to calculate to obtain the first target mechanical property parameter value. According to the embodiment of the invention, the mechanical property parameter prediction model of the steel is established by analyzing the historical production data in the previous steel production process, and the mechanical property parameter of the steel is obtained by automatic calculation according to the hot rolling parameter of the steel and the mechanical property prediction model, so that the calculation speed is high, a sample does not need to be cut for testing, the labor and material costs are saved, and the determination efficiency of the steel property parameter is improved.

In some embodiments, the mechanical property parameter may be any one of yield strength, tensile strength, and elongation.

The above examples are further illustrated by taking the mechanical property parameters as yield strength as examples.

1. Respectively acquiring historical production data in a secondary server database, a process quality system, a tertiary server database and a materialized view database, matching the data in four data storage servers by taking a steel coil number and a furnace number as keywords, and storing the data in a one-to-one correspondence manner to acquire historical production data;

2. and constructing an initial sample set according to the historical production data, wherein the initial sample set comprises hot rolling parameters carrying labels, and the labels carried by the hot rolling parameters carrying the labels are corresponding yield strength values.

3. According to Hotelling T²Preprocessing the initial sample set, and removing abnormal samples in the initial sample set to form an intermediate sample set;

4. selecting initial hot rolling parameters which have large influence on yield strength by adopting a linear regression model to form a target hot rolling parameter set, and extracting data of the target hot rolling parameter set and corresponding yield strength data from the intermediate sample set to form a training sample set;

5. constructing a yield strength prediction model by adopting a Bayesian neural network model, and training the yield strength prediction model by utilizing a training sample set to obtain a trained yield strength prediction model;

6. and inputting the hot rolling parameters into the trained yield strength prediction model to obtain the predicted value of the yield strength.

Referring to fig. 7, the yield strength of the steel is predicted by the yield strength prediction model, and the actual yield strength of the steel is detected, so that the fault tolerance value is ± 6, the accuracy of the yield strength prediction model can reach 87.21% through statistics, the accuracy is high, and the yield strength of the steel can be well predicted. Referring to fig. 8 and 9, the accuracy of predicting the tensile strength is 93.28%, and the accuracy of predicting the elongation is 94.7%. Therefore, the mechanical property parameter prediction method has high accuracy in predicting the mechanical property parameters, and can well realize the prediction of the mechanical property parameters.

If the system needs to simultaneously predict the yield strength, the tensile strength and the elongation, three mechanical property prediction models can be respectively established corresponding to the three mechanical property parameters, so that the simultaneous prediction of the yield strength, the tensile strength and the elongation is realized. Referring to fig. 10 and 11, the online prediction system for the mechanical properties of the hot-rolled strip steel can simultaneously predict the yield strength, the tensile strength and the elongation.

In some embodiments, the training sample set may further include mechanical property parameters carrying labels, and the labels carried by the mechanical property parameters carrying labels are corresponding hot rolling parameters;

referring to fig. 2, the method for predicting parameters of a steel material may further include:

step S104: constructing a hot rolling parameter prediction model, and training the hot rolling parameter prediction model by using a training sample set to obtain a trained hot rolling parameter prediction model;

in some embodiments, the hot rolling parameter prediction model may be a bayesian neural network model.

Step S105: and acquiring a second target mechanical property parameter value, and inputting the second target mechanical property parameter value into the trained hot rolling parameter prediction model to obtain a second target hot rolling parameter value.

The second target mechanical property parameter value is a target mechanical property parameter value set by the individualized requirement of the steel, and the second target hot rolling parameter value is a hot rolling parameter which is obtained by prediction and can produce the steel with the target mechanical property parameter value.

