CN112668234B - Intelligent control method for converter steelmaking end point - Google Patents
Intelligent control method for converter steelmaking end point Download PDFInfo
- Publication number
- CN112668234B CN112668234B CN202011414366.5A CN202011414366A CN112668234B CN 112668234 B CN112668234 B CN 112668234B CN 202011414366 A CN202011414366 A CN 202011414366A CN 112668234 B CN112668234 B CN 112668234B
- Authority
- CN
- China
- Prior art keywords
- model
- vector
- subsystem
- end point
- data
- 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
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P10/00—Technologies related to metal processing
- Y02P10/25—Process efficiency
Landscapes
- Carbon Steel Or Casting Steel Manufacturing (AREA)
Abstract
The invention provides an intelligent control method for a converter steelmaking end point, which is realized by the following subsystems: 1) And a data preprocessing subsystem: data are collected from the database, data preprocessing is carried out, and the end point carbon content and the input variable of the temperature prediction subsystem model are determined through independence and correlation analysis, so that the model accuracy is ensured; 2) Molten steel end point prediction subsystem: predicting the endpoint carbon content and endpoint temperature of converter steelmaking by adopting a non-parallel support vector regression algorithm based on wavelet weights; 3) Oxygen blowing amount and auxiliary material calculating subsystem: combining a whale optimization algorithm and an incremental calculation method, calculating an optimization error according to output feedback of a prediction model, and calculating the addition amount of auxiliary materials such as oxygen blowing amount, lime, light burned dolomite and the like required in the converting stage on the premise of ensuring the minimum optimization error; 4) Model update subsystem: and updating and upgrading the prediction subsystem regularly according to the actual production condition. Can realize one-key steelmaking of the converter.
Description
Technical Field
The invention relates to the technical field of converter steelmaking, in particular to an intelligent control method for a converter steelmaking end point.
Background
Converter steelmaking is a main steelmaking mode in China, and the quality of products is directly influenced by the control of end point components and temperature of molten steel. Since the smelting process of steelmaking is an extremely complex physicochemical reaction process, the endpoint control of converter steelmaking is a research focus and difficulty in the field of ferrous metallurgy, and the research of control problems thereof goes through several stages of empirical control, static control, dynamic control and intelligent control.
With the rapid development of computer technology and Internet of things technology, a good foundation is laid for realizing one-key steelmaking. The intelligent control is a current research hot spot problem, the current mainstream control mode is to build a model based on a neural network, establish a prediction model of a converter endpoint, calculate oxygen blowing amount and auxiliary material amount required by smelting by combining a mechanism or a mathematical method on the basis of the model, and further realize endpoint control of the converter. The model has the defects that the model is easy to fall into a local minimum value, which leads to the situation that a global optimal solution cannot be obtained in the modeling process, brings a plurality of inconveniences for searching optimal parameters of the model, has low modeling efficiency and is unfavorable for realizing one-key steelmaking. The new modeling thought is explored, the learning performance and efficiency of the intelligent model are improved, the full-automatic control of the converter is further realized, the on-site operation mode is effectively standardized, the splashing occurrence rate is reduced, the converting time is shortened, the consumption of slag formers is reduced, the primary carbon drawing rate of converter steelmaking is improved, and the method is a great innovation in the field of converter steelmaking.
Disclosure of Invention
In order to solve the technical problems of the background technology, the invention provides an intelligent control method for the steelmaking end point of a converter, which can realize one-key steelmaking of the converter.
In order to achieve the above purpose, the invention is realized by adopting the following technical scheme:
the intelligent control method of the converter steelmaking end point is realized by the following subsystems: the system comprises a data preprocessing subsystem, a molten steel end point prediction subsystem, an oxygen blowing amount and auxiliary material calculation subsystem and a model updating subsystem;
1) And a data preprocessing subsystem: data are collected from the database, data preprocessing is carried out, and the end point carbon content and the input variable of the temperature prediction subsystem model are determined through independence and correlation analysis, so that the model accuracy is ensured;
2) Molten steel end point prediction subsystem: predicting the endpoint carbon content and endpoint temperature of converter steelmaking by adopting a non-parallel support vector regression algorithm (Wavelet transform based weightedNPSVR, WTWNPSVR) based on wavelet weights;
3) Oxygen blowing amount and auxiliary material calculating subsystem: combining a whale optimization algorithm and an incremental calculation method, calculating an optimization error according to output feedback of a prediction model, and calculating the addition amount of auxiliary materials such as oxygen blowing amount, lime, light burned dolomite and the like required in the converting stage on the premise of ensuring the minimum optimization error;
4) Model update subsystem: and updating and upgrading the prediction subsystem regularly according to the actual production condition.
Further, the data preprocessing method of the data preprocessing subsystem specifically comprises the following steps:
step 1-1: reading n groups of converter data from a database, preprocessing the converter data, removing irrelevant information such as smelting numbers, classes, furnace length names and the like, and obtaining n groups of converter data sets with m characteristic variables;
step 1-2: constructing an evaluation problem according to n groups of converter data and m characteristic variables, and determining a reference sequence and a comparison sequence; the original evaluation matrix is:
wherein F is i ＝[f i (1),...,f i (k),...,f i (n)]A comparison sequence for the ith feature variable, f i (k) The ith characteristic variable of the kth group of converter data;
determining a reference sequence R according to the evaluation purpose and the index condition 0 ：
R 0 ＝(r 0 (1),...,r 0 (k),...,r 0 (n)) (2)
For converter data, R 0 Refers to the output sequence of the model, namely the end point carbon content data y C Or end point temperature data y T ，r 0 (k) Output variables of the kth group of converter data;
step 1-3: normalization of reference sequences R 0 Comparing the sequences F to obtain matrix dimensionless data Y;
wherein Y is 0 ＝(y 0 (1),...,y 0 (k),...,y 0 (n)) is a normalized reference sequence;
step 1-4: calculating a difference sequence omega; the difference sequence being the absolute value of the difference between the element of each comparison sequence and the element of the corresponding reference sequence, i.e
Step 1-5: determining the maximum p and minimum v values in the sequence of differences, i.e
Step 1-6: calculating the association coefficient of each comparison sequence and the corresponding element of the reference sequence, namely y i (k) And y 0 (k) Correlation coefficient between：
Wherein ρ ε (0, 1) is an adjustable parameter;
step 1-7: calculating the average value of the correlation coefficient of each feature to form a correlation sequence, i.e
The larger the association coefficient is, the larger the influence of the corresponding input factors on the output variable is explained;
step 1-8: according to the grey correlation coefficient gamma 0 (i) Arranging from large to small, selecting l (l is less than or equal to m) features with larger influence on output variables, and primarily taking the l (l is less than or equal to m) features as corresponding input variables;
step 1-9: adopting a partial correlation analysis method to perform independent analysis on the obtained input variables so as to ensure that the input variables are mutually independent or have smaller correlation; any two input variables x i And x j G (g.ltoreq.l-2) order bias correlation coefficient betweenThe calculation can be performed by the following equation:
wherein the right side of equation (8) represents the partial correlation coefficient of g-1; the partial correlation coefficient is a statistic which can truly reflect the correlation between two variables; if the partial correlation coefficient between the two variables is smaller, the correlation between the two variables is smaller, even uncorrelated;
step 1-10: through mechanism analysis and combination of correlation and independence analysis, d (d is less than or equal to l) influencing factors can be finally determined as the endpoint carbon content y C Or end point temperature y T The input variable of the predictive model is defined as x= [ x ] 1 ,x 2 ,...,x d ] T ∈R d×1 The method comprises the steps of carrying out a first treatment on the surface of the The endpoint carbon content or endpoint temperature is used as the output variable of the predictive model.
