CN117150832B - Real-time prediction method and device for cross section shape of hot-rolled digital twin strip steel - Google Patents
Real-time prediction method and device for cross section shape of hot-rolled digital twin strip steel Download PDFInfo
- Publication number
- CN117150832B CN117150832B CN202311436735.4A CN202311436735A CN117150832B CN 117150832 B CN117150832 B CN 117150832B CN 202311436735 A CN202311436735 A CN 202311436735A CN 117150832 B CN117150832 B CN 117150832B
- Authority
- CN
- China
- Prior art keywords
- sparse
- model
- sindy
- strip steel
- time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 93
- 239000010959 steel Substances 0.000 title claims abstract description 93
- 238000000034 method Methods 0.000 title claims abstract description 48
- 238000005096 rolling process Methods 0.000 claims abstract description 57
- 238000004519 manufacturing process Methods 0.000 claims abstract description 48
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 30
- 239000011159 matrix material Substances 0.000 claims description 66
- 239000013598 vector Substances 0.000 claims description 33
- 238000007781 pre-processing Methods 0.000 claims description 14
- 230000009466 transformation Effects 0.000 claims description 12
- 238000012952 Resampling Methods 0.000 claims description 6
- 238000005452 bending Methods 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 102220568473 Dual specificity mitogen-activated protein kinase kinase 1_S222A_mutation Human genes 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims 2
- 230000008569 process Effects 0.000 abstract description 13
- 238000005098 hot rolling Methods 0.000 abstract description 8
- 230000015556 catabolic process Effects 0.000 abstract description 6
- 230000007547 defect Effects 0.000 abstract description 6
- 238000006731 degradation reaction Methods 0.000 abstract description 6
- 238000005457 optimization Methods 0.000 abstract description 6
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 abstract description 4
- 229910052742 iron Inorganic materials 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 17
- 230000015654 memory Effects 0.000 description 6
- 238000003860 storage Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 241000135164 Timea Species 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000005312 nonlinear dynamic Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
- B21B—ROLLING OF METAL
- B21B37/00—Control devices or methods specially adapted for metal-rolling mills or the work produced thereby
- B21B37/28—Control of flatness or profile during rolling of strip, sheets or plates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mechanical Engineering (AREA)
- Control Of Metal Rolling (AREA)
Abstract
The invention discloses a real-time prediction method and device for the cross section shape of a hot-rolled digital twin strip steel, and relates to the technical field of digital twin systems. Comprising the following steps: constructing a hot continuous rolling digital twin production line, and obtaining set parameters; adopting a dynamic modal decomposition DMD algorithm to optimize a nonlinear system dynamics sparse recognition algorithm SINDY model, and establishing a strip steel cross section shape prediction model DMD-SINDY; and obtaining a strip steel cross section shape prediction result of the hot continuous rolling digital twin production line according to the set parameters and the strip steel cross section shape prediction model. The invention fully utilizes the data measured by the multifunctional convexity meter, establishes a high-precision prediction model of the cross section shape of the strip steel, and simultaneously combines a hot continuous rolling digital twin production line to realize the prediction and display of the cross section shape of the strip steel after virtual rolling. According to the invention, iterative optimization of the set parameters can be realized on the hot-rolling digital twin production line, the defects and degradation phenomena of hot-rolled products caused by unreasonable process parameter setting are avoided, and the production cost of iron and steel enterprises is reduced.
Description
Technical Field
The invention relates to the technical field of digital twin systems, in particular to a method and a device for predicting the cross section shape of hot rolled digital twin strip steel in real time.
Background
The cross-sectional shape of the strip is one of the important indexes for determining the quality of hot rolled strip products. The convexity meter is used for on-line monitoring the cross section shape of the strip steel on site and realizing the real-time feedback control of the cross section shape, but the cross section quality of the strip steel can not be ensured before the roll gap of the rolling mill and under the condition that the feedback control can not be input. In particular for hot continuous rolling, each plate shape adjusting mechanism of the rolling mill should have a correct preset value before the head of the strip steel enters the rolling mill so as to ensure the plate shape of the rolled strip steel before the closed-loop feedback control model is put into operation and serve as a starting point of closed-loop feedback control. The accuracy of the preset control is related to the yield of the strip steel, and meanwhile, the preset value is also an initial value of the feedback control, so that the convergence speed and the accuracy of the plate shape feedback control module for adjusting the plate shape to reach the target value are directly affected.
Disclosure of Invention
The invention provides the hot rolling method for the hot rolling product, which aims at solving the problems of defects and degradation phenomena of the hot rolling product caused by unreasonable process parameter setting.
In order to solve the technical problems, the invention provides the following technical scheme:
in one aspect, the invention provides a method for predicting the cross-sectional shape of a hot rolled digital twin strip steel in real time, which is realized by electronic equipment and comprises the following steps:
s1, constructing a hot continuous rolling digital twin production line, and obtaining set parameters of the hot continuous rolling digital twin production line.
S2, adopting a dynamic modal decomposition DMD algorithm, optimizing a nonlinear system dynamics sparse recognition algorithm SINDY model, and establishing a strip steel cross-section shape prediction model DMD-SINDY.