In the implementation of the invention, the training sample set can also comprise mechanical property parameters carrying labels, and the labels carried by the mechanical property parameters carrying the labels are corresponding hot rolling parameters; and finally, inputting the second target mechanical property parameters into the trained hot rolling parameter prediction model to calculate to obtain second target hot rolling parameter values. Corresponding to the embodiment, the embodiment of the invention also analyzes historical production data in the previous steel production process to establish a hot rolling parameter prediction model of the steel, automatically calculates the hot rolling parameters of the steel according to the mechanical property parameters of the steel and the trained hot rolling parameter prediction model to obtain the hot rolling parameters of the steel, has high calculation speed, can predict the corresponding hot rolling parameters according to the mechanical property requirements, quickly obtains the optimal hot rolling parameters according to the mechanical property requirements, and realizes the intelligent design of the hot rolling parameters of the steel. The mechanical property requirement can be met without continuously adjusting hot rolling parameters, labor and material costs are saved, and the determining efficiency of the steel hot rolling parameters is improved.

The above prediction of the hot rolling parameters of the steel and the prediction of the mechanical property parameters are not described herein again.

Fig. 12 is a parameter prediction apparatus 3 for a steel material according to an embodiment of the present invention, including:

theparameter acquisition module 31 is used for acquiring historical production data in the steel production process and constructing a training sample set according to the historical production data; the training sample set comprises hot rolling parameters carrying labels, and the labels carried by the hot rolling parameters carrying the labels are corresponding mechanical property parameters;

the firstmodel training module 32 is used for constructing a mechanical property prediction model, and training the mechanical property prediction model by using hot rolling parameters carrying labels to obtain a trained mechanical property prediction model;

and a firstresult output module 33, configured to obtain a first target hot rolling parameter value, and input the first target hot rolling parameter value into the trained mechanical property prediction model to obtain the first target mechanical property parameter value.

In some embodiments, the firstmodel training module 32 may include:

the intermediate sample set constructing unit is used for constructing an initial sample set according to historical production data and preprocessing the initial sample set to obtain an intermediate sample set;

and the training sample set construction unit is used for determining a target hot rolling parameter set influencing the mechanical property parameters according to the intermediate sample set, and extracting data of the target hot rolling parameter set and data of the corresponding mechanical property parameters from the intermediate sample set to construct a training sample set.

In some embodiments, the intermediate sample set construction unit may include:

a first initial parameter determining subunit, configured to determine a control limit of the initial sample set and a hotelling T of each sample in the initial sample set²A value;

a first judging subunit for judging whether the Hotelling T of each sample is satisfied²If the values are not greater than the control limit of the initial sample set, the initial sample set is an intermediate sample set;

a second judging subunit, configured to judge whether there is Hotelling T in the initial sample set²Removing the target samples in the initial sample set to obtain a first sample set if the value of the target samples is larger than the control limit of the initial sample set, taking the first sample set as a new initial sample set, and continuously determining the control limit of the initial sample set and the Hotelling T of each sample in the initial sample set²And (4) value.

In some embodiments, the intermediate sample set comprises a correspondence of mechanical property parameters to an initial set of hot rolling parameters, the initial set of hot rolling parameters comprising a plurality of initial hot rolling parameters; the training sample set construction unit may include:

the first model determining subunit is used for establishing a unary linear regression model corresponding to the mechanical property parameters and each initial hot rolling parameter according to the corresponding relation between the mechanical property parameters and the initial hot rolling parameter set, and calculating an unary linear regression decision coefficient of each initial hot rolling parameter according to each unary linear regression model and the intermediate sample set;

the intermediate hot rolling parameter set determining sub-unit is used for selecting an initial hot rolling parameter with the maximum unitary linear regression decision coefficient, recording the initial hot rolling parameter as a first hot rolling parameter, and removing the first hot rolling parameter from the initial hot rolling parameter set to obtain an intermediate hot rolling parameter set and a corresponding relation between the mechanical property parameter and the intermediate hot rolling parameter set;

the second model determining subunit is used for establishing a binary linear regression model corresponding to each hot rolling parameter in the mechanical property parameter, the first hot rolling parameter and the intermediate hot rolling parameter set respectively according to the corresponding relation between the mechanical property parameter and the intermediate hot rolling parameter set, and calculating a binary linear regression decision coefficient of each hot rolling parameter in the intermediate hot rolling parameter set according to each binary linear regression model and the intermediate sample set;