Further, in the molten steel endpoint prediction subsystem, the method for predicting the endpoint carbon content and the endpoint temperature of the converter steelmaking specifically comprises the following steps:
step 2-1: reading converter steelmaking input data of Step 1-10, endpoint carbon content data and endpoint temperature data, and determining the quantity L of training samples in a uniform sampling mode 1 (L 1 <n)；
Step 2-2: establishing a converter steelmaking endpoint information prediction model (WWNPSVR) with anti-noise performance; the model is based on NPSVR, and introduces a parameter v 1 And v 2 Meanwhile, the weight among samples is considered; assume a data set isFor Gaussian kernel function, let ∈ ->To input training samples, y= [ y ] 1 ,...,y n ] T ∈R n For outputting training samples, ++>For the ith training sample, the objective function of the algorithm may be described as:
wherein c i (i=1,..4) > 0 is a penalty parameter, v 1 ,v 2 ,ε 1 ,ε 2 Epsilon is more than or equal to 0 and is an adjustable parameter eta 1 ，ξ 1 ，η 2 ，/>And xi 2 Is a relaxation variable, W.epsilon.R n×1 Is the weight vector of the sample, [ w ] 1 ；b 1 ]And [ w ] 2 ；b 2 ]For the augmentation vector, e= [1, ], 1] T ∈R n×1 ；
Taking the objective function (9) as an example for explanation, the objective function (10) has an explanation similar to that of (9); the purpose of the first and second terms in the constraint is to determine two hyperplanesAndso that as many training samples as possible are positioned between the two hyperplanes; the first term of the objective function is a regularization term derived from a standard support vector regression algorithm; the purpose of constraint third item is to let training samples go to the lower boundary hyperplane +.>Is at least epsilon 1 In other words, the training samples are located as much as possibleUpper side of (2); second and third terms of the objective function are used to minimize the relaxation factor +.>And xi 1 And epsilon 1 The width of the pipeline, and the coefficient vector W is a penalty vector of a relaxation factor;
deducing the model can obtain corresponding dual problems:
wherein alpha is 1 ＝[α 11 ,....,α 1n ] T ,β 1 ＝[β 11 ,...,β 1n ] T ,α 2 ＝[α 21 ,....,α 2n ] T ,/>And beta 2 ＝[β 21 ,...,β 2n ] T Is a Lagrangian multiplier vector;
step2-3: based on wavelet transformation theory, determining weight vector W= [ d ] of sample 1 ,d 2 ,···,d n ] T ，d i The representation weight coefficient can be obtained by the following formula:
wherein exp (·) is an e exponential function, δ is the standard deviation of a Gaussian function, r i Is the distance between the original data and the wavelet transformed and noise reduced data;
step 2-4: initializing parameter c of WWNPSVR prediction model i (i＝1,...,4),v 1 ,v 2 ,ε 1 ,ε 2 Epsilon and delta;
step 2-5: training by using a converter steelmaking training sample set; by solving the equations (11) and (12), the optimal solution vector α can be obtained 1 ，β 1 ，α 2 ，/>And beta 2 ；
Step 2-6: by substituting the optimal solution into the following equations (14) and (15), w can be found 1 ,b 1 And w 2 ,b 2 ；
In I S k The I represents the number of support vectors;
step 2-7: will w 1 ,b 1 And w 2 ,b 2 The result of (2) is substituted into formula (16) to obtain the final carbon content regression function f C (x) Or a temperature regression function f T (x)；
Step 2-8: substituting training samples into the regression function f C (x) Or f T (x) Obtaining a predicted value of the end point carbon content or the end point temperature; calculating indexes such as model precision, end point hit rate and the like, if the indexes reach a set value, completing model establishment, otherwise, updating parameters c of the WWNPSVR prediction model 1 ,c 2 ,c 3 ,c 4 ,v 1 ,v 2 ,ε 1 ,ε 2 And epsilon, delta, repeating Step 2-5 to Step 2-7 until the index reaches a set value, and completing the model establishment.
Further, the oxygen blowing amount and auxiliary material calculating method of the oxygen blowing amount and auxiliary material calculating subsystem specifically comprises the following steps:
step 3-1: the objective function of the following oxygen addition optimization problem is established as follows:
wherein, ffit (x) = (fC/T (x) -D C/T ) 2 Is called a fitness function, and x is defined by oxygen blowing amount, lime addition amount and light weightInput vector composed of variables such as adding amount of burnt dolomite, f C/T (x) D is the carbon temperature predicted value of the prediction subsystem in Step2-8 C/T Target value for end point carbon content or end point temperature;
step 3-2: setting the number j of whales and the maximum number N of iterations max The method comprises the steps of carrying out a first treatment on the surface of the Taking variables such as oxygen blowing amount, lime addition amount and the like as optimization variables, carrying out normalization treatment on each variable, and mapping to [1, -1 ]]Generating an initial random solution with equal quantity according to the set whale group number j;
step3-3: combining each group of initial solution variables with initial information of molten iron to obtain x, and substituting the x into a prediction subsystem in Step2-8 to obtain predicted values f of carbon content and temperature C/T (x)；
Step3-4: according to the fitness function f in equation (17) fit (x) Calculating the fitness of each group of solutions, and storing the optimal vector x with the minimum current fitness * ；
Step3-5: using the whale group optimization strategy of equation (18), if the current number of iterations is less than N max Then update a, r 1 E, eta, k and p, determining a solution required for the next iteration, detecting whether a solution exceeding a search space exists, if so, mapping the solution to a random position in a feasible domain, and repeating the steps Step3-4 and Step 3-5; otherwise, returning to the optimal solution to finish whale group optimization of the oxygen amount to obtain an optimal vector B of the oxygen blowing amount and the auxiliary material adding amount * ＝invnorm(x * ) Invnorm represents the inverse normalization process;
where t is the current iteration number, x (t) is the current position vector of the whale of the seat, λ= |2r 1 ·x rand -x(t)|，x rand Is the random position of whale, r 1 Is [ -1,1]Random vectors within an interval, η is a constant, E is a coefficient vector, x * (t) a position vector representing the current optimal solution, lambda z Representing the distance between whale and prey, a is a variable that decreases from 2 to 0, and k is [ -1,1]Interval ofRandom vectors within, z represents a probability variable;
step3-6: using equation (17), an incremental calculation model is built, and the search interval of the oxygen blowing amount is set to be [ M ] 1 ,N 1 ]Step length of l 1 The method comprises the steps of carrying out a first treatment on the surface of the Setting the search interval of the adding amount of auxiliary materials as [ M ] 2 ,N 2 ]Step length of l 2 ；
Step3-7: setting initial values of oxygen blowing amount and auxiliary material adding amount in the same group of samples optimized by whale group as [ M ] 1 ,M 2 ]Then combining the model input vector x with molten iron information, substituting the model input vector x into a prediction subsystem in Step2-8 to obtain a predicted value f of carbon content and temperature C/T (x) The method comprises the steps of carrying out a first treatment on the surface of the Obtaining fitness value according to formula (17)Preserving vector C * ＝x；