S3, obtaining a strip steel cross section shape prediction result of the hot continuous rolling digital twin production line according to the set parameters and the strip steel cross section shape prediction model.
Optionally, in S2, a dynamic modal decomposition DMD algorithm is adopted to optimize a nonlinear system dynamics sparse recognition algorithm SINDy model, and a strip steel cross-sectional shape prediction model DMD-SINDy is built, which includes:
s21, collecting historical data in the hot continuous rolling production process, and preprocessing the historical data.
The historical data comprise actual measured rolling force, bending force difference of rolling force at two sides and convexity actual measured values.
S22, establishing a nonlinear system dynamics sparse recognition algorithm SINDY model.
S23, adopting a dynamic modal decomposition DMD algorithm to optimize a nonlinear system dynamics sparse recognition algorithm SINDY model to obtain a strip steel cross-section shape prediction model DMD-SINDY.
Optionally, preprocessing the history data in S21 includes:
the history data is subjected to linear transformation as shown in the following formula (1):
（1）
wherein,representing the linearly transformed data, ++>Representing data before linear transformation ++>Representing the amount of history data.
Optionally, establishing a nonlinear system dynamics sparse recognition algorithm SINDy model in S22 includes:
s221, constructing a space-time matrix according to the preprocessed historical dataAnd sparse function base->。
S222, time space matrixAnd sparse function base->And generating a sparse model through sparse regression.
S223, determining a sparse solution according to a sequential least squares regression methodAccording to the sparse model and the sparse solution +.>And establishing a nonlinear system dynamics sparse recognition algorithm SINDY model.
Optionally, constructing the spatiotemporal matrix in S221Comprising:
carrying out space-time conversion resampling on the preprocessed historical data according to a space sequence, and establishing a space-time matrixAs shown in the following (2):
（2）
wherein,representing spatial sequence,/->Representing the history after preprocessing, +.>Representing the matrix transpose.
Optionally, the sparse model in S222As shown in the following (3):
（3）
wherein,representing a sparse function base, ++>Indicating a lean fluffing.
Optionally, the nonlinear system dynamics sparse recognition algorithm SINDY model in S223As shown in (4) below:
（4）
wherein,representing dynamic constraints defining the equations of motion of the system, +.>Indicating a thin break, a->Representing matrix transpose->Representation->Vector of element sign functions.
Optionally, the optimizing the SINDy model of the nonlinear system dynamics sparse recognition algorithm by adopting a dynamic modal decomposition DMD algorithm in S23 includes:
s231, setting a matrix composed of preprocessed historical data, wherein elements in the matrix comprise first time sequence data vectors evolving along with timeAnd +/with the first time sequence data vector>Second time-series data vector with linear relation +.>And->Wherein->Is a state matrix.
S232, for the first time sequence data vectorPerforming simplified singular value decomposition SVD to obtain state matrix +.>Is defined in the specification.
S233, decomposing SVD and state matrix according to the simplified singular valueAnd obtaining historical data at any time point.
Optionally, historical data at any point in time in S233As shown in (5) below:
（5）
wherein,representing the total number of modalities>Representing eigenvectors>Represents a sparse knob at time point t, +.>The modal amplitude of each mode is shown.
In another aspect, the present invention provides a device for predicting the cross-sectional shape of a hot rolled digital twin strip steel in real time, which is applied to a method for predicting the cross-sectional shape of a hot rolled digital twin strip steel in real time, the device comprising:
the acquisition module is used for constructing a hot continuous rolling digital twin production line and acquiring set parameters of the hot continuous rolling digital twin production line.
And the input module is used for optimizing a nonlinear system dynamics sparse recognition algorithm SINDY model by adopting a dynamic modal decomposition DMD algorithm and establishing a strip steel cross-section shape prediction model DMD-SINDY.
And the output module is used for obtaining a strip steel cross section shape prediction result of the hot continuous rolling digital twin production line according to the set parameters and the strip steel cross section shape prediction model.
Optionally, the input module is further configured to:
s21, collecting historical data in the hot continuous rolling production process, and preprocessing the historical data.
The historical data comprise actual measured rolling force, bending force difference of rolling force at two sides and convexity actual measured values.
S22, establishing a nonlinear system dynamics sparse recognition algorithm SINDY model.
S23, adopting a dynamic modal decomposition DMD algorithm to optimize a nonlinear system dynamics sparse recognition algorithm SINDY model to obtain a strip steel cross-section shape prediction model DMD-SINDY.
Optionally, the input module is further configured to:
the history data is subjected to linear transformation as shown in the following formula (1):
（1）
wherein,representing the linearly transformed data, ++>Representing data before linear transformation ++>Representing the amount of history data.
Optionally, the input module is further configured to:
s221, according to pretreatmentAfter history data, constructing a space-time matrixAnd sparse function base->。
S222, time space matrixAnd sparse function base->And generating a sparse model through sparse regression.