the third judging subunit is used for selecting each first hot rolling parameter to form a target hot rolling parameter set if the binary linear regression decision coefficients of each hot rolling parameter in the intermediate hot rolling parameter set are the same;

and a fourth judging subunit, configured to, if the binary linear regression decision coefficients of the hot rolling parameters in the intermediate hot rolling parameter set are not identical, select the hot rolling parameter with the largest binary linear regression decision coefficient, record the hot rolling parameter as the second hot rolling parameter, use the second hot rolling parameter as the new first hot rolling parameter, use the intermediate hot rolling parameter set as the new initial hot rolling parameter, and continue to perform the step of removing the first hot rolling parameter from the initial hot rolling parameter set to obtain the intermediate hot rolling parameter set and the corresponding relationship between the mechanical property parameter and the intermediate hot rolling parameter set.

In some embodiments, the target hot rolling parameter set comprises: a finish rolling outlet thickness (FDH), a finish rolling outlet temperature (FDT), a rough rolling outlet temperature (RDT), a Coiling Temperature (CT), a furnace discharge temperature (SRT), a reduction rate (epsilon), a coiling temperature set value (CET), a vanadium content (V), a silicon content (Si) and a sulfur content (S).

In some embodiments, the training sample set may further include mechanical property parameters carrying labels, and the labels carried by the mechanical property parameters carrying labels are corresponding hot rolling parameters;

the steel material parameter prediction device 3 may further include:

the second model training module 34 is configured to construct a hot rolling parameter prediction model, and train the hot rolling parameter prediction model by using the training sample set to obtain a trained hot rolling parameter prediction model;

and the secondresult output module 35 is configured to obtain a second target mechanical property parameter value, and input the second target mechanical property parameter value into the trained hot rolling parameter prediction model to obtain a second target hot rolling parameter value.

In some embodiments, the hot rolling parameter prediction model may be a bayesian neural network model.

Fig. 13 is a schematic block diagram of a terminal device according to an embodiment of the present invention. As shown in fig. 13, theterminal device 4 of this embodiment includes: one ormore processors 40, amemory 41, and acomputer program 42 stored in thememory 41 and executable on theprocessors 40. Theprocessor 40 implements the steps in the above-described embodiment of the method for predicting parameters of steel products, such as steps S101 to S103 shown in fig. 1, when executing thecomputer program 42. Alternatively, theprocessor 40 implements the functions of the modules/units in the above-described steel parameter prediction apparatus embodiment, for example, the functions of themodules 31 to 33 shown in fig. 12, when executing thecomputer program 42.

Illustratively, thecomputer program 42 may be partitioned into one or more modules/units that are stored in thememory 41 and executed by theprocessor 40 to accomplish the present application. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, which are used to describe the execution process of thecomputer program 42 in theterminal device 4. For example, thecomputer program 42 may be partitioned into a parameter acquisition module, a first model training module, and a first result output module.

Other modules or units can refer to the description of the embodiment shown in fig. 12, and are not described in detail here.

Theterminal device 4 includes, but is not limited to, aprocessor 40 and amemory 41. It will be understood by those skilled in the art that fig. 13 is only one example of a terminal device, and does not constitute a limitation toterminal device 4, and may include more or less components than those shown, or combine some components, or different components, for example,terminal device 4 may also include an input device, an output device, a network access device, a bus, etc.

TheProcessor 40 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

Thememory 41 may be an internal storage unit of the terminal device, such as a hard disk or a memory of the terminal device. Thememory 41 may also be an external storage device of the terminal device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, which are provided on the terminal device. Further, thememory 41 may also include both an internal storage unit of the terminal device and an external storage device. Thememory 41 is used for storing thecomputer program 42 and other programs and data required by the terminal device. Thememory 41 may also be used to temporarily store data that has been output or is to be output.

In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or illustrated in a certain embodiment.

Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

In the embodiments provided in the present application, it should be understood that the disclosed terminal device and method may be implemented in other ways. For example, the above-described terminal device embodiments are merely illustrative, and for example, the division of the modules or units is only one logical function division, and there may be other divisions when actually implemented, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The integrated modules/units, if implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium. Based on such understanding, all or part of the flow in the method of the embodiments described above may be implemented by a computer program, which may be stored in a computer readable storage medium and used by a processor to implement the steps of the embodiments of the methods described above. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, etc. It should be noted that the computer readable medium may contain other components which may be suitably increased or decreased as required by legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer readable media which may not include electrical carrier signals and telecommunications signals in accordance with legislation and patent practice.

It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present invention, and are intended to be included within the scope of the present invention.

Claims (10)

Translated fromChinese

1.一种钢材的参数预测方法，其特征在于，包括：1. a parameter prediction method of steel, is characterized in that, comprises:获取钢材生产过程中的历史生产数据，并根据所述历史生产数据构建训练样本集；所述训练样本集包括携带标签的热轧参数，所述携带标签的热轧参数所携带的标签为对应的力学性能参数；Obtain historical production data in the steel production process, and construct a training sample set according to the historical production data; the training sample set includes hot rolling parameters with labels, and the labels carried by the hot rolling parameters with labels are corresponding mechanical properties parameters;构建力学性能预测模型，利用所述携带标签的热轧参数对所述力学性能预测模型进行训练，得到训练完成的力学性能预测模型；constructing a mechanical property prediction model, and using the labeled hot rolling parameters to train the mechanical property prediction model to obtain a trained mechanical property prediction model;获取第一目标热轧参数值，并将所述第一目标热轧参数值输入所述训练完成的力学性能预测模型得到第一目标力学性能参数值。A first target hot rolling parameter value is acquired, and the first target hot rolling parameter value is input into the trained mechanical property prediction model to obtain a first target mechanical property parameter value.2.如权利要求1所述的钢材的参数预测方法，其特征在于，所述训练样本集还包括携带标签的力学性能参数，所述携带标签的力学性能参数所携带的标签为对应的热轧参数；2 . The method for predicting parameters of steel according to claim 1 , wherein the training sample set further comprises a mechanical property parameter carrying a label, and the label carried by the mechanical property parameter carrying a label is the corresponding hot rolling. 3 . parameter;所述钢材的参数预测方法还包括：The method for predicting the parameters of the steel further includes:构建热轧参数预测模型，利用所述训练样本集对所述热轧参数预测模型进行训练，得到训练完成的热轧参数预测模型；constructing a hot-rolling parameter prediction model, and using the training sample set to train the hot-rolling parameter prediction model to obtain a trained hot-rolling parameter prediction model;获取第二目标力学性能参数值，并将所述第二目标力学性能参数值输入所述训练完成的热轧参数预测模型得到第二目标热轧参数值。Obtaining a second target mechanical property parameter value, and inputting the second target mechanical property parameter value into the trained hot rolling parameter prediction model to obtain a second target hot rolling parameter value.3.如权利要求1所述的钢材的参数预测方法，其特征在于，所述根据所述历史生产数据构建训练样本集，包括：3. The method for predicting parameters of steel as claimed in claim 1, wherein the constructing a training sample set according to the historical production data comprises:根据所述历史生产数据构建初始样本集，并对所述初始样本集进行预处理得到中间样本集；Construct an initial sample set according to the historical production data, and preprocess the initial sample set to obtain an intermediate sample set;根据所述中间样本集确定影响力学性能参数的目标热轧参数集合，并在所述中间样本集中提取所述目标热轧参数集合的数据及对应的力学性能参数的数据构建训练样本集。