Step3-8: according to the search interval l 1 And l 2 Gradually updating the values of the oxygen blowing amount and the auxiliary material adding amount to form a new model output variable x new Substituting into a prediction subsystem in Step2-8 to obtain a predicted value f of carbon content and temperature C/T (x new ) The method comprises the steps of carrying out a first treatment on the surface of the Obtaining fitness value f according to formula (17) fit (x new ) If (3)Then save vector C * ＝x new And fitness valueWhen the oxygen blowing amount reaches N 1 The addition amount of the auxiliary materials reaches N 2 Executing Step 3-9, otherwise repeating Step3-8;
step 3-9: optimal solution B derived from Steps 3-5 and 3-8 * And C * Extracting oxygen blowing amount and auxiliary material adding amount to form new vectorAnd->Substitution formula->Obtaining an optimal vector A of oxygen blowing amount and auxiliary material adding amount, wherein K 1 And K 2 Is a weight coefficient;
step3-10: by adjusting K 1 And K 2 The final oxygen blowing amount and the auxiliary material adding amount A are obtained * The method comprises the steps of carrying out a first treatment on the surface of the And transmitting the final results of the oxygen blowing amount and the auxiliary material calculation subsystem to a database, then reading data from the database by the PLC system, determining the blowing mode and the blowing stopping time according to the optimal results, and finally sending a control instruction to the converter executing mechanism to complete the whole blowing process.
Further, the updating and upgrading method of the model updating subsystem specifically comprises the following steps:
step4-1: when the newly added smelting data volume reaches L 2 At the time, setting the population vector omega of whale group 0 ＝[Ω 1 ,...,Ω l ] T Wherein Ω i ＝[c 1,i ,c 2,i ,c 3,i ,c 4,i ,ε 1,i ,ε 2,i ,ε i ,δ i ] T I=1, the combination of the first and second parts, l, setting the maximum iteration times T max The method comprises the steps of carrying out a first treatment on the surface of the Defining a model parameter optimization fitness function as follows:
in the method, in the process of the invention,size (U, 1) represents the vector x in the matrix U i Number of (A)>Is a predicted value of carbon content and temperature, y (x i ) Is the actual value of carbon content and temperature, τ is the upper error bound, L 1 For the number of training samples determined in Step 2-1;
step 4-2: for each set of parameters, using equation (19)Calculate fitness +.>Obtaining the current optimal fitness f * (Ω * ) Optimum parameter Ω * ；
Step 4-3: if the current iteration number is less than T max Update Ω * Obtaining the next set of parameters omega i Using equation (19), the fitness f (Ω) i )；
Step 4-4: if f * (Ω * )≤f(Ω i ) Let omega * ＝Ω i Andreturning to Step 4-3; otherwise, directly returning to Step 4-3;
step 4-5: if the current iteration number is equal to T max Then the global optimum parameter omega is output * The method comprises the steps of carrying out a first treatment on the surface of the And (5) updating the model.
Compared with the prior art, the invention has the beneficial effects that:
1. the molten steel endpoint prediction subsystem adopts a non-parallel support vector machine algorithm (Wavelet transform based weightedNPSVR, WTWNPSVR) based on wavelet weights to realize the prediction of endpoint carbon content and endpoint temperature of converter steelmaking. Wtwnplsvr is a new method of pattern recognition based on statistical learning theory that seeks the best compromise between model complexity and learning ability based on limited sample information to obtain the best generalization performance. The WWNPSVR algorithm can overcome the inherent defects of the neural network in the modeling process, and has the following advantages: 1) The method is specially used for the condition of limited samples, and aims to obtain an optimal solution under the existing information, and not just an optimal value when the number of samples tends to infinity; 2) The NPSVR algorithm is finally converted into two secondary convex planning problems, and in theory, the obtained solution is necessarily the global optimal solution, so that the problem that the neural network prediction model is easy to fall into a local minimum value in the modeling process is solved; 3) The algorithm converts the actual problem into a high-dimensional characteristic space through nonlinear transformation, and constructs a linear discriminant function in the high-dimensional space to replace the nonlinear discriminant function in the original space, so that the special property ensures that the model has better prediction performance, and meanwhile, the dimension problem is skillfully solved, and the dimension disaster is avoided. 4) The WWNPSVR adopts a wavelet transformation method to establish a prediction model, so that adverse effects of noise can be restrained, the accuracy of end point prediction is improved, the yield of target products is further improved, and energy conservation and emission reduction are realized.
2. The oxygen blowing amount and auxiliary material calculating subsystem adopts a combined calculation model, combines the advantages of whale group optimization and increment calculation, can further improve the calculation accuracy of oxygen amount, improves the primary carbon drawing rate of converter steelmaking, and can save labor cost for enterprises.
3. The model updating subsystem adopts Whale Optimization Algorithm (WOA) to optimize parameters in the prediction model, so that updating and upgrading of the prediction subsystem are realized. The value of the parameter has important significance on the generalization capability and stability of the model. Usually, the optimal parameters of the model are selected by adopting a grid search method, but the grid search method is slow and long in running time. In order to overcome the problem, the invention combines the advantages of quick convergence speed, simple adjustment parameters and the like of the WOA algorithm, and can update and upgrade the molten steel end point prediction subsystem regularly, thereby effectively avoiding the problem of sinking into local optimum, improving the updating efficiency of the model and having reliability.
Drawings
FIG. 1 is a diagram of a converter steelmaking intelligent control model structure according to the invention;
FIG. 2 is a flow chart of a molten steel endpoint prediction subsystem of the present invention;
FIG. 3 is a flow chart of a model update subsystem of the present invention.
Detailed Description
The following detailed description of the embodiments of the invention is provided with reference to the accompanying drawings.
As shown in fig. 1, the intelligent control method of the converter steelmaking end point is realized by the following subsystems: the system comprises a data preprocessing subsystem, a molten steel end point prediction subsystem, an oxygen blowing amount and auxiliary material calculation subsystem and a model updating subsystem.