S223, determining a sparse solution according to a sequential least squares regression methodAccording to the sparse model and the sparse solution +.>And establishing a nonlinear system dynamics sparse recognition algorithm SINDY model.
Optionally, the input module is further configured to:
carrying out space-time conversion resampling on the preprocessed historical data according to a space sequence, and establishing a space-time matrixAs shown in the following (2):
（2）
wherein,representing spatial sequence,/->Representing the history after preprocessing, +.>Representing the matrix transpose.
Optionally, a sparse modelAs shown in the following (3):
（3）
wherein,representing a sparse function base, ++>Indicating a lean fluffing.
Alternatively, a nonlinear system dynamics sparse recognition algorithm SINDY modelAs shown in (4) below:
（4）
wherein,representing dynamic constraints defining the equations of motion of the system, +.>Indicating a thin break, a->Representing matrix transpose->Representation->Vector of element sign functions.
Optionally, the input module is further configured to:
s231, set byA matrix of preprocessed historical data, the elements in the matrix comprising first time-ordered data vectors evolving over timeAnd +/with the first time sequence data vector>Second time-series data vector with linear relation +.>And->Wherein->Is a state matrix.
S232, for the first time sequence data vectorPerforming simplified singular value decomposition SVD to obtain state matrix +.>Is defined in the specification.
S233, decomposing SVD and state matrix according to the simplified singular valueAnd obtaining historical data at any time point.
Optionally, historical data at any point in timeAs shown in (5) below:
（5）
wherein,representing the total number of modalities>Representing eigenvectors>Represents a sparse knob at time point t, +.>The modal amplitude of each mode is shown.
In one aspect, an electronic device is provided, the electronic device includes a processor and a memory, the memory stores at least one instruction, and the at least one instruction is loaded and executed by the processor to implement the hot rolled digital twin strip steel cross-sectional shape real-time prediction method.
In one aspect, a computer readable storage medium having stored therein at least one instruction loaded and executed by a processor to implement the above-described method of real-time predicting a cross-sectional shape of a hot rolled digital twin strip steel is provided.
Compared with the prior art, the technical scheme has at least the following beneficial effects:
according to the scheme, in order to obtain the optimal technological parameters of plate shape setting, the high-precision hot rolled strip steel cross section shape prediction model is built, the hot continuous rolling digital twin production line is developed, virtual hot rolling of the strip steel under any technological parameter setting is realized, and the strip steel cross section shape of the strip steel is reflected in real time through the strip steel digital twin model. Based on the method and the control precision requirement of the cross section shape of the strip steel on the production site, the iterative optimization of the process setting parameters can be realized, the defects and degradation phenomena of hot rolled products caused by unreasonable process parameter setting are avoided, and the production cost of steel enterprises is reduced.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method for predicting the cross-sectional shape of a hot rolled digital twin strip steel in real time according to an embodiment of the invention;
FIG. 2 is a sequential threshold least squares regression step diagram provided by an embodiment of the present invention;
FIG. 3 shows the actual rolling results of the cross section of the strip steel provided by the embodiment of the invention;
FIG. 4 is a graph showing the predicted result of SINDY provided by an embodiment of the present invention;
FIG. 5 is a predicted DMD-SINDY result provided by an embodiment of the present invention;
FIG. 6 is a block diagram of a real-time prediction device for the cross-sectional shape of a hot rolled digital twin strip steel provided by an embodiment of the invention;
fig. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. It will be apparent that the described embodiments are some, but not all, embodiments of the invention. All other embodiments, which can be made by a person skilled in the art without creative efforts, based on the described embodiments of the present invention fall within the protection scope of the present invention.
As shown in fig. 1, the embodiment of the invention provides a real-time prediction method for the cross-sectional shape of a hot rolled digital twin strip steel, which can be realized by electronic equipment. The flow chart of the method for predicting the cross-sectional shape of the hot rolled digital twin strip steel in real time as shown in fig. 1 comprises the following steps:
s1, constructing a hot continuous rolling digital twin production line, and obtaining set parameters of the hot continuous rolling digital twin production line.
In a possible implementation mode, a high-precision hot continuous rolling digital twin production line is constructed, and a technological process parameter setting interface is provided.
S2, optimizing a SINDY (Sparse Identification of Nonlinear Dynamics) model by adopting a DMD (Dynamic Mode Decomposition) dynamic modal decomposition algorithm, and establishing a strip steel cross-section shape prediction model DMD-SINDY.
Optionally, the step S2 may include the following steps S21-23:
s21, collecting historical data in the hot continuous rolling production process, and preprocessing the historical data.
The historical data comprise actual measured rolling force, bending force difference of rolling force at two sides and convexity actual measured values.
Furthermore, the collected production data cannot be directly used for modeling, certain preprocessing is needed to remove the data, and the data is distributed to the range of [0,1] by carrying out linear transformation on the one-dimensional column vector sample data.
Specifically, the history data is subjected to linear transformation as shown in the following formula (1):
（1）
wherein,representing the linearly transformed data, ++>Representing data before linear transformation ++>Representing the maximum data in the data sample, +.>Representing the smallest data in the data sample, +.>Representing the amount of history data.