A target hot rolling parameter set affecting mechanical performance parameters is determined according to the intermediate sample set, and data of the target hot rolling parameter set and corresponding mechanical performance parameter data are extracted from the intermediate sample set to construct a training sample set.4.如权利要求3所述的钢材的参数预测方法，其特征在于，所述对所述初始样本集进行预处理得到中间样本集，包括：4. The method for predicting parameters of steel according to claim 3, wherein the preprocessing of the initial sample set to obtain an intermediate sample set comprises:确定所述初始样本集的控制限及所述初始样本集中的各个样本的霍特林T²值；determining the control limit of the initial sample set and the Hotelling T² value for each sample in the initial sample set;若所述各个样本的霍特林T²值均不大于所述初始样本集的控制限，则所述初始样本集为所述中间样本集；If the Hotelling T² value of each sample is not greater than the control limit of the initial sample set, the initial sample set is the intermediate sample set;若所述初始样本集中存在霍特林T²值大于所述初始样本集的控制限的目标样本，则去除所述初始样本集中的所述目标样本，得到第一样本集，并将所述第一样本集作为新的初始样本集，并继续执行所述确定所述初始样本集的控制限及所述初始样本集中的各个样本的霍特林T²值的步骤。If there is a target sample whose Hotelling T² value is greater than the control limit of the initial sample set, remove the target sample from the initial sample set to obtain a first sample set, and use the The first sample set is used as a new initial sample set, and the step of determining the control limit of the initial sample set and the Hotelling T² value of each sample in the initial sample set is continued.5.如权利要求3所述的钢材的参数预测方法，其特征在于，所述中间样本集包括力学性能参数与初始热轧参数集合的对应关系，所述初始热轧参数集合包括多个初始热轧参数；所述根据所述中间样本集确定影响力学性能参数的目标热轧参数集合，包括：5 . The method for predicting parameters of steel according to claim 3 , wherein the intermediate sample set includes a corresponding relationship between mechanical property parameters and an initial hot rolling parameter set, and the initial hot rolling parameter set includes a plurality of initial hot rolling parameters. 6 . rolling parameters; the target hot rolling parameter set that affects mechanical performance parameters is determined according to the intermediate sample set, including:根据所述力学性能参数与初始热轧参数集合的对应关系，建立力学性能参数与各个初始热轧参数分别对应的一元线性回归模型，并根据各个一元线性回归模型和所述中间样本集计算各个初始热轧参数的一元线性回归决定系数；According to the corresponding relationship between the mechanical performance parameters and the initial hot rolling parameter set, a univariate linear regression model corresponding to the mechanical performance parameters and each initial hot rolling parameter is established, and each initial hot rolling model is calculated according to each univariate linear regression model and the intermediate sample set. Univariate linear regression coefficient of determination of hot rolling parameters;选取一元线性回归决定系数最大的初始热轧参数，记为第一热轧参数，并将所述第一热轧参数从所述初始热轧参数集合中去除，得到中间热轧参数集合以及力学性能参数与所述中间热轧参数集合的对应关系；Select the initial hot rolling parameter with the largest coefficient of determination of univariate linear regression, record it as the first hot rolling parameter, and remove the first hot rolling parameter from the initial hot rolling parameter set to obtain an intermediate hot rolling parameter set and mechanical properties The corresponding relationship between the parameters and the set of intermediate hot rolling parameters;根据所述力学性能参数与所述中间热轧参数集合的对应关系，建立力学性能参数、所述第一热轧参数和所述中间热轧参数集合中的各个热轧参数分别对应的二元线性回归模型，并根据各个二元线性回归模型和所述中间样本集计算所述中间热轧参数集合中的各个热轧参数的二元线性回归决定系数；According to the corresponding relationship between the mechanical property parameter and the intermediate hot rolling parameter set, a binary linear corresponding to the mechanical property parameter, the first hot rolling parameter and each hot rolling parameter in the intermediate hot rolling parameter set is established. regression model, and calculate the binary linear regression coefficient of determination of each hot rolling parameter in the intermediate hot rolling parameter set according to each binary linear regression model and the intermediate sample set;若所述中间热轧参数集合中的各个热轧参数的二元线性回归决定系数均相同，则选取的各个第一热轧参数组成所述目标热轧参数集合；If the binary linear regression coefficient of determination of each hot rolling parameter in the intermediate hot rolling parameter set is the same, then each selected first hot rolling parameter constitutes the target hot rolling parameter set;若所述中间热轧参数集合中的各个热轧参数的二元线性回归决定系数不完全相同，则选取二元线性回归决定系数最大的热轧参数，记为第二热轧参数，并将所述第二热轧参数作为新的第一热轧参数，将所述中间热轧参数集合作为新的初始热轧参数，继续执行所述将所述第一热轧参数从所述初始热轧参数集合中去除，得到中间热轧参数集合以及力学性能参数与所述中间热轧参数集合的对应关系的步骤。