1) And a data preprocessing subsystem: data are collected from the database, data preprocessing is carried out, and the end point carbon content and the input variable of the temperature prediction subsystem model are determined through independence and correlation analysis, so that the model accuracy is ensured; the method specifically comprises the following steps:
step 1-1: reading n groups of converter data from a database, preprocessing the converter data, removing irrelevant information such as smelting numbers, classes, furnace length names and the like, and obtaining n groups of converter data sets with m characteristic variables;
step 1-2: constructing an evaluation problem according to n groups of converter data and m characteristic variables, and determining a reference sequence and a comparison sequence; the original evaluation matrix is:
wherein F is i ＝[f i (1),...,f i (k),...,f i (n)]A comparison sequence for the ith feature variable, f i (k) The ith characteristic variable of the kth group of converter data;
determining a reference sequence R according to the evaluation purpose and the index condition 0 ：
R 0 ＝(r 0 (1),...,r 0 (k),...,r 0 (n)) (2)
For converter data, R 0 Refers to the output sequence of the model, namely the end point carbon content data y C Or end point temperature data y T ，r 0 (k) Output variables of the kth group of converter data;
step 1-3: normalization of reference sequences R 0 Comparing the sequences F to obtain matrix dimensionless data Y;
wherein Y is 0 ＝(y 0 (1),...,y 0 (k),...,y 0 (n)) is a normalized reference sequence;
step 1-4: calculating a difference sequence omega; the difference sequence being the absolute value of the difference between the element of each comparison sequence and the element of the corresponding reference sequence, i.e
Step 1-5: determining the maximum p and minimum v values in the sequence of differences, i.e
Step 1-6: calculating the association coefficient of each comparison sequence and the corresponding element of the reference sequence, namely y i (k) And y 0 (k) Correlation coefficient between：
Wherein ρ ε (0, 1) is an adjustable parameter;
step 1-7: calculating the average value of the correlation coefficient of each feature to form a correlation sequence, i.e
The larger the association coefficient is, the larger the influence of the corresponding input factors on the output variable is explained;
step 1-8: according to the grey correlation coefficient gamma 0 (i) Arranging from large to small, selecting l (l is less than or equal to m) features with larger influence on output variables, and primarily taking the l (l is less than or equal to m) features as corresponding input variables;
step 1-9: adopting a partial correlation analysis method to perform independent analysis on the obtained input variables so as to ensure that the input variables are mutually independent or have smaller correlation; any two input variables x i And x j G (g.ltoreq.l-2) order bias correlation coefficient betweenThe calculation can be performed by the following equation:
wherein the right side of equation (8) represents the partial correlation coefficient of g-1; the partial correlation coefficient is a statistic which can truly reflect the correlation between two variables; if the partial correlation coefficient between the two variables is smaller, the correlation between the two variables is smaller, even uncorrelated;
step 1-10: through mechanism analysis and combination of correlation and independence analysis, d (d is less than or equal to l) influencing factors can be finally determined as the endpoint carbon content y C Or end point temperature y T The input variable of the predictive model is defined as x= [ x ] 1 ,x 2 ,...,x d ] T ∈R d×1 The method comprises the steps of carrying out a first treatment on the surface of the The endpoint carbon content or endpoint temperature is used as the output variable of the predictive model.
2) Molten steel end point prediction subsystem: predicting the endpoint carbon content and endpoint temperature of converter steelmaking by adopting a non-parallel support vector regression algorithm (Wavelet transformbasedweightedNPSVR, WTWNPSVR) based on Wavelet weights; as shown in fig. 2, the method specifically comprises the following steps:
step 2-1: reading converter steelmaking input data of Step 1-10, endpoint carbon content data and endpoint temperature data, and determining the quantity L of training samples in a uniform sampling mode 1 (L 1 <n)；
Step 2-2: establishing a converter steelmaking endpoint information prediction model (WWNPSVR) with anti-noise performance; the model is based on NPSVR, and introduces a parameter v 1 And v 2 Is the same as that ofWhen, the weights between samples are considered; assume a data set isFor Gaussian kernel function, let ∈ ->To input training samples, y= [ y ] 1 ,...,y n ] T ∈R n For outputting training samples, ++>For the ith training sample, the objective function of the algorithm may be described as:
wherein c i (i=1,..4) > 0 is a penalty parameter, v 1 ,v 2 ,ε 1 ,ε 2 Epsilon is more than or equal to 0 and is an adjustable parameter eta 1 ，ξ 1 ，η 2 ，/>And xi 2 Is a relaxation variable, W.epsilon.R n×1 Is the weight vector of the sample, [ w ] 1 ；b 1 ]And [ w ] 2 ；b 2 ]For the augmentation vector, e= [1, ], 1] T ∈R n×1 ；
Taking the objective function (9) as an example for explanation, the objective function (10) has an explanation similar to that of (9); the purpose of the first and second terms in the constraint is to determine two hyperplanesAndso that as many training samples as possible are positioned between the two hyperplanes; the first term of the objective function is a regularization term, derived from standard SVR; the third term of constraint is to make training samples to the lower boundary hyperplaneIs at least epsilon 1 In other words, the training samples are made to lie as far as possible +.>Upper side of (2); second and third terms of the objective function are used to minimize the relaxation factor η 1 ，/>And xi 1 And epsilon 1 The width of the pipeline, and the coefficient vector W is a penalty vector of a relaxation factor;
deducing the model can obtain corresponding dual problems:
wherein alpha is 1 ＝[α 11 ,....,α 1n ] T ,β 1 ＝[β 11 ,...,β 1n ] T ,α 2 ＝[α 21 ,....,α 2n ] T ,/>And beta 2 ＝[β 21 ,...,β 2n ] T Is a Lagrangian multiplier vector;
step2-3: based on wavelet transformation theory, determining weight vector W= [ d ] of sample 1 ,d 2 ,···,d n ] T ，d i The representation weight coefficient can be obtained by the following formula:
wherein exp (·) is an e exponential function, δ is the standard deviation of a Gaussian function, r i Is the distance between the original data and the wavelet transformed and noise reduced data;
step 2-4: initializing parameter c of WWNPSVR prediction model i (i＝1,...,4),v 1 ,v 2 ,ε 1 ,ε 2 Epsilon and delta;
step 2-5: training by using a converter steelmaking training sample set; by solving the equations (11) and (12), the optimal solution vector α can be obtained 1 ，β 1 ，α 2 ，/>And beta 2 ；
Step 2-6: by substituting the optimal solution into the following equations (14) and (15), w can be found 1 ,b 1 And w 2 ,b 2 ；
/>
In I S k The I represents the number of support vectors;
Step 2-7: will w 1 ,b 1 And w 2 ,b 2 The result of (2) is substituted into formula (16) to obtain the final carbon content regression function f C (x) Or a temperature regression function f T (x)；
Step 2-8: substituting training samples into the regression function f C (x) Or f T (x) Obtaining a predicted value of the end point carbon content or the end point temperature; calculating indexes such as model precision, end point hit rate and the like, if the indexes reach a set value, completing model establishment, otherwise, updating parameters c of the WWNPSVR prediction model 1 ,c 2 ,c 3 ,c 4 ,v 1 ,v 2 ,ε 1 ,ε 2 And epsilon, delta, repeating Step 2-5 to Step 2-7 until the index reaches a set value, and completing the model establishment.