S22, establishing a nonlinear system dynamics sparse recognition algorithm SINDY model.
Optionally, the step S22 may include the following steps S221 to S223:
s221, constructing a space-time matrix according to the preprocessed historical dataAnd sparse function base->。
Wherein a space-time matrix is constructedComprising:
acquisition of time-series data sets and spatial sequence of dataResampling the space-time conversion to establish a space-time matrix>As shown in the following (2):
（2）
wherein,representing the length of the strip.
Further, constructing a sparse function baseIt consists of a number of candidate functions in x columns, as shown in (3) below:
（3）
wherein, byFor example, the state variable +.>The second nonlinear term of (a) is specifically as follows:
（4）
wherein,representing data over a spatial sequence.
S222, constructing a space-time matrix according to the real-time production data collected from the siteAnd input variable matrix +.>. Matrix arrayCombining into dictionary matrix->Generating a sparse model through sparse regression:
（5）
wherein,representing a sparse function base, ++>Indicating a thin break, a->Is +.>Representing sparse vectors of coefficients. If it is determined->The model of each row of control equations may be constructed as follows:
（6）
wherein,is a vector of x element sign functions, different from the data matrix +.>. The overall model can be expressed as follows:
（7）
s223, as shown in FIG. 2, selecting a sequential least squares regression method to determine a sparse solution of the regression problemAnd (3) forcibly setting the weight lower than the threshold value to 0, then carrying out least square on the rest characteristics, iterating for a plurality of times, and further obtaining the final regression solution.
S23, adopting a dynamic modal decomposition DMD algorithm to optimize a nonlinear system dynamics sparse recognition algorithm SINDY model to obtain a strip steel cross-section shape prediction model DMD-SINDY.
Optionally, the step S23 may include the following steps S231 to S233:
s231, setting a matrix composed of preprocessed historical data, wherein elements in the matrix comprise first time sequence data vectors evolving along with timeAnd +/with the first time sequence data vector>Second time-series data vector with linear relation +.>And->Wherein->Is a state matrix.
In one possible embodiment, it is assumed that the matrix of measurement data of the actual hot rolling process is:
（8）
the evolution rule of the system is described by a linear relation:
（9）
the two sets of time-evolving time-series data vectors are respectively:
（10）
（11）
wherein the method comprises the steps ofAnd->Are all +.>：
（12）
For a pair ofSimplified SVD (Singular Value Decomposition ) was performed with:
（13）
wherein,is->Is a data matrix of (a); />Is->Is a diagonal array of (a); />Is->Is a rectangular array of (a) and (b). Order theThe method can obtain:
（14）
thenIs +.>And satisfies:
（15）
wherein,representation sparsenessAnd (5) a knob.
S232, for the first time sequence data vectorSimplified singular value decomposition SVD is carried out to obtain an original system state matrix +.>Modality of (1)>The method comprises the following steps:
（16）
s233, based on the dynamic mode decomposition process, the sampled data at any time point may be expressed as:
（17）
wherein,representing the total number of modalities>Representing eigenvectors>Represents a sparse knob at time point t, +.>The modal amplitude of each mode is shown.
S3, obtaining a strip steel cross section shape prediction result of the hot continuous rolling digital twin production line according to the set parameters and the strip steel cross section shape prediction model.
In a possible implementation mode, manually set parameters of the digital twin production line are input into a strip steel cross-sectional shape prediction model to obtain the cross-sectional shape of the strip steel digital twin model.
Further, the prediction accuracy of the cross-sectional shape of the hot-rolled digital twin strip steel is evaluated.
The evaluation indexes of the prediction precision of the cross section shape of the hot rolling digital twin strip steel specifically comprise:
(1) MAE (Mean Absolute Error ):
（18）
(2) RMSE (Mean Squa re Error, root mean square error):
（19）
wherein,is the number of samples; />Is the expected value; />Is a model predictive value. Where smaller MAE and RMSE indicate higher prediction accuracy of the model.
In the implementation case of a 2250 hot continuous rolling mill, the thickness range of the strip steel of the product line is 1.2-25.4 mm, and the width specification range is 800-2130 mm. The convexity detector at the outlet of the end frame is provided with 60 measuring channels, and the convexity detector can be combined to form the cross-sectional profile of the strip steel through a plurality of strip steel thickness values obtained in each measuring channel. The method of the invention is implemented as follows:
step S1: the collected historical data parameters of the hot rolling production process are shown in table 1, and specific values are shown in table 2. And pre-processes the data in table 2.
TABLE 1
TABLE 2
Step S2: and establishing a strip steel cross-section shape prediction model DMD-SINDY by adopting a dynamic modal decomposition optimization nonlinear system dynamics sparse recognition algorithm.
Step S3: and constructing a high-precision hot continuous rolling digital twin production line and providing a technological process parameter setting interface.