If the binary linear regression coefficient of determination of each hot rolling parameter in the intermediate hot rolling parameter set is not exactly the same, the hot rolling parameter with the largest binary linear regression coefficient of determination is selected and recorded as the second hot rolling parameter. The second hot rolling parameter is taken as the new first hot rolling parameter, the set of intermediate hot rolling parameters is taken as the new initial hot rolling parameter, and the process of changing the first hot rolling parameter from the initial hot rolling parameter is continued. The step of removing from the set to obtain the set of intermediate hot rolling parameters and the corresponding relationship between the mechanical property parameters and the set of intermediate hot rolling parameters.6.如权利要求3至5任一项所述的钢材的参数预测方法，其特征在于，所述目标热轧参数集合包括：精轧出口厚度、精轧出口温度、粗轧出口温度、卷取温度、加热炉出炉温度、压下率、卷取温度设定值、钒含量、硅含量及硫含量。6. The method for predicting parameters of steel according to any one of claims 3 to 5, wherein the target hot rolling parameter set comprises: finish rolling outlet thickness, finish rolling outlet temperature, rough rolling outlet temperature, coiling Temperature, furnace release temperature, reduction ratio, coiling temperature set point, vanadium content, silicon content and sulfur content.7.如权利要求1至5任一项所述的钢材的参数预测方法，其特征在于，所述力学性能预测模型为贝叶斯神经网络模型。7 . The method for predicting parameters of steel according to claim 1 , wherein the mechanical property prediction model is a Bayesian neural network model. 8 .8.一种钢材的参数预测装置，其特征在于，包括：8. A parameter prediction device for steel, characterized in that, comprising:参数获取模块，获取钢材生产过程中的历史生产数据，并根据所述历史生产数据构建训练样本集；所述训练样本集包括携带标签的热轧参数，所述携带标签的热轧参数所携带的标签为对应的力学性能参数；The parameter acquisition module acquires historical production data in the steel production process, and constructs a training sample set according to the historical production data; the training sample set includes hot rolling parameters with labels, and the hot rolling parameters with labels carry The label is the corresponding mechanical property parameter;第一模型训练模块，用于构建力学性能预测模型，利用所述携带标签的热轧参数对所述力学性能预测模型进行训练，得到训练完成的力学性能预测模型；The first model training module is used to construct a mechanical property prediction model, and the mechanical property prediction model is trained by using the label-carrying hot rolling parameters to obtain a trained mechanical property prediction model;第一结果输出模块，用于获取第一目标热轧参数值，并将所述第一目标热轧参数值输入所述训练完成的力学性能预测模型得到第一目标力学性能参数值。The first result output module is used to obtain the first target hot rolling parameter value, and input the first target hot rolling parameter value into the trained mechanical property prediction model to obtain the first target mechanical property parameter value.9.一种终端设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机程序，其特征在于，所述处理器执行所述计算机程序时实现如权利要求1至7任一项所述的钢材的参数预测方法的步骤。9. A terminal device, comprising a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor implements the computer program as claimed in the claims when executing the computer program Steps of the method for predicting parameters of steel according to any one of 1 to 7.10.一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现如权利要求1至7任一项所述的钢材的参数预测方法的步骤。10. A computer-readable storage medium storing a computer program, characterized in that, when the computer program is executed by a processor, the processing of the steel material according to any one of claims 1 to 7 is realized. The steps of the parameter prediction method.

Images