CN113850020A - Continuous annealing tension setting method and device based on neural network - Google Patents
Continuous annealing tension setting method and device based on neural network Download PDFInfo
- Publication number
- CN113850020A CN113850020A CN202111089354.4A CN202111089354A CN113850020A CN 113850020 A CN113850020 A CN 113850020A CN 202111089354 A CN202111089354 A CN 202111089354A CN 113850020 A CN113850020 A CN 113850020A
- Authority
- CN
- China
- Prior art keywords
- tension
- strip steel
- value
- steel
- strip
- 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.)
- Granted
Links
- 238000000137 annealing Methods 0.000 title claims abstract description 69
- 238000000034 method Methods 0.000 title claims abstract description 68
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 29
- 239000010959 steel Substances 0.000 claims abstract description 278
- 229910000831 Steel Inorganic materials 0.000 claims abstract description 275
- 238000004519 manufacturing process Methods 0.000 claims abstract description 138
- 238000003062 neural network model Methods 0.000 claims abstract description 42
- 238000012549 training Methods 0.000 claims abstract description 25
- 239000002436 steel type Substances 0.000 claims abstract description 18
- 230000008569 process Effects 0.000 claims abstract description 14
- 238000007781 pre-processing Methods 0.000 claims abstract description 10
- 238000004364 calculation method Methods 0.000 claims description 18
- 238000012545 processing Methods 0.000 claims description 12
- 238000010606 normalization Methods 0.000 claims description 10
- 238000005097 cold rolling Methods 0.000 claims description 6
- 239000002994 raw material Substances 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 4
- 230000010354 integration Effects 0.000 claims description 2
- 230000003068 static effect Effects 0.000 abstract description 4
- 230000008859 change Effects 0.000 abstract description 3
- 238000010438 heat treatment Methods 0.000 description 12
- 238000010586 diagram Methods 0.000 description 9
- 238000001816 cooling Methods 0.000 description 8
- 230000015654 memory Effects 0.000 description 8
- 238000004590 computer program Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 7
- 238000002791 soaking Methods 0.000 description 4
- 238000003860 storage Methods 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 3
- 230000009471 action Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 229910001209 Low-carbon steel Inorganic materials 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 239000010960 cold rolled steel Substances 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000006748 scratching Methods 0.000 description 1
- 230000002393 scratching effect Effects 0.000 description 1
- 238000005482 strain hardening Methods 0.000 description 1
Images
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
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Strategic Management (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Molecular Biology (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Pure & Applied Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Educational Administration (AREA)
- Development Economics (AREA)
- Mathematical Optimization (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Computing Systems (AREA)
- Computer Hardware Design (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Medical Informatics (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
Abstract
The invention discloses a continuous annealing tension setting method and a continuous annealing tension setting device based on a neural network, wherein the method comprises the following steps: acquiring historical production data of the strip steel; classifying the steel type, the thickness of the strip steel and the width of the strip steel according to a preset classification rule, and calculating the average tension value of each tension section and the strip shape characteristic value of the cold-rolled strip steel corresponding to each strip steel; preprocessing the data to construct a data set; constructing a neural network model and training by adopting the data set; obtaining tension set values of all sections in current production by using a trained neural network model; and calculating to obtain a final tension set value by using a weighted sliding average method based on the tension value obtained by the model and the historical production tension value. The invention can quickly and automatically set the static tension of each section in the continuous annealing process, automatically correct the static tension along with the change of the working condition of the continuous annealing furnace, and effectively improve the production stability, the production efficiency and the product quality of the continuous annealing.
Description
Technical Field
The invention relates to the technical field of strip steel production, in particular to a continuous annealing tension setting method and device based on a neural network.
Background
The cold-rolled steel strip needs to be annealed before finishing to achieve the purposes of eliminating work hardening and improving mechanical properties. The tension control of the strip steel plays a key role in controlling the stable through plate of the unit and the scratch on the surface of the strip steel. The current continuous annealing tension setting mainly adopts a meter reading mode, data in the table is determined by an empirical formula, so that the tension of the strip steel in the continuous annealing furnace is not matched with the working condition, the quality and the efficiency of a produced product are influenced, and even accidents such as deviation can occur in severe cases. Therefore, it is necessary to develop a new method for setting the continuous annealing tension.
Disclosure of Invention
The invention provides a continuous annealing tension setting method and device based on a neural network, and aims to solve the technical problems that the tension of strip steel in a continuous annealing furnace is not matched with the working condition by the conventional continuous annealing tension setting method, so that the quality and the efficiency of a produced product are influenced, and even deviation and other accidents occur in serious cases.