Step S4: taking a certain coil of steel as an example, inputting manual setting parameters of a digital twin production line into a DMD-SINDY model, and predicting to obtain the cross section shape of the strip steel digital twin model. In order to better compare the effectiveness of the method, the actual measured plate shape after rolling with the same technological parameters is selected to be compared with the SINDY predicted plate shape, and the specific results are shown in figures 3-5;
step S5: the prediction accuracy of the cross-sectional shape of the hot-rolled digital twin strip steel is evaluated, and the evaluation results are shown in table 3:
TABLE 3 Table 3
As can be seen from FIGS. 3-5 and Table 3, the method of the invention has the highest prediction precision and can meet the use requirements of the production site. By means of the hot continuous rolling digital twin platform constructed by the invention, virtual simulation rolling can be carried out, and the cross section shape of the strip steel under different hot rolling technological parameter settings can be obtained. The method has important significance for realizing iterative optimization of the set parameters of the rolling process, avoiding the defects and degradation phenomena of the hot rolled products caused by unreasonable process parameter setting and reducing the production cost of iron and steel enterprises.
In the embodiment of the invention, in order to obtain the optimal technological parameters of plate shape setting, the invention establishes a high-precision hot rolled strip steel cross section shape prediction model, develops a hot continuous rolling digital twin production line, realizes the virtual hot rolling of the strip steel under any technological parameter setting, and reflects the strip steel cross section shape in real time through the strip steel digital twin model. Based on the method and the control precision requirement of the cross section shape of the strip steel on the production site, the iterative optimization of the process setting parameters can be realized, the defects and degradation phenomena of hot rolled products caused by unreasonable process parameter setting are avoided, and the production cost of steel enterprises is reduced.
As shown in fig. 6, an embodiment of the present invention provides a real-time prediction apparatus 600 for cross-sectional shape of a hot-rolled digital twin strip steel, the apparatus 600 being applied to implement a real-time prediction method for cross-sectional shape of a hot-rolled digital twin strip steel, the apparatus 600 comprising:
the obtaining module 610 is configured to construct a hot continuous rolling digital twin line, and obtain setting parameters of the hot continuous rolling digital twin line.
And the input module 620 is used for optimizing a nonlinear system dynamics sparse recognition algorithm SINDY model by adopting a dynamic modal decomposition DMD algorithm to establish a strip steel cross-sectional shape prediction model DMD-SINDY.
And the output module 630 is used for obtaining a strip steel cross section shape prediction result of the hot continuous rolling digital twin production line according to the set parameters and the strip steel cross section shape prediction model.
Optionally, the input module 620 is further configured to:
s21, collecting historical data in the hot continuous rolling production process, and preprocessing the historical data.
The historical data comprise actual measured rolling force, bending force difference of rolling force at two sides and convexity actual measured values.
S22, establishing a nonlinear system dynamics sparse recognition algorithm SINDY model.
S23, adopting a dynamic modal decomposition DMD algorithm to optimize a nonlinear system dynamics sparse recognition algorithm SINDY model to obtain a strip steel cross-section shape prediction model DMD-SINDY.
Optionally, the input module 620 is further configured to:
the history data is subjected to linear transformation as shown in the following formula (1):
（1）
wherein the method comprises the steps of，Representing the linearly transformed data, ++>Representing data before linear transformation ++>Representing the amount of history data.
Optionally, the input module 620 is further configured to:
s221, constructing a space-time matrix according to the preprocessed historical dataAnd sparse function base->。
S222, time space matrixAnd sparse function base->And generating a sparse model through sparse regression.
S223, determining a sparse solution according to a sequential least squares regression methodAccording to the sparse model and the sparse solution +.>And establishing a nonlinear system dynamics sparse recognition algorithm SINDY model.
Optionally, the input module 620 is further configured to:
carrying out space-time conversion resampling on the preprocessed historical data according to a space sequence, and establishing a space-time matrixAs shown in the following (2)：
（2）
Wherein,representing spatial sequence,/->Representing the history after preprocessing, +.>Representing the matrix transpose.
Optionally, a sparse modelAs shown in the following (3):
（3）
wherein,representing a sparse function base, ++>Indicating a lean fluffing.
Alternatively, a nonlinear system dynamics sparse recognition algorithm SINDY modelAs shown in (4) below:
（4）
wherein,representing dynamic constraints defining the equations of motion of the system, +.>Indicating a thin break, a->Representing matrix transpose->Representation->Vector of element sign functions.
Optionally, the input module 620 is further configured to:
s231, setting a matrix composed of preprocessed historical data, wherein elements in the matrix comprise first time sequence data vectors evolving along with timeAnd +/with the first time sequence data vector>Second time-series data vector with linear relation +.>And (2) andwherein->Is a state matrix.
S232, for the first time sequence data vectorPerforming simplified singular value decomposition SVD to obtain state matrix +.>Is defined in the specification.
S233, decomposing SVD and state matrix according to the simplified singular valueIs used for the mode of the (c),historical data of any time point is obtained.
Optionally, historical data at any point in timeAs shown in (5) below:
（5）
wherein,representing the total number of modalities>Representing eigenvectors>Represents a sparse knob at time point t, +.>The modal amplitude of each mode is shown.