3) Oxygen blowing amount and auxiliary material calculating subsystem: combining a whale optimization algorithm and an incremental calculation method, calculating an optimization error according to output feedback of a prediction model, and calculating the addition amount of auxiliary materials such as oxygen blowing amount, lime, light burned dolomite and the like required in the converting stage on the premise of ensuring the minimum optimization error; the method specifically comprises the following steps:
step 3-1: the objective function of the following oxygen addition optimization problem is established as follows:
wherein f fit (x)＝(f C/T (x)-D C/T ) 2 Is called a fitness function, x is an input variable composed of variables such as oxygen blowing amount, lime addition amount, light burned dolomite addition amount and the like, f C/T (x) D is the carbon temperature predicted value of the prediction subsystem in Step2-8 C/T Target value for end point carbon content or end point temperature;
step 3-2: setting the number j of whales and the maximum number N of iterations max The method comprises the steps of carrying out a first treatment on the surface of the The variables such as oxygen blowing amount, lime adding amount and the like are used as optimization variablesNormalizing the variables, mapping to [1, -1 ]]Generating an initial random solution with equal quantity according to the set whale group number j;
step3-3: combining each group of initial solution variables with initial information of molten iron to obtain x, and substituting the x into a prediction subsystem in Step2-8 to obtain predicted values f of carbon content and temperature C/T (x)；
Step3-4: according to the fitness function f in equation (17) fit (x) Calculating the fitness of each group of solutions, and storing the optimal vector x with the minimum current fitness * ；
Step3-5: using the whale group optimization strategy of equation (18), if the current number of iterations is less than N max Then update a, r 1 E, eta, k and p, determining a solution required for the next iteration, detecting whether a solution exceeding a search space exists, if so, mapping the solution to a random position in a feasible domain, and repeating the steps Step3-4 and Step 3-5; otherwise, returning to the optimal solution to finish whale group optimization of the oxygen amount to obtain an optimal vector B of the oxygen blowing amount and the auxiliary material adding amount * ＝invnorm(x * ) Invnorm represents the inverse normalization process;
where t is the current iteration number, x (t) is the current position vector of the whale of the seat, λ= |2r 1 ·x rand -x(t)|，x rand Is the random position of whale, r 1 Is [ -1,1]Random vectors within an interval, η is a constant, E is a coefficient vector, x * (t) a position vector representing the current optimal solution, lambda z Representing the distance between whale and prey, a is a variable that decreases from 2 to 0, and k is [ -1,1]Random vectors within the interval, z representing probability variables;
step3-6: using equation (17), an incremental calculation model is built, and the search interval of the oxygen blowing amount is set to be [ M ] 1 ,N 1 ]Step length of l 1 The method comprises the steps of carrying out a first treatment on the surface of the Setting the search interval of the adding amount of auxiliary materials as [ M ] 2 ,N 2 ]Step length of l 2 ；
Step3-7: setting initial values of oxygen blowing amount and auxiliary material adding amount in the same group of samples optimized by whale group as [ M ] 1 ,M 2 ]Then combining the model input vector x with molten iron information, substituting the model input vector x into a prediction subsystem in Step2-8 to obtain a predicted value f of carbon content and temperature C/T (x) The method comprises the steps of carrying out a first treatment on the surface of the Obtaining fitness value according to formula (17)Preserving vector C * ＝x；
Step3-8: according to the search interval l 1 And l 2 Gradually updating the values of the oxygen blowing amount and the auxiliary material adding amount to form a new model output variable x new Substituting into a prediction subsystem in Step2-8 to obtain a predicted value f of carbon content and temperature C/T (x new ) The method comprises the steps of carrying out a first treatment on the surface of the Obtaining fitness value f according to formula (17) fit (x new ) If (3)Then save vector C * ＝x new And fitness valueWhen the oxygen blowing amount reaches N 1 The addition amount of the auxiliary materials reaches N 2 Executing Step 3-9, otherwise repeating Step3-8;
step 3-9: optimal solution B derived from Steps 3-5 and 3-8 * And C * Extracting oxygen blowing amount and auxiliary material adding amount to form new vectorAnd->Substitution formula->Obtaining an optimal vector A of oxygen blowing amount and auxiliary material adding amount, wherein K 1 And K 2 Is a weight coefficient;
step3-10: general purpose medicineOverregulation K 1 And K 2 The final oxygen blowing amount and the auxiliary material adding amount A are obtained * The method comprises the steps of carrying out a first treatment on the surface of the And transmitting the final results of the oxygen blowing amount and the auxiliary material calculation subsystem to a database, then reading data from the database by the PLC system, determining the blowing mode and the blowing stopping time according to the optimal results, and finally sending a control instruction to the converter executing mechanism to complete the whole blowing process.
4) Model update subsystem: according to actual production conditions, updating and upgrading the prediction subsystem regularly; as shown in fig. 3, the specific steps include the following:
step4-1: when the newly added smelting data volume reaches L 2 At the time, setting the population vector omega of whale group 0 ＝[Ω 1 ,...,Ω l ] T Wherein Ω i ＝[c 1,i ,c 2,i ,c 3,i ,c 4,i ,ε 1,i ,ε 2,i ,ε i ,δ i ] T I=1, the combination of the first and second parts, l, setting the maximum iteration times T max The method comprises the steps of carrying out a first treatment on the surface of the Defining a model parameter optimization fitness function as follows:
in the method, in the process of the invention,size (U, 1) represents the vector x in the matrix U i Number of (A)>Is a predicted value of carbon content and temperature, y (x i ) Is the actual value of carbon content and temperature, τ is the upper error bound, L 1 For the number of training samples determined in Step 2-1;
step 4-2: for each set of parameters, using equation (19)Calculate fitness +.>Obtaining the current optimal fitness f * (Ω * ) Optimum parameter Ω * ；/>
Step 4-3: if the current iteration number is less than T max Update Ω * Obtaining the next set of parameters omega i Using equation (19), the fitness f (Ω) i )；
Step 4-4: if f * (Ω * )≤f(Ω i ) Let omega * ＝Ω i Andreturning to Step 4-3; otherwise, directly returning to Step 4-3;
step 4-5: if the current iteration number is equal to T max Then the global optimum parameter omega is output * The method comprises the steps of carrying out a first treatment on the surface of the And (5) updating the model.
The above examples are implemented on the premise of the technical scheme of the present invention, and detailed implementation manners and specific operation processes are given, but the protection scope of the present invention is not limited to the above examples. The methods used in the above examples are conventional methods unless otherwise specified.