In order to solve the technical problems, the invention provides the following technical scheme:
in one aspect, the present invention provides a method for setting continuous annealing tension based on a neural network, including:
acquiring historical production data of the strip steel; wherein the historical production data of the strip steel comprises: the steel type, the width, the thickness and the shape value of each cold rolling channel plate of the raw material strip steel, and the speed of each tension section set furnace roller, the production speed of the strip steel and the tension set value of each tension section during continuous annealing production;
grading the steel type, the strip steel thickness and the strip steel width in the strip steel historical production data according to a preset grading rule to obtain a steel type grading value, a strip steel thickness grading value and a strip steel width grading value corresponding to each strip steel; calculating the tension setting average value of each tension section corresponding to each strip steel and the strip shape characteristic value of the cold-rolled strip steel;
preprocessing a steel grade grading value, a strip steel thickness grading value, a strip steel width grading value, a cold-rolled strip steel plate shape characteristic value, a set furnace roller speed of each tension section, a strip steel production speed and a set average value of tension of each tension section, and constructing a data set for model training by using preprocessed data;
constructing a neural network model and training by adopting the data set; the neural network model takes a steel grade grading value, a strip steel thickness grading value, a strip steel width grading value, a cold-rolled strip steel plate shape characteristic value, a furnace roller speed set for each tension section and a strip steel production speed as input, and a tension average value set for each tension section as output;
obtaining tension set values of all sections in the continuous annealing furnace in the current production by using the trained neural network model;
and calculating to obtain a final tension set value by using a weighted sliding average method based on the tension set values of all the sections obtained by the neural network model and the tension set average value of all the tension sections in the historical production of the strip steel.
Further, the step of classifying the steel type, the strip steel thickness and the strip steel width in the strip steel historical production data according to a preset classification rule to obtain a steel type classification value, a strip steel thickness classification value and a strip steel width classification value corresponding to each strip steel includes:
dividing strip steels with different steel grades into four steel families and a default classification according to the steel quality; dividing the band steel width grading range according to the production frequency of each band steel width interval; wherein, the higher the production frequency is, the finer the corresponding width interval is graded; according to the thickness of the strip steel, uniformly dividing the thickness of the strip steel into grading ranges from thin to thick;
and numbering the graded strip steel according to the steel grade grading result, the strip steel width grading result and the strip steel thickness grading result, and taking the steel grade grading value, the strip steel thickness grading value and the strip steel width grading value corresponding to the strip steel.
Further, the calculating the set average value of the tension of each tension section corresponding to each strip steel includes:
and calculating the average value of the tension of each section of the strip steel during the production of each coil of the strip steel in the time range from the strip head of each strip steel entering the continuous annealing furnace area to the strip tail leaving the continuous annealing furnace area, and taking the average value as the set average value of the tension of each tension section corresponding to the strip steel.
Further, the method for calculating the strip shape characteristic value of the cold-rolled strip steel comprises the following steps:
respectively integrating the strip shape primary term coefficient curves corresponding to 50m sections at the head and the tail of the strip steel and three sections in the middle of the strip steel, and carrying out weighted summation on the integration results of the three sections to obtain a strip shape characteristic value I of the cold-rolled strip steel, wherein the formula is as follows:
wherein l represents the length of the strip, a1(l) Is a 50m plate-shaped first-order coefficient curve of the head of the strip steel, a2(l) Is a 50m plate-shaped first-order coefficient curve a of the tail part of the strip steel3(l) Is a primary coefficient curve of the strip shape in the middle of the strip steel.
Further, the preprocessing comprises normalization processing, noise point removal and abnormal data removal.
Further, the normalization processing is maximum and minimum normalization processing.
Further, the calculating of the tension set values obtained based on the neural network model and the tension set average value of each tension section in the historical production of the strip steel by using a weighted sliding average method to obtain the final tension set value includes:
according to a first-order moving average model, each section of tension set value T obtained by calculation of a neural network modelnewSetting average value T of tension of each tension section in historical productionoldWeighted summation is carried out to obtain the final tension set value of each tension section of each specification strip steel
The formula is as follows:
where γ is a weight value of the weighted sum.
Further, the historical production data of the strip steel further comprises: historical average production speed of the production team, average strip steel quality rating of the production team and average accident rate of the production team;
γ is calculated by the following formula:
wherein, VaFor the current team historical average production speed, VmaxAverage production speed, Q, of teams of highest technical levelaAnd grading the average quality of the strip steel produced by the current team, wherein eta is the average accident rate of the current team.
Further, after the final tension set value is obtained through calculation, the continuous annealing tension setting method further comprises the following steps:
and taking the final tension set value obtained by calculation as a tension control value of actual production, and putting the final tension set value obtained by calculation into a training set of the neural network model for self-learning training.