In the embodiment of the invention, in order to obtain the optimal technological parameters of plate shape setting, the invention establishes a high-precision hot rolled strip steel cross section shape prediction model, develops a hot continuous rolling digital twin production line, realizes the virtual hot rolling of the strip steel under any technological parameter setting, and reflects the strip steel cross section shape in real time through the strip steel digital twin model. Based on the method and the control precision requirement of the cross section shape of the strip steel on the production site, the iterative optimization of the process setting parameters can be realized, the defects and degradation phenomena of hot rolled products caused by unreasonable process parameter setting are avoided, and the production cost of steel enterprises is reduced.
Fig. 7 is a schematic structural diagram of an electronic device 700 according to an embodiment of the present invention, where the electronic device 700 may have a relatively large difference due to different configurations or performances, and may include one or more processors (central processing units, CPU) 701 and one or more memories 702, where at least one instruction is stored in the memories 702, and the at least one instruction is loaded and executed by the processors 701 to implement the following method for predicting the cross-sectional shape of a hot rolled digital twin strip steel in real time:
s1, constructing a hot continuous rolling digital twin production line, and obtaining set parameters of the hot continuous rolling digital twin production line.
S2, adopting a dynamic modal decomposition DMD algorithm, optimizing a nonlinear system dynamics sparse recognition algorithm SINDY model, and establishing a strip steel cross-section shape prediction model DMD-SINDY.
S3, obtaining a strip steel cross section shape prediction result of the hot continuous rolling digital twin production line according to the set parameters and the strip steel cross section shape prediction model.
In an exemplary embodiment, a computer readable storage medium, such as a memory including instructions executable by a processor in a terminal to perform the above-described method of real-time predicting a cross-sectional shape of a hot rolled digital twin strip steel, is also provided. For example, the computer readable storage medium may be ROM, random Access Memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program for instructing relevant hardware, where the program may be stored in a computer readable storage medium, and the storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
The foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and scope of the invention are intended to be included within the scope of the invention.
Claims (3)
1. A method for predicting the cross-sectional shape of a hot rolled digital twin strip steel in real time, which is characterized by comprising the following steps:
s1, constructing a hot continuous rolling digital twin production line, and acquiring set parameters of the hot continuous rolling digital twin production line;
s2, adopting a dynamic modal decomposition DMD algorithm, optimizing a nonlinear system dynamics sparse recognition algorithm SINDY model, and establishing a strip steel cross-section shape prediction model DMD-SINDY;
s3, obtaining a strip steel cross section shape prediction result of the hot continuous rolling digital twin production line according to the set parameters and the strip steel cross section shape prediction model;
in the step S2, a dynamic modal decomposition DMD algorithm is adopted to optimize a nonlinear system dynamics sparse recognition algorithm SINDY model, and a strip steel cross section shape prediction model DMD-SINDY is established, which comprises the following steps:
s21, collecting historical data of a hot continuous rolling production process, and preprocessing the historical data;
the history data comprises actually measured rolling force, bending force difference of rolling force at two sides and actually measured convexity values;
s22, establishing a nonlinear system dynamics sparse recognition algorithm SINDY model;
s23, optimizing a nonlinear system dynamics sparse recognition algorithm SINDY model by adopting a dynamic modal decomposition DMD algorithm to obtain a strip steel cross-section shape prediction model DMD-SINDY;
the step S22 of establishing a nonlinear system dynamics sparse recognition algorithm SINDY model comprises the following steps:
s221, constructing a space-time matrix according to the preprocessed historical dataAnd sparse function base->；
S222, for the space-time matrixAnd sparse function base->Generating a sparse model through sparse regression;
s223, determining a sparse solution according to a sequential least squares regression methodAccording to the sparse model and the sparse solution +.>Establishing a nonlinear system dynamics sparse recognition algorithm SINDY model;
construction of a space-time matrix in S221Comprising:
carrying out space-time conversion resampling on the preprocessed historical data according to a space sequence, and establishing a space-time matrixAs shown in (1) below:
（1）
wherein,representing spatial sequence,/->Representing the history after preprocessing, +.>Representing a matrix transpose;
the sparse model in S222As shown in the following (2):
（2）
wherein,representing a sparse function base, ++>Represents lean fluffing;
the nonlinear system dynamics sparse recognition algorithm SINDY model in the S223As shown in the following (3):
（3）
wherein,representing dynamic constraints defining the equations of motion of the system, +.>Indicating a thin break, a->Representing matrix transpose->Representation ofVector of element sign function;
in the step S23, a dynamic modal decomposition DMD algorithm is adopted to optimize the nonlinear system dynamics sparse recognition algorithm SINDy model, which includes:
s231, setting a matrix composed of preprocessed historical data, wherein elements in the matrix comprise first time sequence data vectors evolving along with timeAnd +/with the first time sequence data vector>Second time-series data vector with linear relation +.>And (2) andwherein->Is a state matrix;
s232, for the first time sequence data vectorPerforming simplified singular value decomposition SVD to obtain state matrix +.>Is a mode of (2);
s233, decomposing SVD and state matrix according to the simplified singular valueThe mode of the system is used for obtaining historical data at any time point;
history data at any time point in the S233As shown in (4) below:
（4）
wherein,representing the total number of modalities>Representing eigenvectors>Indicating a point in timetLower sparse knob, ++>The modal amplitude of each mode is shown.