Claims (1)
1. The intelligent control method for the converter steelmaking end point is characterized by being realized by the following subsystems: the system comprises a data preprocessing subsystem, a molten steel end point prediction subsystem, an oxygen blowing amount and auxiliary material calculation subsystem and a model updating subsystem;
1) And a data preprocessing subsystem: data are collected from the database, data preprocessing is carried out, and the end point carbon content and the input variable of the temperature prediction subsystem model are determined through independence and correlation analysis, so that the model accuracy is ensured;
2) Molten steel end point prediction subsystem: predicting the endpoint carbon content and endpoint temperature of converter steelmaking by adopting a non-parallel support vector regression algorithm based on wavelet weights;
3) Oxygen blowing amount and auxiliary material calculating subsystem: combining a whale optimization algorithm and an incremental calculation method, calculating an optimization error according to output feedback of a prediction model, and calculating the addition amount of auxiliary materials such as oxygen blowing amount, lime, light burned dolomite and the like required in the converting stage on the premise of ensuring the minimum optimization error;
4) Model update subsystem: according to actual production conditions, updating and upgrading the prediction subsystem regularly;
the data preprocessing method of the data preprocessing subsystem specifically comprises the following steps:
step 1-1: reading n groups of converter data from a database, preprocessing the converter data, removing irrelevant information such as smelting numbers, classes, furnace length names and the like, and obtaining n groups of converter data sets with m characteristic variables;
step 1-2: constructing an evaluation problem according to n groups of converter data and m characteristic variables, and determining a reference sequence and a comparison sequence; the original evaluation matrix is:
wherein F is i ＝[f i (1),...,f i (k),...,f i (n)]A comparison sequence for the ith feature variable, f i (k) The ith characteristic variable of the kth group of converter data;
determining a reference sequence R according to the evaluation purpose and the index condition 0 ：
R 0 ＝(r 0 (1),...,r 0 (k),...,r 0 (n)) (2)
For converter data, R 0 Refers to the output sequence of the model, namely the end point carbon content data y C Or end point temperature data y T ，r 0 (k) Output variables of the kth group of converter data;
step 1-3: normalization of reference sequences R 0 Comparing the sequences F to obtain matrix dimensionless data Y;
wherein Y is 0 ＝(y 0 (1),...,y 0 (k),...,y 0 (n)) is a normalized reference sequence;
step 1-4: calculating a difference sequence omega; the difference sequence being the absolute value of the difference between the element of each comparison sequence and the element of the corresponding reference sequence, i.e
Step 1-5: determining the maximum p and minimum v values in the sequence of differences, i.e
Step 1-6: calculating the association coefficient of each comparison sequence and the corresponding element of the reference sequence, namely y i (k) And y 0 (k) Correlation coefficient between
Wherein ρ ε (0, 1) is an adjustable parameter;
step 1-7: calculating the average value of the correlation coefficient of each feature to form a correlation sequence, i.e
The larger the association coefficient is, the larger the influence of the corresponding input factors on the output variable is explained;
step 1-8: according to the grey correlation coefficient gamma 0 (i) Arranging from large to small, selecting l (l is less than or equal to m) features with larger influence on output variables, and preliminarily taking the features as corresponding input variables；
Step 1-9: adopting a partial correlation analysis method to perform independent analysis on the obtained input variables so as to ensure that the input variables are mutually independent or have smaller correlation; any two input variables x i And x j G (g.ltoreq.l-2) order bias correlation coefficient betweenThe calculation can be performed by the following equation:
wherein the right side of equation (8) represents the partial correlation coefficient of g-1; the partial correlation coefficient is a statistic which can truly reflect the correlation between two variables; if the partial correlation coefficient between the two variables is smaller, the correlation between the two variables is smaller, even uncorrelated;
step 1-10: through mechanism analysis and combination of correlation and independence analysis, d (d is less than or equal to l) influencing factors can be finally determined as the endpoint carbon content y C Or end point temperature y T The input variable of the predictive model is defined as x= [ x ] 1 ,x 2 ,...,x d ] T ∈R d×1 The method comprises the steps of carrying out a first treatment on the surface of the The end point carbon content or the end point temperature is used as an output variable of the prediction model;
in the molten steel end point predicting subsystem, the method for predicting the end point carbon content and the end point temperature of converter steelmaking specifically comprises the following steps:
step 2-1: reading converter steelmaking input data of Step 1-10, endpoint carbon content data and endpoint temperature data, and determining the quantity L of training samples in a uniform sampling mode 1 (L 1 <n)；
Step 2-2: establishing a converter steelmaking endpoint information prediction model with anti-noise performance; the model is based on NPSVR, and introduces a parameter v 1 And v 2 Meanwhile, the weight among samples is considered; assume a data set is For Gaussian kernel function, let ∈ ->To input training samples, y= [ y ] 1 ,...,y n ] T ∈R n For outputting training samples, ++>For the ith training sample, the objective function of the algorithm may be described as:
wherein c i (i=1,..4) > 0 is a penalty parameter, v 1 ,v 2 ,ε 1 ,ε 2 Epsilon is more than or equal to 0 and is an adjustable parameter eta 1 ,ξ 1 ,η 2 ,And xi 2 Is a relaxation variable, W.epsilon.R n×1 Is the weight vector of the sample, [ w ] 1 ；b 1 ]And [ w ] 2 ；b 2 ]For the augmentation vector, e= [1, ], 1] T ∈R n×1 ；
Taking the objective function (9) as an example for explanation, the objective function (10) has an explanation similar to that of (9); first term in constraint condition andthe second item is aimed at determining two hyperplanesAnd->So that as many training samples as possible are positioned between the two hyperplanes; the first term of the objective function is a regularization term, derived from standard SVR; the purpose of constraint third item is to let training samples go to the lower boundary hyperplane +.>Is at least epsilon 1 In other words, the training samples are made to lie as far as possible +.>Upper side of (2); second and third terms of the objective function are used to minimize the relaxation factor η 1 ,/>And xi 1 And epsilon 1 The width of the pipeline, and the coefficient vector W is a penalty vector of a relaxation factor;
deducing the model can obtain corresponding dual problems:
wherein alpha is 1 ＝[α 11 ,....,α 1n ] T ,β 1 ＝[β 11 ,...,β 1n ] T ,α 2 ＝[α 21 ,....,α 2n ] T ,And beta 2 ＝[β 21 ,...,β 2n ] T Is a Lagrangian multiplier vector;
step2-3: based on wavelet transformation theory, determining weight vector W= [ d ] of sample 1 ,d 2 ,···,d n ] T ，d i The representation weight coefficient can be obtained by the following formula:
wherein exp (·) is an e exponential function, δ is the standard deviation of a Gaussian function, r i Is the distance between the original data and the wavelet transformed and noise reduced data;
step 2-4: initializing parameter c of converter steelmaking end point information prediction model of Step 2-2 i (i＝1,...,4),v 1 ,v 2 ,ε 1 ,ε 2 Epsilon and delta;
step 2-5: training by using a converter steelmaking training sample set; by solving the equations (11) and (12), the optimal solution vector α can be obtained 1 ,β 1 ,α 2 ,/>And beta 2 ；
Step 2-6: by substituting the optimal solution into the following equations (14) and (15), w can be found 1 ,b 1 And w 2 ,b 2 ；
In I S k The I represents the number of support vectors;
step 2-7: will w 1 ,b 1 And w 2 ,b 2 The result of (2) is substituted into formula (16) to obtain the final carbon content regression function f C (x) Or a temperature regression function f T (x)；
Step 2-8: substituting training samples into the regression function f C (x) Or f T (x) Obtaining a predicted value of the end point carbon content or the end point temperature; calculating indexes such as model precision, end hit rate and the like, if the indexes reach a set value, completing model establishment, otherwise, updating parameters c of a converter steelmaking end information prediction model of Step 2-2 1 ,c 2 ,c 3 ,c 4 ,v 1 ,v 2 ,ε 1 ,ε 2 And (3) repeating Step 2-5 to Step 2-7 until the index reaches a set value, epsilon and delta, and completing model establishment;
the oxygen blowing amount and auxiliary material calculating method of the oxygen blowing amount and auxiliary material calculating subsystem specifically comprises the following steps:
step 3-1: the objective function of the following oxygen addition optimization problem is established as follows:
wherein f fit (x)＝(f C/T (x)-D C/T ) 2 Called fitness function, x is an input vector composed of variables such as oxygen blowing amount, lime addition amount, light burned dolomite addition amount and the like, f C/T (x) D is the carbon temperature predicted value of the prediction subsystem in Step2-8 C/T For the end point carbon contentOr a target value of the end point temperature;
step 3-2: setting the number j of whales and the maximum number N of iterations max The method comprises the steps of carrying out a first treatment on the surface of the Taking variables such as oxygen blowing amount, lime addition amount and the like as optimization variables, carrying out normalization treatment on each variable, and mapping to [1, -1 ]]Generating an initial random solution with equal quantity according to the set whale group number j;
step3-3: combining each group of initial solution variables with initial information of molten iron to obtain x, and substituting the x into a prediction subsystem in Step2-8 to obtain predicted values f of carbon content and temperature C/T (x)；
Step3-4: according to the fitness function f in equation (17) fit (x) Calculating the fitness of each group of solutions, and storing the optimal vector x with the minimum current fitness * ；
Step3-5: using the whale group optimization strategy of equation (18), if the current number of iterations is less than N max Then update a, r 1 E, eta, k and p, determining a solution required for the next iteration, detecting whether a solution exceeding a search space exists, if so, mapping the solution to a random position in a feasible domain, and repeating the steps Step3-4 and Step 3-5; otherwise, returning to the optimal solution to finish whale group optimization of the oxygen amount to obtain an optimal vector B of the oxygen blowing amount and the auxiliary material adding amount * ＝invnorm(x * ) Invnorm represents the inverse normalization process;
where t is the current iteration number, x (t) is the current position vector of the whale of the seat, λ= |2r 1 ·x rand -x(t)|，x rand Is the random position of whale, r 1 Is [ -1,1]Random vectors within an interval, η is a constant, E is a coefficient vector, x * (t) a position vector representing the current optimal solution, lambda z Representing the distance between whale and prey, a is a variable that decreases from 2 to 0, and k is [ -1,1]Random vectors within the interval, z representing probability variables;
step3-6: establishing an incremental calculation model by using a formula (17), and setting the search of oxygen blowing quantityThe cable spacing is [ M ] 1 ,N 1 ]Step length of l 1 The method comprises the steps of carrying out a first treatment on the surface of the Setting the search interval of the adding amount of auxiliary materials as [ M ] 2 ,N 2 ]Step length of l 2 ；
Step3-7: setting initial values of oxygen blowing amount and auxiliary material adding amount in the same group of samples optimized by whale group as [ M ] 1 ,M 2 ]Then combining the model input vector x with molten iron information, substituting the model input vector x into a prediction subsystem in Step2-8 to obtain a predicted value f of carbon content and temperature C/T (x) The method comprises the steps of carrying out a first treatment on the surface of the Obtaining fitness value according to formula (17)Preserving vector C * ＝x；
Step3-8: according to the search interval l 1 And l 2 Gradually updating the values of the oxygen blowing amount and the auxiliary material adding amount to form a new model output variable x new Substituting into a prediction subsystem in Step2-8 to obtain a predicted value f of carbon content and temperature C/T (x new ) The method comprises the steps of carrying out a first treatment on the surface of the Obtaining fitness value f according to formula (17) fit (x new ) If (3)Then save vector C * ＝x new And fitness valueWhen the oxygen blowing amount reaches N 1 The addition amount of the auxiliary materials reaches N 2 Executing Step 3-9, otherwise repeating Step3-8;
step 3-9: optimal solution B derived from Steps 3-5 and 3-8 * And C * Extracting oxygen blowing amount and auxiliary material adding amount to form new vectorAnd->Substitution formula->Obtaining an optimal vector A of oxygen blowing amount and auxiliary material adding amount, wherein K 1 And K 2 Is a weight coefficient;
step3-10: by adjusting K 1 And K 2 The final oxygen blowing amount and the auxiliary material adding amount A are obtained * The method comprises the steps of carrying out a first treatment on the surface of the Transmitting the final results of the oxygen blowing amount and the auxiliary material calculation subsystem to a database, then reading data from the database by a PLC system, determining a blowing mode and blowing stopping time according to the optimal results, and finally sending a control instruction to a converter executing mechanism to complete the whole blowing process;
the updating and upgrading method of the model updating subsystem specifically comprises the following steps:
step4-1: when the newly added smelting data volume reaches L 2 At the time, setting the population vector omega of whale group 0 ＝[Ω 1 ,...,Ω l ] T Wherein Ω i ＝[c 1,i ,c 2,i ,c 3,i ,c 4,i ,ε 1,i ,ε 2,i ,ε i ,δ i ] T I=1, the combination of the first and second parts, l, setting the maximum iteration times T max The method comprises the steps of carrying out a first treatment on the surface of the Defining a model parameter optimization fitness function as follows:
in the method, in the process of the invention,size (U, 1) represents the vector x in the matrix U i Number of (A)>Is a predicted value of carbon content and temperature, y (x i ) Is the actual value of carbon content and temperature, τ is the upper error bound, L 1 For the number of training samples determined in Step 2-1;
step 4-2: for each group of parameters, using equation (19)Number of digitsCalculate fitness +.>Obtaining the current optimal fitness f * (Ω * ) Optimum parameter Ω * ；
Step 4-3: if the current iteration number is less than T max Update Ω * Obtaining the next set of parameters omega i Using equation (19), the fitness f (Ω) i )；
Step 4-4: if f * (Ω * )≤f(Ω i ) Let omega * ＝Ω i Andreturning to Step 4-3; otherwise, directly returning to Step 4-3;
step 4-5: if the current iteration number is equal to T max Then the global optimum parameter omega is output * The method comprises the steps of carrying out a first treatment on the surface of the And (5) updating the model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011414366.5A CN112668234B (en) | 2020-12-07 | 2020-12-07 | Intelligent control method for converter steelmaking end point |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011414366.5A CN112668234B (en) | 2020-12-07 | 2020-12-07 | Intelligent control method for converter steelmaking end point |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112668234A CN112668234A (en) | 2021-04-16 |
| CN112668234B true CN112668234B (en) | 2023-08-15 |
Family
ID=75401327
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011414366.5A Active CN112668234B (en) | 2020-12-07 | 2020-12-07 | Intelligent control method for converter steelmaking end point |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112668234B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115058555B (en) * | 2022-07-19 | 2023-06-30 | 北京博谦工程技术有限公司 | Intelligent soft measurement method and system for measuring carbon content of converter endpoint |
| CN115456264B (en) * | 2022-09-01 | 2023-07-25 | 北京科技大学 | Endpoint carbon content and endpoint temperature prediction method for small and medium-sized converter |
| CN117851816A (en) * | 2024-03-07 | 2024-04-09 | 北京科技大学 | Method for predicting steelmaking end point component of converter |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102766723A (en) * | 2012-06-29 | 2012-11-07 | 石家庄钢铁有限责任公司 | Limestone slagging steelmaking lance position control method for basic oxygen furnaces |
| CN103778466A (en) * | 2013-12-30 | 2014-05-07 | 浙江中控研究院有限公司 | Converter steel-making temperature modeling forecasting method based on vector error and system thereof |
| CN106148637A (en) * | 2015-04-10 | 2016-11-23 | 南京理工大学 | The pneumatic steelmaking carbon content dynamic detection system of high stable |
| CN108676955A (en) * | 2018-05-02 | 2018-10-19 | 中南大学 | A kind of BOF Steelmaking Endpoint carbon content and temprature control method |
| CN109447346A (en) * | 2018-10-26 | 2019-03-08 | 冶金自动化研究设计院 | Based on gray prediction and neural network ensemble model converter oxygen consumption prediction technique |
| CN109886456A (en) * | 2019-01-14 | 2019-06-14 | 西安工程大学 | Debris flow prediction method based on PCA Yu mixed kernel function LSSVR |
| CN111915080A (en) * | 2020-07-30 | 2020-11-10 | 中冶华天工程技术有限公司 | Raw fuel cost optimal proportioning method based on molten iron quality constraint |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8002871B2 (en) * | 2008-02-01 | 2011-08-23 | Honeywell International Inc. | Methods and apparatus for an oxygen furnace quality control system |
| US20200133254A1 (en) * | 2018-05-07 | 2020-04-30 | Strong Force Iot Portfolio 2016, Llc | Methods and systems for data collection, learning, and streaming of machine signals for part identification and operating characteristics determination using the industrial internet of things |
-
2020
- 2020-12-07 CN CN202011414366.5A patent/CN112668234B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102766723A (en) * | 2012-06-29 | 2012-11-07 | 石家庄钢铁有限责任公司 | Limestone slagging steelmaking lance position control method for basic oxygen furnaces |
| CN103778466A (en) * | 2013-12-30 | 2014-05-07 | 浙江中控研究院有限公司 | Converter steel-making temperature modeling forecasting method based on vector error and system thereof |
| CN106148637A (en) * | 2015-04-10 | 2016-11-23 | 南京理工大学 | The pneumatic steelmaking carbon content dynamic detection system of high stable |
| CN108676955A (en) * | 2018-05-02 | 2018-10-19 | 中南大学 | A kind of BOF Steelmaking Endpoint carbon content and temprature control method |
| CN109447346A (en) * | 2018-10-26 | 2019-03-08 | 冶金自动化研究设计院 | Based on gray prediction and neural network ensemble model converter oxygen consumption prediction technique |
| CN109886456A (en) * | 2019-01-14 | 2019-06-14 | 西安工程大学 | Debris flow prediction method based on PCA Yu mixed kernel function LSSVR |
| CN111915080A (en) * | 2020-07-30 | 2020-11-10 | 中冶华天工程技术有限公司 | Raw fuel cost optimal proportioning method based on molten iron quality constraint |
Non-Patent Citations (1)
| Title |
|---|
| End-point prediction of 260 tons basic oxygen furnace (BOF) steelmaking based on WNPSVR and WOA;Liming Liu等;《Journal of Intelligent & Fuzzy Systems》;2923-2937 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112668234A (en) | 2021-04-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112668234B (en) | Intelligent control method for converter steelmaking end point | |
| CN108845501B (en) | Blast furnace molten iron quality self-adaptive optimization control method based on lazy learning | |
| CN104899425A (en) | Variable selection and forecast method of silicon content in molten iron of blast furnace | |
| CN112949894B (en) | Output water BOD prediction method based on simplified long-short-term memory neural network | |
| CN111310348A (en) | Material constitutive model prediction method based on PSO-LSSVM | |
| CN107085371A (en) | Crude(oil)unit economic model forecast control method based on data-driven | |
| Fu et al. | A building energy consumption prediction method based on integration of a deep neural network and transfer reinforcement learning | |
| CN110908361B (en) | Fermentation process soft measurement method based on online twin support vector regression | |
| CN110245398B (en) | Soft measurement deep learning method for thermal deformation of air preheater rotor | |
| CN112947627A (en) | Temperature control method based on DDPG-fuzzy PID | |
| CN117289652A (en) | Numerical control machine tool spindle thermal error modeling method based on multi-universe optimization | |
| CN115310348A (en) | Stacking-based grouting amount integrated agent prediction model and prediction method | |
| CN109493921B (en) | Multi-agent model-based normal pressure rectification process modeling method | |
| CN111798023B (en) | Comprehensive coke ratio prediction method in steelmaking sintering production | |
| CN113204918A (en) | Modeling method for production energy consumption of cement kiln | |
| CN106647250B (en) | Based on offline optimization/online lookup table mode double-layer structure forecast Control Algorithm | |
| CN109101683B (en) | Model updating method for pyrolysis kettle of coal quality-based utilization and clean pretreatment system | |
| CN116880201A (en) | Water network channel state control system based on fuzzy neural network | |
| CN110909492A (en) | Sewage treatment process soft measurement method based on extreme gradient lifting algorithm | |
| CN116502750A (en) | Urban fuel gas daily load prediction method based on GRA-PSO-LSSVM | |
| CN113223634B (en) | Blast furnace molten iron silicon content prediction method based on two-dimensional self-attention enhanced GRU model | |
| CN114740713A (en) | Multi-objective optimization control method for wet flue gas desulfurization process | |
| CN103135444A (en) | Steel production energy consumption immune prediction control model | |
| CN116127345B (en) | Converter steelmaking process mode design method based on deep clustering generation countermeasure network | |
| CN113593657B (en) | Cement free calcium soft measurement system with quality target as guiding semi-supervised learning |
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 |