On the other hand, the invention also provides a continuous annealing tension setting device based on the neural network, which comprises the following components:
the data acquisition module is used for acquiring historical production data of the strip steel; wherein the historical production data of the strip steel comprises: the steel type, the width, the thickness and the shape value of each cold rolling channel plate of the raw material strip steel, and the speed of each tension section set furnace roller, the production speed of the strip steel and the tension set value of each tension section during continuous annealing production;
the data grading and characteristic value calculating module is used for grading the steel grade, the strip steel thickness and the strip steel width in the strip steel historical production data acquired by the data acquiring module according to a preset grading rule to obtain a steel grade grading value, a strip steel thickness grading value and a strip steel width grading value corresponding to each strip steel; calculating the tension setting average value of each tension section corresponding to each strip steel and the strip shape characteristic value of the cold-rolled strip steel;
the data set construction module is used for preprocessing a steel grade grading value, a strip steel thickness grading value, a strip steel width grading value, a cold-rolled strip steel plate characteristic value, a set furnace roller speed of each tension section, a strip steel production speed and a set average tension value of each tension section, and constructing a data set for model training by using the preprocessed data;
the neural network model building and training module is used for building a neural network model and training by adopting the data set built by the data set building module; the neural network model takes a steel grade grading value, a strip steel thickness grading value, a strip steel width grading value, a cold-rolled strip steel plate shape characteristic value, a furnace roller speed set for each tension section and a strip steel production speed as input, and a tension average value set for each tension section as output;
the tension calculation module is used for obtaining tension set values of all sections in the continuous annealing furnace in the current production by utilizing the trained neural network model; and calculating to obtain a final tension set value by using a weighted sliding average method based on the tension set values of all the sections obtained by the neural network model and the tension set average value of all the tension sections in the historical production of the strip steel.
In yet another aspect, the present invention also provides an electronic device comprising a processor and a memory; wherein the memory has stored therein at least one instruction that is loaded and executed by the processor to implement the above-described method.
In yet another aspect, the present invention also provides a computer-readable storage medium having at least one instruction stored therein, the instruction being loaded and executed by a processor to implement the above method.
The technical scheme provided by the invention has the beneficial effects that at least:
aiming at the actual production situation that an operator can frequently adjust the tension set values of the strip steels with different specifications according to the production situation because the production situation has more disturbance and uncertain factors in the field actual production process, the invention divides the specifications of the strip steels by combining the field experience and the production frequency of the strip steels with different specifications in a longer time, learns the tension set values again by using a neural network algorithm and adjusts the self-learning coefficients according to the production capacity of a team to improve the model precision in order to learn the field experience and provide a better reference for the production operation in combination with the production line. Therefore, the technical effect of quickly and automatically setting the static tension of each section in the continuous annealing process is realized, and automatic correction is carried out along with the change of the working condition of the continuous annealing furnace, so that the continuous annealing production stability, the production efficiency and the product quality can be effectively 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 description of the embodiments will be briefly introduced 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 creative efforts.
Fig. 1 is a flowchart of a method for setting a continuous annealing tension based on a neural network according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a neural network model provided in an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
First embodiment
The embodiment provides a continuous annealing tension setting method based on a neural network, which is characterized in that tension of each section in a continuous annealing furnace is calculated and set by establishing a neural network model, and the model is continuously optimized through self-learning. The method may be implemented by an electronic device. The execution flow of the method is shown in fig. 1, and comprises the following steps:
and S1, acquiring historical production data of the strip steel.
Specifically, in this embodiment, the step S1 is specifically: acquiring quality data of cold-rolled strip steel, continuous annealing production historical data and the condition of a production team on the day from a cold-rolling and continuous annealing data acquisition system; wherein, cold-rolled strip steel quality data includes: the steel type, the width and the thickness of the strip steel of the raw material strip steel and the shape value of each cold-rolled channel plate; the continuous annealing production historical data comprises: setting furnace roller speed, strip steel production speed and actual tension set values of all sections at each tension section during continuous annealing production; production team conditions include: historical average production speed of a production team, average strip quality rating of production of the production team, and average accident rate of the production team.
S2, classifying steel types, strip steel thicknesses and strip steel widths in strip steel historical production data according to preset classification rules to obtain steel type classification values, strip steel thickness classification values and strip steel width classification values corresponding to the strip steels; and calculating the tension setting average value of each tension section corresponding to each strip steel and the strip shape characteristic value of the cold-rolled strip steel.
Specifically, in the present embodiment, the purpose of S2 is to divide and number the steel grade, thickness, and width of the continuous annealing production according to the historical production frequency; and calculating the tension setting average value of each tension section and the strip shape characteristic value of the cold-rolled strip steel in the period from the charging to the discharging of the historical production coil, wherein the implementation process comprises the following steps:
and calculating the average value of the tension setting of each section of the strip steel at each gear. The method specifically comprises the following steps: and calculating the mean value of the tension of each section when each roll of strip steel is produced according to the time range from the strip head of each gear of strip steel entering the continuous annealing furnace area to the strip tail leaving the continuous annealing furnace area. Taking a certain historical production roll as an example, the production roll has the roll number H1110043000000, the thickness of 1.183mm, the width of 1283mm, and the steel grade of DC 04. The average tension of each stage of the historical production is shown in table 1.
TABLE 1 tension average tension value of each section of historical production
| Name of tension section | Tension average value/kN | Name of tension section | Tension average value/kN |
| Tension of inlet section | 14.01 | Over-aged I-stage tension | 9.45 |
| Heating tension of section I | 10.01 | Over-aged II section tension | 9.44 |
| Heating tension in section II | 10.03 | Over-aged III-stage tension | 9.26 |
| Heating tension in stage III | 9.05 | Tension of final cooling section | 10.15 |
| Tension of soaking section | 9.12 | Tension of outlet section | 14.00 |
| Tension of fast cooling section | 10.99 |
The strip steels with different steel grades are divided into four steel families and a default classification according to the steel quality. Including IF steel, mild steel, HSS steel (tensile strength less than 590MPa) and HSS steel (tensile strength greater than or equal to 590MPa), corresponding to four steel families of TENS-0,. TENS-1,. TENS-2 and. TENS-3, respectively, as shown in Table 2.
TABLE 2 Steel family division rules
Due to the fact that omission occurs in steel grade counting, steel grades missing in a table can be automatically recorded in the TENS-4 default classification during production, and the steel grades in the TENS-4 need to be deleted regularly and added to the corresponding classification.
Obtaining the number records of the production rolls in each width interval by counting the actual secondary tension setting data of each roll of steel production in a past period of time, and dividing the width grading range according to the production frequency of each strip steel width interval; wherein, the width interval with higher production frequency is divided into more fine grades; the thickness grading is divided uniformly from thin to thick to obtain the grading rule of thickness and width as shown in table 3.
TABLE 3 Width and thickness grading rules
The specification class of the strip steel can be determined through steel type-thickness grading-width grading, each specification is numbered on the basis, and the numbering naming mode is steel type number + thickness number + width number; wherein, the steel grade family is numbered from 0 to 4, corresponding to tension-0 to tension-4, and the thickness is numbered as follows: the (0, 0.6) grade number is 06, the (1.0, 1.5) grade number is 15, the width number is from 1 to 6 according to the grade interval from narrow to wide, taking a certain strip steel as an example, the steel species is tension-1, the thickness grade is (0, 0.6), the width grade is (0,1200), and the strip steel specification number is 1061.
After grading, the steel family number of the current coil strip steel is 0, the thickness grading is (1.0, 1.5), the thickness grading number is 15, and the width grading is 2, so that the specification number of the coil strip steel is 0152.
And performing secondary fitting calculation on data of each channel of the cold-rolled strip shape instrument, wherein because the strip steel with poor head and tail strip shape quality is more prone to deviation accidents in a continuous annealing furnace, when calculating the strip shape characteristic value of the cold-rolled strip steel, the first-order coefficient curve integrals respectively corresponding to 50m sections of the head and the tail of the strip steel and three sections of the middle of the strip steel are subjected to weighted summation, and the strip shape characteristic value I can be represented through the integral values. The specific formula is as follows:
wherein l represents the length of the strip, a1(l) Is a 50m plate-shaped first-order coefficient curve of the head of the strip steel, a2(l) Is a 50m plate-shaped first-order coefficient curve a of the tail part of the strip steel3(l) Is a primary coefficient curve of the strip shape in the middle of the strip steel.
Taking a certain roll of strip steel as an example, the characteristic value of the head plate shape is 3.41, the characteristic value of the middle plate shape is 1.61, and the characteristic value of the tail plate shape is 1.68, and the weighted characteristic value of the plate shape is 2.358 after calculation.
And S3, preprocessing the steel grade grading value, the strip steel thickness grading value, the strip steel width grading value, the cold-rolled strip steel shape characteristic value, the set furnace roller speed of each tension section, the strip steel production speed and the set average value of the tension of each tension section, and constructing a data set for model training by using the preprocessed data.
Specifically, in this embodiment, the above-mentioned purpose of S3 is to establish a database with the steel grade value, the strip thickness grade value, the strip width grade value, the cold-rolled strip shape characteristic value, the set furnace roller speed of each tension section, the strip production speed and the set average value of each tension section, and preprocess the data therein to make it more suitable for model training. The preprocessing process of the data comprises normalization processing, noise point removal and abnormal data removal, wherein the normalization adopts maximum and minimum normalization processing. The method can reduce the influence of different dimensions on the prediction precision, and the formula is as follows:
wherein, max { xjDenotes the largest data in the dataset, min { x }jDenotes the smallest data in the data set. y isiThe results after the normalization process.
Further, in this embodiment, the processing of removing noise and abnormal data is performed manually, and the purpose is to remove data of deviation accidents during production.
S4, constructing a neural network model and training by adopting the data set; the neural network model takes a steel grade grading value, a strip steel thickness grading value, a strip steel width grading value, a cold-rolled strip steel plate characteristic value, a furnace roller speed set for each tension section and a strip steel production speed as input, and a tension setting average value of each tension section as output.
Specifically, in this embodiment, the neural network model used is a BP neural network model, and the structure thereof is shown in fig. 2. The construction process of the BP neural network model is as follows:
determining the number of neural network layers, the number of input and output nodes, and calculating the number of hidden layer neuron nodes; specifically, the present embodiment is determined by using an empirical formula, where the empirical formula is:
wherein h is the number of hidden layer nodes, m is the number of input layer nodes, n is the number of output layer nodes, and a is an adjusting constant between 1 and 10.
Initializing the BP neural network, including determining a weight interval, a calculation precision value and a learning rate;
the process of model training is as follows: data is transmitted from the input layer to the output layer after passing through the hidden layer, if the result is not consistent, the error signal is transmitted in reverse direction, and the weight of each unit is corrected by the error; and continuously circulating until the output error is smaller than the set precision value or the learning times are reached.
Specifically, in this embodiment, a general Sigmoid function is selected as an activation function, the number of input nodes is 6, the number of output nodes is 1, and the number of hidden layer nodes is 10. The weight adjustment rate of the BP neural network is 0.003, the threshold adjustment rate is 0.008, the maximum training frequency is 600, the allowed error of a single sample is 0.001, and the allowed error of each iteration is 0.005.
And S5, obtaining tension set values of all sections in the continuous annealing furnace in the current production by using the trained neural network model.
After the BP neural network is trained, the parameters of the strip steel with the same specification are substituted, and the obtained results are shown in the following table, taking the characteristic value of the plate shape as 3.415 as an example.
TABLE 4 prediction of average tension value of each tension
| Name of tension section | Tension average value/kN | Name of tension section | Tension average value/kN |
| Tension of inlet section | 14.58 | Over-aged I-stage tension | 9.78 |
| Heating tension of section I | 12.63 | Over-aged II section tension | 9.05 |
| Heating tension in section II | 10.82 | Over-aged III-stage tension | 9.35 |
| Heating tension in stage III | 9.87 | Tension of final cooling section | 12.13 |
| Tension of soaking section | 9.55 | Tension of outlet section | 15.02 |
| Tension of fast cooling section | 13.65 |
And S6, calculating to obtain the final tension set value by using a weighted sliding average method based on the tension set values of all the sections obtained by the neural network model and the tension set average value of all the tension sections in the historical production of the strip steel.
Specifically, in this embodiment, the implementation process of S6 is as follows:
according to a first-order moving average model, each section of tension set value T obtained by calculation of a neural network modelnewSetting average value T of tension of each tension section in historical productionoldWeighting and summing according to the weight gamma to obtain the final tension set value of each tension section of the strip steel with each specification
The formula is as follows:
wherein the weight gamma is expressed as a self-learning speed and is determined by the production level of the team. The production level of the team has influence on the continuous annealing production stability, when the technical level of operators is higher, the tension value manually adjusted is more suitable for the working condition in the furnace, the quality of the produced strip steel is higher, and the self-learning coefficient is higher at the moment. Therefore, in the data processing, according to the historical production conditions of different teams, the production level of the teams is evaluated by taking the production speed, the average production strip steel quality and the accident rate as indexes. The technical level calculation formula of the team is as follows:
wherein the weight gamma is the team skillOperative level, VaFor the current team historical average production speed, VmaxAverage production speed, Q, of teams of highest technical levelaAverage strip quality rating for the current team production, determined from the following table, 0<Qa<1, eta is the average accident rate of the current team.
TABLE 5 average production strip quality rating
Taking a certain team as an example, the ratio of the speed to the highest production speed is 80%, the average production strip steel quality is rated as 0.85, the accident rate is 1%, and the finally determined weight is 0.40.
Specifically, in the present embodiment, there are 11 tension sections in the continuous annealing furnace, wherein each section calculates the tension set value by the above method. The tension section includes: a furnace inlet tension roller, a heating I section, a heating II section, a heating III section, a soaking section, a fast cooling section, an overaging I section, an overaging II section, an overaging III section, a final cooling section and a furnace outlet tension roller. The final tension of the strip is shown in the following table.
TABLE 6 setting of static tension values for each stage
| Name of tension section | Tension average value/kN | Name of tension section | Tension average value/kN |
| Tension of inlet section | 14.24 | Over-aged I-stage tension | 9.58 |
| Heating tension of section I | 11.06 | Over-aged II section tension | 9.28 |
| Heating tension in section II | 10.33 | Over-aged III-stage tension | 9.30 |
| Heating tension in stage III | 9.35 | Tension of final cooling section | 10.95 |
| Tension of soaking section | 9.22 | Tension of outlet section | 14.41 |
| Tension of fast cooling section | 12.06 |
Further, after the final tension set value is obtained through calculation, the continuous annealing tension setting method further comprises the following steps:
and S7, taking the final tension set value obtained by calculation as a tension control value of actual production, putting the final tension set value obtained by calculation into a training set of the neural network model, and carrying out self-learning training.
After the continuous annealing tension setting method based on the neural network provided by the embodiment is adopted, the strip steel deviation number of a continuous annealing furnace of a production line 2130 of a certain factory is reduced by 5%, the yield is improved by 12% compared with the yield in the same period, and the surface scratching condition of the strip steel is obviously reduced. Therefore, the continuous annealing tension setting method can be proved to be capable of self-learning along with the change of the working condition in the continuous annealing furnace, and the stability, efficiency and product quality of continuous annealing production are effectively improved.
Second embodiment
The embodiment provides a continuous annealing tension setting device based on a neural network, which comprises:
the data acquisition module is used for acquiring historical production data of the strip steel; wherein the historical production data of the strip steel comprises: the steel type, the width, the thickness and the shape value of each cold rolling channel plate of the raw material strip steel, and the speed of each tension section set furnace roller, the production speed of the strip steel and the tension set value of each tension section during continuous annealing production;
the data grading and characteristic value calculating module is used for grading the steel grade, the strip steel thickness and the strip steel width in the strip steel historical production data acquired by the data acquiring module according to a preset grading rule to obtain a steel grade grading value, a strip steel thickness grading value and a strip steel width grading value corresponding to each strip steel; calculating the tension setting average value of each tension section corresponding to each strip steel and the strip shape characteristic value of the cold-rolled strip steel;
the data set construction module is used for preprocessing a steel grade grading value, a strip steel thickness grading value, a strip steel width grading value, a cold-rolled strip steel plate characteristic value, a set furnace roller speed of each tension section, a strip steel production speed and a set average tension value of each tension section, and constructing a data set for model training by using the preprocessed data;
the neural network model building and training module is used for building a neural network model and training by adopting the data set built by the data set building module; the neural network model takes a steel grade grading value, a strip steel thickness grading value, a strip steel width grading value, a cold-rolled strip steel plate shape characteristic value, a furnace roller speed set for each tension section and a strip steel production speed as input, and a tension average value set for each tension section as output;
the tension calculation module is used for obtaining tension set values of all sections in the continuous annealing furnace in the current production by utilizing the trained neural network model; and calculating to obtain a final tension set value by using a weighted sliding average method based on the tension set values of all the sections obtained by the neural network model and the tension set average value of all the tension sections in the historical production of the strip steel.
The continuous annealing tension setting device based on the neural network of the present embodiment corresponds to the continuous annealing tension setting method based on the neural network of the first embodiment; the functions realized by the functional modules in the continuous annealing tension setting device based on the neural network of the present embodiment correspond to the flow steps in the continuous annealing tension setting method based on the neural network of the first embodiment one by one; therefore, it is not described herein.
Third embodiment
The present embodiment provides an electronic device, which includes a processor and a memory; wherein the memory has stored therein at least one instruction that is loaded and executed by the processor to implement the method of the first embodiment.
The electronic device may have a relatively large difference due to different configurations or performances, and may include one or more processors (CPUs) and one or more memories, where at least one instruction is stored in the memory, and the instruction is loaded by the processor and executes the method.
Fourth embodiment
The present embodiment provides a computer-readable storage medium, in which at least one instruction is stored, and the instruction is loaded and executed by a processor to implement the method of the first embodiment. The computer readable storage medium may be, among others, ROM, random access memory, CD-ROM, magnetic tape, floppy disk, optical data storage device, and the like. The instructions stored therein may be loaded by a processor in the terminal and perform the above-described method.
Furthermore, it should be noted that the present invention may be provided as a method, apparatus or computer program product. Accordingly, embodiments of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present invention may take the form of a computer program product embodied on one or more computer-usable storage media having computer-usable program code embodied in the medium.
Embodiments of the present invention are described with reference to flowchart illustrations and/or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, embedded processor, or other programmable data processing terminal to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks. These computer program instructions may also be loaded onto a computer or other programmable data processing terminal to cause a series of operational steps to be performed on the computer or other programmable terminal to produce a computer implemented process such that the instructions which execute on the computer or other programmable terminal provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It is further noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. The terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or terminal that comprises the element.
Finally, it should be noted that while the above describes a preferred embodiment of the invention, it will be appreciated by those skilled in the art that, once the basic inventive concepts have been learned, numerous changes and modifications may be made without departing from the principles of the invention, which shall be deemed to be within the scope of the invention. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the embodiments of the invention.
Claims (10)
1. A continuous annealing tension setting method based on a neural network is characterized by comprising the following steps:
acquiring historical production data of the strip steel; wherein the historical production data of the strip steel comprises: the steel type, the width, the thickness and the shape value of each cold rolling channel plate of the raw material strip steel, and the speed of each tension section set furnace roller, the production speed of the strip steel and the tension set value of each tension section during continuous annealing production;
grading the steel type, the strip steel thickness and the strip steel width in the strip steel historical production data according to a preset grading rule to obtain a steel type grading value, a strip steel thickness grading value and a strip steel width grading value corresponding to each strip steel; calculating the tension setting average value of each tension section corresponding to each strip steel and the strip shape characteristic value of the cold-rolled strip steel;
preprocessing a steel grade grading value, a strip steel thickness grading value, a strip steel width grading value, a cold-rolled strip steel plate shape characteristic value, a set furnace roller speed of each tension section, a strip steel production speed and a set average value of tension of each tension section, and constructing a data set for model training by using preprocessed data;
constructing a neural network model and training by adopting the data set; the neural network model takes a steel grade grading value, a strip steel thickness grading value, a strip steel width grading value, a cold-rolled strip steel plate shape characteristic value, a furnace roller speed set for each tension section and a strip steel production speed as input, and a tension average value set for each tension section as output;
obtaining tension set values of all sections in the continuous annealing furnace in the current production by using the trained neural network model;
and calculating to obtain a final tension set value by using a weighted sliding average method based on the tension set values of all the sections obtained by the neural network model and the tension set average value of all the tension sections in the historical production of the strip steel.
2. The method for setting the continuous annealing tension based on the neural network of claim 1, wherein the step of classifying the steel type, the strip thickness and the strip width in the strip steel historical production data according to a preset classification rule to obtain the steel type classification value, the strip thickness classification value and the strip width classification value corresponding to each strip steel comprises:
dividing strip steels with different steel grades into four steel families and a default classification according to the steel quality; dividing the band steel width grading range according to the production frequency of each band steel width interval; wherein, the higher the production frequency is, the finer the corresponding width interval is graded; according to the thickness of the strip steel, uniformly dividing the thickness of the strip steel into grading ranges from thin to thick;
and numbering the graded strip steel according to the steel grade grading result, the strip steel width grading result and the strip steel thickness grading result, and taking the steel grade grading value, the strip steel thickness grading value and the strip steel width grading value corresponding to the strip steel.
3. The method for setting the tension of the neural network-based annealing tension of claim 1, wherein the calculating the average tension setting value of each tension section corresponding to each strip steel comprises:
and calculating the average value of the tension of each section of the strip steel during the production of each coil of the strip steel in the time range from the strip head of each strip steel entering the continuous annealing furnace area to the strip tail leaving the continuous annealing furnace area, and taking the average value as the set average value of the tension of each tension section corresponding to the strip steel.
4. The method for setting the continuous annealing tension based on the neural network as claimed in claim 1, wherein the method for calculating the strip shape characteristic value of the cold-rolled strip steel comprises the following steps:
respectively integrating the strip shape primary term coefficient curves corresponding to 50m sections at the head and the tail of the strip steel and three sections in the middle of the strip steel, and carrying out weighted summation on the integration results of the three sections to obtain a strip shape characteristic value I of the cold-rolled strip steel, wherein the formula is as follows:
wherein l represents the length of the strip, a1(l) Is a 50m plate-shaped first-order coefficient curve of the head of the strip steel, a2(l) Is a 50m plate-shaped first-order coefficient curve a of the tail part of the strip steel3(l) Is a primary coefficient curve of the strip shape in the middle of the strip steel.
5. The neural network-based continuous annealing tension setting method of claim 1, wherein the preprocessing includes normalization processing, denoising and outlier data removal.
6. The neural-network-based continuous annealing tension setting method of claim 5, wherein the normalization process is a maximum-minimum normalization process.
7. The method of claim 1, wherein the calculating of the final tension setting value by using a weighted moving average method based on the tension setting values of the tension sections obtained by the neural network model and the tension setting average value of the tension sections in the historical production of the strip steel comprises:
according to a first-order moving average model, each section of tension set value T obtained by calculation of a neural network modelnewSetting average value T of tension of each tension section in historical productionoldWeighted summation is carried out to obtain the final tension set value of each tension section of each specification strip steel
The formula is as follows:
where γ is a weight value of the weighted sum.
8. The neural network-based continuous annealing tension setting method of claim 7, wherein the historical production data of the strip steel further comprises: historical average production speed of the production team, average strip steel quality rating of the production team and average accident rate of the production team;
γ is calculated by the following formula:
wherein, VaFor the current team historical average production speed, VmaxAverage production speed, Q, of teams of highest technical levelaAnd grading the average quality of the strip steel produced by the current team, wherein eta is the average accident rate of the current team.
9. The neural-network-based continuous annealing tension setting method of any one of claims 1-8, wherein after calculating the final tension set value, the continuous annealing tension setting method further comprises:
and taking the final tension set value obtained by calculation as a tension control value of actual production, and putting the final tension set value obtained by calculation into a training set of the neural network model for self-learning training.
10. A continuous annealing tension setting device based on a neural network is characterized by comprising:
the data acquisition module is used for acquiring historical production data of the strip steel; wherein the historical production data of the strip steel comprises: the steel type, the width, the thickness and the shape value of each cold rolling channel plate of the raw material strip steel, and the speed of each tension section set furnace roller, the production speed of the strip steel and the tension set value of each tension section during continuous annealing production;
the data grading and characteristic value calculating module is used for grading the steel grade, the strip steel thickness and the strip steel width in the strip steel historical production data acquired by the data acquiring module according to a preset grading rule to obtain a steel grade grading value, a strip steel thickness grading value and a strip steel width grading value corresponding to each strip steel; calculating the tension setting average value of each tension section corresponding to each strip steel and the strip shape characteristic value of the cold-rolled strip steel;
the data set construction module is used for preprocessing a steel grade grading value, a strip steel thickness grading value, a strip steel width grading value, a cold-rolled strip steel plate characteristic value, a set furnace roller speed of each tension section, a strip steel production speed and a set average tension value of each tension section, and constructing a data set for model training by using the preprocessed data;
the neural network model building and training module is used for building a neural network model and training by adopting the data set built by the data set building module; the neural network model takes a steel grade grading value, a strip steel thickness grading value, a strip steel width grading value, a cold-rolled strip steel plate shape characteristic value, a furnace roller speed set for each tension section and a strip steel production speed as input, and a tension average value set for each tension section as output;
the tension calculation module is used for obtaining tension set values of all sections in the continuous annealing furnace in the current production by utilizing the trained neural network model; and calculating to obtain a final tension set value by using a weighted sliding average method based on the tension set values of all the sections obtained by the neural network model and the tension set average value of all the tension sections in the historical production of the strip steel.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111089354.4A CN113850020B (en) | 2021-09-16 | 2021-09-16 | Continuous annealing tension setting method and device based on neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111089354.4A CN113850020B (en) | 2021-09-16 | 2021-09-16 | Continuous annealing tension setting method and device based on neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113850020A true CN113850020A (en) | 2021-12-28 |
| CN113850020B CN113850020B (en) | 2023-01-31 |
Family
ID=78974432
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111089354.4A Active CN113850020B (en) | 2021-09-16 | 2021-09-16 | Continuous annealing tension setting method and device based on neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113850020B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114854978A (en) * | 2022-04-06 | 2022-08-05 | 武汉钢铁有限公司 | Method and device for predicting strip steel deviation value |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6119112A (en) * | 1997-11-19 | 2000-09-12 | International Business Machines Corporation | Optimum cessation of training in neural networks |
| CN103886373A (en) * | 2014-04-18 | 2014-06-25 | 天津理工大学 | High-speed cold-rolling mill third octave flutter prediction method based on BP neural network |
| CN108229072A (en) * | 2018-02-10 | 2018-06-29 | 东北大学 | Continuous annealing soaking pit strip tension On-line Measuring Method based on data parsing |
| CN112029990A (en) * | 2020-07-16 | 2020-12-04 | 宝钢湛江钢铁有限公司 | Automatic deviation rectifying control method for cold-rolled strip steel continuous annealing heating furnace |
-
2021
- 2021-09-16 CN CN202111089354.4A patent/CN113850020B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6119112A (en) * | 1997-11-19 | 2000-09-12 | International Business Machines Corporation | Optimum cessation of training in neural networks |
| CN103886373A (en) * | 2014-04-18 | 2014-06-25 | 天津理工大学 | High-speed cold-rolling mill third octave flutter prediction method based on BP neural network |
| CN108229072A (en) * | 2018-02-10 | 2018-06-29 | 东北大学 | Continuous annealing soaking pit strip tension On-line Measuring Method based on data parsing |
| CN112029990A (en) * | 2020-07-16 | 2020-12-04 | 宝钢湛江钢铁有限公司 | Automatic deviation rectifying control method for cold-rolled strip steel continuous annealing heating furnace |
Non-Patent Citations (2)
| Title |
|---|
| 孙磊等: "带钢平整张力神经网络预设定模型研究", 《轧钢》 * |
| 张洁露等: "基于BP神经网络的平整机伸长率控制技术", 《冶金自动化》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114854978A (en) * | 2022-04-06 | 2022-08-05 | 武汉钢铁有限公司 | Method and device for predicting strip steel deviation value |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113850020B (en) | 2023-01-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109351785B (en) | Rolling force optimization method and device | |
| CN104517162B (en) | A kind of continuous annealing product hardness Online integration learns forecasting procedure | |
| US10949582B2 (en) | Classification optimization method for steel grade family of steel rolling model | |
| CN113850020B (en) | Continuous annealing tension setting method and device based on neural network | |
| KR101432436B1 (en) | Apparatus and method for prediction of influent flow rate and influent components using nearest neighbor method | |
| CN115035106B (en) | Strip steel defect intelligent detection method | |
| CN114897227A (en) | Multi-steel-grade mechanical property forecasting method based on improved random forest algorithm | |
| CN113102516B (en) | Hot continuous rolling strip steel head width prediction method integrating rolling mechanism and deep learning | |
| CN114818456B (en) | Prediction method and optimization method for full-length deformation resistance of cold continuous rolling strip steel | |
| CN115815342A (en) | Cold rolling force prediction method based on mechanism and data fusion model | |
| CN115034437A (en) | Hot rolled plate convexity prediction method based on improved XGboost | |
| CN115169453A (en) | Hot continuous rolling width prediction method based on density clustering and depth residual error network | |
| CN117724433B (en) | Cold rolling production feedforward control method based on multichannel distributed deep integration prediction | |
| KR20190078334A (en) | Apparatus for setting schedule of continuous rolling pass | |
| CN111598328A (en) | Power load prediction method considering epidemic situation events | |
| CN117010277A (en) | Data-driven-based plate strip steel variable-specification continuous annealing deviation prediction and decision method | |
| CN112598026B (en) | Continuous annealing production process fault diagnosis method based on unbalanced data deep learning | |
| CN116090678B (en) | Data processing method, device and equipment | |
| JP4280351B2 (en) | Rolling control model learning device | |
| CN116689502A (en) | Roll bending force parameter setting method of finishing mill group process of hot continuous rolling mill based on SA-SCSO-SCN neural network | |
| CN116108932A (en) | Method for establishing fusion model of steel production process data and mechanism | |
| CN113695404B (en) | Strip steel hot continuous rolling width control method | |
| CN112861394B (en) | Cold-rolling roller grinding amount optimization method and system based on genetic particle swarm optimization | |
| CN114611376B (en) | Continuous annealing deviation prediction method and device based on neural network | |
| CN112862284A (en) | Method and system for evaluating rigidity precision of hot-rolled strip mill |
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 |