2. The method according to claim 1, wherein preprocessing the history data in S21 includes:
and (3) performing linear transformation on the historical data, wherein the linear transformation is shown in the following formula (5):
（5）
wherein,representing the linearly transformed data, ++>Representing data before linear transformation ++>Representing the amount of history data.
3. A device for predicting the cross-sectional shape of a hot rolled digital twin strip steel in real time, which is characterized by comprising:
the acquisition module is used for constructing a hot continuous rolling digital twin production line and acquiring set parameters of the hot continuous rolling digital twin production line;
the input module is used for optimizing a nonlinear system dynamics sparse recognition algorithm SINDY model by adopting a dynamic modal decomposition DMD algorithm and establishing a strip steel cross-section shape prediction model DMD-SINDY;
the output module is used for obtaining a strip steel cross section shape prediction result of the hot continuous rolling digital twin production line according to the set parameters and the strip steel cross section shape prediction model;
the method for establishing the strip steel cross-section shape prediction model DMD-SINDY by adopting a dynamic modal decomposition DMD algorithm and optimizing a nonlinear system dynamics sparse recognition algorithm SINDY model comprises the following steps:
s21, collecting historical data of a hot continuous rolling production process, and preprocessing the historical data;
the history data comprises actually measured rolling force, bending force difference of rolling force at two sides and actually measured convexity values;
s22, establishing a nonlinear system dynamics sparse recognition algorithm SINDY model;
s23, optimizing a nonlinear system dynamics sparse recognition algorithm SINDY model by adopting a dynamic modal decomposition DMD algorithm to obtain a strip steel cross-section shape prediction model DMD-SINDY;
the step S22 of establishing a nonlinear system dynamics sparse recognition algorithm SINDY model comprises the following steps:
s221, constructing a space-time matrix according to the preprocessed historical dataAnd sparse function base->；
S222, for the space-time matrixAnd sparse function base->Generating a sparse model through sparse regression;
s223, determining a sparse solution according to a sequential least squares regression methodAccording to the sparse model and the sparse solution +.>Establishing a nonlinear system dynamics sparse recognition algorithm SINDY model;
construction of a space-time matrix in S221Comprising:
carrying out space-time conversion resampling on the preprocessed historical data according to a space sequence, and establishing a space-time matrixAs shown in (1) below:
（1）
wherein,representing spatial sequence,/->Representing the history after preprocessing, +.>Representing a matrix transpose;
the sparse model in S222As shown in the following (2):
（2）
wherein,representing a sparse function base, ++>Represents lean fluffing;
the nonlinear system dynamics sparse recognition algorithm SINDY model in the S223As shown in the following (3):
（3）
wherein,representing dynamic constraints defining the equations of motion of the system, +.>Indicating a thin break, a->Representing matrix transpose->Representation ofVector of element sign function;
in the step S23, a dynamic modal decomposition DMD algorithm is adopted to optimize the nonlinear system dynamics sparse recognition algorithm SINDy model, which includes:
s231, setting a matrix composed of preprocessed historical data, wherein elements in the matrix comprise first time sequence data vectors evolving along with timeAnd +/with the first time sequence data vector>Second time-series data vector with linear relation +.>And (2) andwherein->Is a state matrix;
s232, for the first time sequence data vectorPerforming simplified singular value decomposition SVD to obtain state matrix +.>Is a mode of (2);
s233, decomposing SVD and state matrix according to the simplified singular valueThe mode of the system is used for obtaining historical data at any time point;
history data at any time point in the S233As shown in (4) below:
（4）
wherein,representing the total number of modalities>Representing eigenvectors>Indicating a point in timetLower sparse knob, ++>The modal amplitude of each mode is shown.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311436735.4A CN117150832B (en) | 2023-11-01 | 2023-11-01 | Real-time prediction method and device for cross section shape of hot-rolled digital twin strip steel |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311436735.4A CN117150832B (en) | 2023-11-01 | 2023-11-01 | Real-time prediction method and device for cross section shape of hot-rolled digital twin strip steel |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117150832A CN117150832A (en) | 2023-12-01 |
| CN117150832B true CN117150832B (en) | 2024-02-23 |
Family
ID=88899249
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311436735.4A Active CN117150832B (en) | 2023-11-01 | 2023-11-01 | Real-time prediction method and device for cross section shape of hot-rolled digital twin strip steel |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117150832B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113333474A (en) * | 2021-04-27 | 2021-09-03 | 山东钢铁股份有限公司 | Strip steel hot-rolled plate shape control method and system based on digital twinning |
| CN113343606A (en) * | 2021-07-01 | 2021-09-03 | 浙江大学 | Method for predicting full flow field from sparse sensor information based on compressed sensing reduced order model |
| CN113504768A (en) * | 2021-08-05 | 2021-10-15 | 东华大学 | High-precision product digital twin computability method for assembly quality prediction |
| WO2022121923A1 (en) * | 2020-12-10 | 2022-06-16 | 东北大学 | Smart modelling method and apparatus of complex industrial process digital twin system, device, and storage medium |
| CN114647904A (en) * | 2022-02-24 | 2022-06-21 | 北京工业大学 | Self-adaptive sparse graph learning bearing life prediction method based on digital twin dictionary |
| CN115034152A (en) * | 2022-05-17 | 2022-09-09 | 浙江大学 | Data-driven fluid-solid coupling system nonlinear order reduction prediction method and device |
| CN115796002A (en) * | 2022-11-01 | 2023-03-14 | 北京科技大学 | Strip steel coil defect prediction and diagnosis method based on deep learning |
| CN116020879A (en) * | 2023-02-15 | 2023-04-28 | 北京科技大学 | Technological parameter-oriented strip steel hot continuous rolling space-time multi-scale process monitoring method and device |
-
2023
- 2023-11-01 CN CN202311436735.4A patent/CN117150832B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022121923A1 (en) * | 2020-12-10 | 2022-06-16 | 东北大学 | Smart modelling method and apparatus of complex industrial process digital twin system, device, and storage medium |
| CN113333474A (en) * | 2021-04-27 | 2021-09-03 | 山东钢铁股份有限公司 | Strip steel hot-rolled plate shape control method and system based on digital twinning |
| CN113343606A (en) * | 2021-07-01 | 2021-09-03 | 浙江大学 | Method for predicting full flow field from sparse sensor information based on compressed sensing reduced order model |
| CN113504768A (en) * | 2021-08-05 | 2021-10-15 | 东华大学 | High-precision product digital twin computability method for assembly quality prediction |
| CN114647904A (en) * | 2022-02-24 | 2022-06-21 | 北京工业大学 | Self-adaptive sparse graph learning bearing life prediction method based on digital twin dictionary |
| CN115034152A (en) * | 2022-05-17 | 2022-09-09 | 浙江大学 | Data-driven fluid-solid coupling system nonlinear order reduction prediction method and device |
| CN115796002A (en) * | 2022-11-01 | 2023-03-14 | 北京科技大学 | Strip steel coil defect prediction and diagnosis method based on deep learning |
| CN116020879A (en) * | 2023-02-15 | 2023-04-28 | 北京科技大学 | Technological parameter-oriented strip steel hot continuous rolling space-time multi-scale process monitoring method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117150832A (en) | 2023-12-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111241657B (en) | Cold-rolled strip steel plate shape prediction method based on deep learning | |
| JP5751045B2 (en) | Plant operating condition optimization system, plant operating condition optimization method, plant operating condition optimization program | |
| US8050792B2 (en) | Method and device for optimization of flatness control in the rolling of a strip | |
| Sun et al. | Data-based flatness prediction and optimization in tandem cold rolling | |
| Dong et al. | Prediction and analysis of key parameters of head deformation of hot-rolled plates based on artificial neural networks | |
| Song et al. | BUAK-AIS: Efficient Bayesian updating with active learning Kriging-based adaptive importance sampling | |
| CN111898247B (en) | Landslide displacement prediction method, landslide displacement prediction equipment and storage medium | |
| CN115841004B (en) | Strip steel hot rolling process mechanical property soft measurement method and device based on multidimensional data | |
| JP4932294B2 (en) | Manufacturing specification determination support system, manufacturing specification determination support method, computer program, and computer-readable recording medium | |
| WO2010008944A2 (en) | Systems and methods for offline and/or online batch monitoring using decomposition and signal approximation approaches | |
| JP5962290B2 (en) | Steel heat transfer coefficient prediction apparatus and cooling control method | |
| CN117150832B (en) | Real-time prediction method and device for cross section shape of hot-rolled digital twin strip steel | |
| Zhao et al. | A novel paradigm of flatness prediction and optimization for strip tandem cold rolling by cloud-edge collaboration | |
| CN112307638B (en) | Capacitor life estimation method and device and electronic equipment | |
| CN113642886A (en) | Method, system, medium and computing device for analyzing utilization rate of planned power grid line | |
| Lee et al. | Integrated model for thermo-mechanical controlled process in rod (or bar) rolling | |
| Lu et al. | Data mining-based flatness pattern prediction for cold rolling process with varying operating condition | |
| Tulupov et al. | Simulation of bar rolling: experience at Nosov Magnitogorsk state technical university | |
| Xue et al. | A novel ensemble of metamodels using improved stepwise metamodel selection and two-layer pointwise ensemble | |
| Duhler et al. | Fast Multi-Order Stochastic Subspace Identification⋆ | |
| Höskuldsson et al. | Path modeling and process control | |
| JPH0895948A (en) | Method and device for time sequential prediction based on trend | |
| Zhu-wen et al. | Optimization and innovative modification of a model used to determine the amount of adjustment of an actuator for flatness control | |
| CN117113520B (en) | Displacement field loading-based transformer substation sound barrier detection method and system | |
| Xiong et al. | A penalized approach to mixed model selection via cross-validation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |